Marta Listewnik
Numer albumu 39330
Phrasal verbs in Welsh in relation to the
Welsh linguistic norm and contact with
English
Czasowniki frazowe w języku walijskim
w odniesieniu do walijskiej normy
językowej i kontaktu z językiem
angielskim
Praca doktorska napisana
na Wydziale Anglistyki
Uniwersytetu im. Adama Mickiewicza w Poznaniu
pod kierunkiem prof. UAM dr. hab. Michaela Hornsby’ego
Poznań, 2018
Acknowledgements
Hoffwn ddiolch i’m goruchwyliwr, yr Athro Michael Hornsby, am ei gymorth a chyngor
dros y blynyddoedd o ysgrifennu’r traethawd hwn a’i barodrwydd i ddarllen y gwaith ac
ateb fy nghwestiynau.
Hoffwn hefyd nodi fy nyled i’r holl gyfranogwyr i’r astudiaeth hon am eu hamser
a pharodrwydd i drafod eu hiaith ag ymchwiliwr o wlad bell: aelodau Cymdeithas yr Iaith
Gymraeg, grŵp ysgrifennu creadigol yn Aberystwyth, staff cylchgrawn Golwg, y Llyfrgell Genedlaethol, Llyfrgell y Dref yn Aberystwyth, Canolfan Bedwyr a chanolfan Cymraeg i Oedolion ym Mangor, ac athrawon ysgol uwchradd a gymerodd ran yn yr ymchwil.
Diolch arbennig i Martyna a Bryn Jones, Mared Thomas, Robin Farrar, Ifor Gruffydd,
Dr. Llion Jones, Owain Schiavone, Awen Schiavone, Hynek Janousek ac Ifor ap Glyn
am eu help gwerthfawr wrth ddod o hyd i gyfranogwyr ac adnoddau ar gyfer yr ymchwil.
Na koniec dziękuję mojej rodzinie i przyjaciołom za ogromne wsparcie, bez którego praca nie zostałaby ukończona, a w szczególności mojemu mężowi za nieskończoną
cierpliwość, bezcenne uwagi oraz entuzjazm, z jakim poświęca swój czas na naukę walijskiego.
2
OŚWIADCZENIE
Ja, niżej podpisana
Marta Listewnik
przedkładam rozprawę doktorską
pt. Phrasal verbs in Welsh in relation to the Welsh linguistic norm and contact with English (Czasowniki frazowe w języku walijskim w odniesieniu do walijskiej normy językowej i kontaktu z językiem angielskim)
na Uniwersytecie im. Adama Mickiewicza w Poznaniu
i oświadczam,
że napisałam ją samodzielnie.
Oznacza to, że przy pisaniu pracy, poza niezbędnymi konsultacjami, nie korzystałam
z pomocy innych osób, a w szczególności nie zlecałam opracowania rozprawy lub jej
istotnych części innym osobom, ani nie odpisywałam tej rozprawy lub jej istotnych części
od innych osób.
Jednocześnie przyjmuję do wiadomości, że gdyby powyższe oświadczenie okazało się
nieprawdziwe, decyzja o wydaniu mi dyplomu zostanie cofnięta.
(miejscowość, data)
(czytelny podpis)
3
Table of contents
ACKNOWLEDGEMENTS ........................................................................................... 2
TABLE OF CONTENTS ............................................................................................... 4
LIST OF TABLES ........................................................................................................ 11
LIST OF FIGURES ...................................................................................................... 14
ABBREVIATIONS OF DICTIONARIES AND CORPORA ................................... 16
INTRODUCTION ........................................................................................................ 17
Phrasal Verbs in Welsh – the present and past .................................................. 18
Phrasal verbs as a language contact feature...................................................... 23
Aims of the research and research questions. .................................................... 25
Structure of the thesis ......................................................................................... 27
Practical implications of the research ................................................................ 28
CHAPTER 1 : MODERN WELSH IN CONTACT WITH ENGLISH –
PERSPECTIVES, IDEOLOGIES AND ISSUES OF STANDARDISATION ....... 30
INTRODUCTION ............................................................................................................ 30
1.1. THE CURRENT SITUATION OF WELSH .................................................................... 31
1.1.1. Welsh between 1900 and present – an overview .......................................... 31
1.1.2. Changes in the linguistic landscape ............................................................. 35
1.2. STANDARD WELSH – LINGUISTIC NORM(S)? .......................................................... 39
1.2.1. Language ideologies, attitudes and issues of standardisation ..................... 40
4
1.2.1.1. Ideologies and attitudes ......................................................................... 40
1.2.1.2. Ideologies of standardisation ................................................................. 44
1.2.2. Standard and minority languages ................................................................. 48
1.2.3. Registers of Welsh – typologies .................................................................... 52
1.2.4. Written standard ........................................................................................... 55
1.2.4.1. Literary Welsh ....................................................................................... 56
1.2.4.2. Official Welsh and semi-formal varieties .............................................. 61
1.2.5. Spoken Standard ........................................................................................... 66
1.2.5.1. Conservative spoken standard – “pulpit” Welsh ................................... 67
1.2.5.2. Cymraeg Byw ......................................................................................... 68
1.2.5.3. Broadcast Welsh .................................................................................... 69
1.2.6. Lexical planning – dictionaries and terminology ......................................... 73
1.2.6.1. Geiriadur Prifysgol Cymru .................................................................... 74
1.2.6.2. Geiriadur yr Academi ............................................................................ 75
1.2.6.3. Other online dictionaries ........................................................................ 76
1.2.6.4. General printed dictionaries ................................................................... 78
1.2.6.5. Terminology dictionaries ....................................................................... 79
1.2.7. Standard and non-standard – further remarks ............................................. 80
1.3. WELSH AND ENGLISH IN CONTACT – THE INFLUENCE OF ENGLISH ON WELSH ...... 85
1.3.1. Language contact, language change and transfer........................................ 86
1.3.1.1. Language contact ................................................................................... 86
1.3.1.2. Contact-induced change and lexical transfer ......................................... 88
1.3.1.3. Borrowing and its integration ................................................................ 90
1.3.1.4. Classification of borrowings .................................................................. 92
1.3.1.5. Borrowing vs code-switching ................................................................ 94
1.3.2. English borrowings in Welsh – literature review ......................................... 96
1.3.3. Ideological stances towards borrowing from English – an overview .......... 99
1.4. CONCLUDING REMARKS ...................................................................................... 104
1.5. METHODOLOGICAL CONSIDERATIONS FOR THE PRESENT STUDY ......................... 107
CHAPTER 2 : PHRASAL VERBS IN WELSH – CROSS-LINGUISTIC
BACKGROUND AND SYNCHRONIC DESCRIPTION ...................................... 110
INTRODUCTION .......................................................................................................... 110
5
2.1. PHRASAL VERBS IN ENGLISH LINGUISTICS – AN OVERVIEW ................................. 111
2.1.1. Definitions ................................................................................................... 112
2.1.2. Popular notions about phrasal verbs.......................................................... 113
2.1.2.1. Informality ........................................................................................... 113
2.1.2.2. Replaceability by a single item ............................................................ 114
2.1.2.3. Idiosyncrasy ......................................................................................... 115
2.2. PHRASAL
VERBS IN LANGUAGE CONTACT SITUATIONS
–
EVIDENCE FROM OTHER
LANGUAGES ............................................................................................................... 117
2.2.1. Phrasal verbs in other languages in contact with English ......................... 117
2.2.2. Phrasal verbs in other Celtic languages..................................................... 118
2.3. PHRASAL VERBS IN WELSH LINGUISTICS ............................................................. 121
2.3.1. Fowkes 1945 ............................................................................................... 121
2.3.2. M. Jones (1979) and A. Thomas (1987)...................................................... 122
2.3.3. P.W. Thomas 1996 ...................................................................................... 123
2.3.4. Heinz 2003 and Asmus and Williams 2014 ................................................ 126
2.3.5. Rottet 2000 and 2005 .................................................................................. 127
2.3.6. Hirata 2012 ................................................................................................. 130
2.3.7. Other studies ............................................................................................... 131
2.3.8. Summary of literature on PVs in Welsh ...................................................... 132
2.4. PHRASAL VERBS IN THE PRESENT STUDY ............................................................. 133
2.4.1. Definition and characteristics of phrasal verbs in Welsh ........................... 134
2.4.2. Semantic classification – Idiomaticity ........................................................ 137
2.4.2.1. Criteria for distinguishing between compositional and idiomatic phrasal
verbs .................................................................................................................. 141
2.4.3. Syntactic classification of phrasal verbs .................................................... 144
2.4.3.1. Adverbial phrasal verbs ....................................................................... 144
2.4.3.2. Phrasal prepositional verbs .................................................................. 145
2.4.3.3. Prepositional verbs ............................................................................... 146
2.4.4. Language contact classification ................................................................. 147
2.4.5. Scope and focus of the present study .......................................................... 148
CHAPTER 3 : PHRASAL VERBS IN WELSH – A CORPUS STUDY ............... 150
3.1. INTRODUCTION ................................................................................................... 150
6
3.1.1. Aims of the study ......................................................................................... 151
3.1.2. Scope of the research .................................................................................. 151
3.2. ANALYSIS OF THE CORPUS .................................................................................. 152
3.2.1. The corpus................................................................................................... 152
3.2.2. Methods ....................................................................................................... 154
3.2.3. Quantitative analysis – results .................................................................... 156
3.2.3.1. Verbs and particles............................................................................... 156
3.2.3.2. Frequency count ................................................................................... 159
3.2.3.3. Syntactic categories ............................................................................. 160
3.2.3.4. Language contact categories and mode of narration ........................... 164
3.2.4. Stylistic markedness .................................................................................... 170
3.2.4.1. Fiction .................................................................................................. 170
3.2.4.2. Press ..................................................................................................... 175
3.3. CORRESPONDENCE
SEMANTICS
OF MEANINGS
–
A COMPARATIVE ANALYSIS OF PARTICLE
................................................................................................................ 180
3.3.1. A cognitive model for analysing the semantics of PV particles .................. 182
3.3.2. The particles ............................................................................................... 184
3.3.2.1. ALLAN and OUT ................................................................................ 184
3.3.2.2. I FYNY and UP ................................................................................... 187
3.3.2.3. YMLAEN, RHAGDDO, AR and ON ................................................. 189
3.3.2.4. I LAWR and DOWN ........................................................................... 192
3.3.2.5. I MEWN, I MEWN I and IN, INTO .................................................... 193
3.3.2.6. YN ÔL and BACK .............................................................................. 195
3.3.2.7. I FFWRDD, YMAITH, OFF, ODDI AR and AWAY, OFF ............... 195
3.3.2.8. HEIBIO and BY, ASIDE ..................................................................... 197
3.3.2.9. O GWMPAS, ROWND and (A)ROUND ........................................... 198
3.3.2.10. DRAW, DROS, DROSODD, and OVER, FOR, AWAY ................. 199
3.3.2.11. DRWODD, TRWY and THROUGH ................................................ 202
3.3.2.12. AR DRAWS and ACROSS ............................................................... 203
3.3.2.13. AR ÔL and AFTER ........................................................................... 204
3.3.2.14. Other prepositional particles .............................................................. 205
3.4. CORPUS STUDY – SUMMARY OF THE FINDINGS AND DISCUSSION ......................... 206
3.5. CONCLUSIONS ..................................................................................................... 212
7
CHAPTER 4 : PHRASAL VERBS IN WELSH DICTIONARIES AND TEACHING
MATERIALS .............................................................................................................. 213
4.1. INTRODUCTION ................................................................................................... 213
4.2. TEACHING MATERIALS AND COURSES ................................................................. 214
4.2.1. Analysis ....................................................................................................... 214
4.2.2. Summary of findings ................................................................................... 221
4.3. BOOKS AND DICTIONARIES OF IDIOMS ................................................................. 223
4.4. CYSILL ................................................................................................................. 225
4.5. DICTIONARIES – CASE STUDIES ........................................................................... 226
4.5.1. Edrych ymlaen ............................................................................................ 229
4.5.2. PVs with ymlaen conveying progress ......................................................... 230
4.5.3. Other construction with ymlaen: dod ymlaen, bod ymlaen, rhoi ymlaen ... 232
4.5.4. Edrych ar ôl ................................................................................................ 233
4.5.5. Native PPs: torri ar draws, dod ar draws and taro ar ............................... 234
4.5.6. Mynd heibio ................................................................................................ 235
4.5.7. Semi-idiomatic PVs: dod draw, edrych yn ôl, mynd allan gyda, torri i mewn
.............................................................................................................................. 235
4.5.8. Pleonastic PVs: eistedd i lawr, tyfu i fyny and torri i lawr ........................ 236
4.5.9. Calques and loanblends: troi i fyny, sortio allan, ffeindio allan ................ 237
4.5.10. Summary of the findings ........................................................................... 238
4.6. CONCLUSIONS ..................................................................................................... 239
CHAPTER 5 : THE ACCEPTABILITY OF PHRASAL VERBS AMONG
PROFESSIONAL SPEAKERS OF WELSH – A FIELD STUDY ........................ 242
5.1. INTRODUCTION ................................................................................................... 242
5.2. STUDY BACKGROUND ......................................................................................... 244
5.2.1. Lexical acceptability ................................................................................... 244
5.2.2. Previous research ....................................................................................... 244
5.2.3. Ideologies of standardisation...................................................................... 246
5.3. STUDY DESIGN .................................................................................................... 248
5.3.1. Aims of the study ......................................................................................... 249
5.3.2. Research hypotheses ................................................................................... 249
5.3.3. The sample .................................................................................................. 250
8
5.3.4. Methods ....................................................................................................... 253
5.3.4.1. The procedure ...................................................................................... 254
5.3.4.2. Questionnaire A (Aim 1) ..................................................................... 255
5.3.4.3. Questionnaire part B (Aim 2) .............................................................. 257
5.3.4.4. Personal questionnaire and interviews (Aim 3) ................................... 259
5.3.5. Qualitative and quantitative aspects of the study ....................................... 260
5.4. RESULTS ............................................................................................................. 261
5.4.1. The profile of the professional speakers ..................................................... 261
5.4.1.1. Extralinguistic data .............................................................................. 261
5.4.1.2. Use of Welsh in daily life .................................................................... 265
5.4.1.3. Reading, writing and use of Welsh-language media ........................... 266
5.4.1.4. Speakers’ perceived competence and confidence ................................ 267
5.4.1.1. Use of dictionaries and software .......................................................... 269
5.4.1.2. Use of borrowings ................................................................................ 271
5.4.1.3. Summary of data .................................................................................. 274
5.4.2. Questionnaire Part A .................................................................................. 275
5.4.3. Questionnaire Part B .................................................................................. 291
5.4.3.1. Results .................................................................................................. 291
5.4.3.2. Extralinguistic variables – statistical analysis ..................................... 298
5.4.4. Interview data ............................................................................................. 308
5.4.4.1. Acceptability and usefulness of borrowings ........................................ 308
5.4.4.2. Ideologies of purism ............................................................................ 316
5.4.4.3. The influence of English on Welsh vocabulary and structures ........... 322
5.4.4.4. Speakers’ views on phrasal verbs in Welsh ......................................... 325
5.4.4.5. Views on standard Welsh .................................................................... 330
5.5. SUMMARY AND DISCUSSION OF THE RESULTS ..................................................... 336
5.6. LIMITATIONS OF THE STUDY................................................................................ 343
5.7. CONCLUSIONS ..................................................................................................... 344
CONCLUSION ........................................................................................................... 347
WELSH PHRASAL VERBS – GENERAL CHARACTERISTICS............................................. 348
Definition and categorisation ............................................................................... 348
Transparent vs. idiomatic ..................................................................................... 349
9
Native vs. borrowed idiomatic phrasal verbs ....................................................... 349
Classification of transferred phrasal verbs .......................................................... 350
Productivity ........................................................................................................... 351
Transitivity ............................................................................................................ 352
Stylistic properties and acceptability .................................................................... 353
Phrasal verbs and linguistic norms and ideologies .............................................. 355
LIMITATIONS OF THE STUDY ...................................................................................... 358
RECOMMENDATIONS FOR FUTURE RESEARCH ............................................................ 360
PRACTICAL IMPLEMENTATION OF THE FINDINGS ........................................................ 361
ABSTRACT ......................................................................................................... 363
STRESZCZENIE ................................................................................................ 365
REFERENCES............................................................................................................ 368
APPENDIX A – THE CORPUS ................................................................................ 396
APPENDIX B – THE QUESTIONNAIRE ............................................................... 398
APPENDIX C – INTERVIEW SCHEDULE ........................................................... 404
10
List of tables
Table 1. Varieties of Welsh - typologies ........................................................................ 55
Table 2. Morphological variation of the preposition trwy ‘through’, after Jones (2013:
210). ....................................................................................................................... 136
Table 3. Dialectal variants of PV particles in Welsh .................................................... 136
Table 4. Dialectal or colloquial variants of particles and their frequencies in the corpus.
................................................................................................................................ 155
Table 5. First recorded uses of borrowed verbs constituting PVs in the corpus according
to GPC. ................................................................................................................... 156
Table 6. The most productive verbs in the corpus. ....................................................... 157
Table 7. Particles constituting PVs in the corpus and their frequencies. ...................... 158
Table 8. The 23 most frequent PVs in the corpus. ........................................................ 159
Table 9. The syntactic distribution of verb-particle constructions. .............................. 160
Table 10. The distribution of PVs in the corpus according to language contact
categorisation. ........................................................................................................ 164
Table 11. The distribution of PVs in narrative and conversation, according to language
contact categories. .................................................................................................. 165
Table 12. Examples of native APVs and PPVs in the corpus....................................... 166
Table 13. Prepositional PVs (PPs) in the corpus classified as native items or loan
renditions. ............................................................................................................... 168
Table 14. Pleonastic PVs in the corpus......................................................................... 169
Table 15. Number of PVs per 1000 words in novels. ................................................... 170
Table 16. Proportion of PVs used in dialogues in the analysed novels. ....................... 171
11
Table 17. Frequency and distribution of PVs in the press sample. ............................... 175
Table 18. Number and percentage of the most frequent PV in the corpus included in
Welsh dictionaries. ................................................................................................. 228
Table 19. The representation of selected PVs in dictionaries of Welsh. ...................... 228
Table 20. Geographical distribution of speakers according to their place of residence.
................................................................................................................................ 264
Table 21. Speakers’ use of Welsh and English in daily life. ........................................ 265
Table 22. Declared readership and the use of Welsh media among the participants. .. 267
Table 23. Dictionaries of Welsh and other tools used by the informants. .................... 270
Table 24. Percentage of speakers accepting the selected lexical items in Part A of the
questionnaire. ......................................................................................................... 275
Table 25. Percentage of speakers accepting PVs and alternative expressions in Part A of
the questionnaire in descending order. ................................................................... 289
Table 26. Part B – perceived acceptability of selected PVs in different registers ........ 292
Table 27. Sentences in the questionnaire containing torri i lawr ‘break down’. .......... 292
Table 28. Sentences in the questionnaire containing troi ymlaen ‘turn on’. ................ 293
Table 29. Sentences in the questionnaire containing troi i fyny ‘turn up’. ................... 294
Table 30. Sentences in the questionnaire containing edrych ar ôl ‘look after’. ........... 295
Table 31. Sentences in the questionnaire containing rhedeg allan o ‘run out of’. ....... 296
Table 32. Sentences in the questionnaire containing ffeindio allan ‘find out’. ............ 297
Table 33. The results of student’s t-test for differences between sexes in Register 4 and
total score with normal distribution. ...................................................................... 299
Table 34. The results of the Mann-Whitney U test for differences between sexes for
Registers 1, 2 and 3 with non-normal distribution ................................................. 299
Table 35. The results of Spearman's test for correlation between age and Questionnaire B
scores for Registers 1-4 and the total score. ........................................................... 300
Table 36. The results of student’s t-test for differences between speakers with secondary
and higher education in the total score. .................................................................. 300
Table 37. The results of the Mann-Whitney U test for differences between speakers with
secondary and higher education in Registers 1-4. .................................................. 301
Table 38. Acceptability of PVs in Part B in spoken registers with regard to speakers’ level
of education. ........................................................................................................... 302
12
Table 39. Acceptability of PVs in Part B in written registers with regard to speakers’ level
of education. ........................................................................................................... 303
Table 40. The results of Part B for Groups B and L compared with the mean. ........... 304
Table 41. The results of student’s t-test for differences between Groups B and L for the
total score and R2. .................................................................................................. 304
Table 42. The results of the Mann-Whitney U test for differences between Groups B and
L for R1, R3 and R4 with non-normal distribution. ............................................... 305
13
List of figures
Figure 1. Exchangeability of verb and particle in English PVs (after Thim 2012: 14) 139
Figure 2. Exchangeability of verb and particle in Welsh PVs. ..................................... 139
Figure 3. Network of meanings of OUT according to Rudzka-Ostyn (2003:41). ........ 184
Figure 4. Network of meanings of UP according to Rudzka-Ostyn (2003: 103). ........ 187
Figure 5. Age of the speakers. ...................................................................................... 262
Figure 6. Geographical distribution of the participants according to their place of
residence. ................................................................................................................ 263
Figure 7. Answers to the question “Does it happen that you cannot think of a Welsh word
you need?”.............................................................................................................. 269
Figure 8. Declared use of borrowings in speech and writing. ...................................... 273
Figure 9. The verbs chosen by the speakers in Sentence 1 (dod yn ôl/dychwelyd). ..... 277
Figure 10. The verbs chosen by the speakers in Sentence 10 (eistedd/eistedd i lawr). 278
Figure 11. The verbs chosen by the speakers in Sentence 2 (tyfu/tyfu i fyny). ............. 279
Figure 12. The verbs chosen by the speakers in Sentence 6 (helpu/ helpu allan/ helpu mas/
cynorthwyo). ........................................................................................................... 280
Figure 13. The verbs chosen by the speakers in Sentence 3 (digwydd/mynd ymlaen). 281
Figure 14. The verbs chosen by the speakers in Sentence 4 (darganfod/gweithio
allan/gweithio mas). ............................................................................................... 282
Figure 15. The verbs chosen by the speakers in Sentence 11 (diffodd/ troi i ffwrdd/troi
bant/troi off). .......................................................................................................... 283
Figure 16. The verbs chosen by the speakers in Sentence 7 (dysgu/ pigo i fyny/pigo lan/
cael crap ar). .......................................................................................................... 284
14
Figure 17. The verbs chosen in Sentence 5 (disgwyl yn eiddgar, edrych ymlaen at,
edrych ymlaen i). .................................................................................................... 286
Figure 18. The verbs chosen in Sentence 12 (sefyll o’r neilltu, eistedd yn ôl). ............ 286
Figure 19. The verbs chosen in Sentence 9 (parhau/bwrw ymlaen/cario ymlaen)....... 288
Figure 20. The verbs chosen in Sentence 8 (dod ar draws/ taro ar). ........................... 288
Figure 21. Part B – perceived acceptability of selected PVs in different registers. ...... 291
Figure 22. Mean results for the total score according to education. ............................. 301
Figure 23. Mean results for R3 scores according to education. .................................... 302
Figure 24. Mean total score in Groups L and B............................................................ 305
Figure 25. Percentile distribution of R1 scores for Groups B and L, ........................... 306
Figure 26. Percentile distribution of R3 scores for Groups B and L. ........................... 306
15
Abbreviations of dictionaries and corpora
CS
Collins Spurrell Welsh Dictionary
CEG
Cronfa Electroneg o Gymraeg [Electronic Corpus of Welsh]
CCCC
Corpws Cyfochrog Cofnod y Cynulliad [Parallel Corpus of the Assembly
Proceeding]
GA
Geiriadur yr Academi: the Welsh Academy English-Welsh Dictionary
GB
Geiriadur Bangor
GM
Y Geiriadur Mawr
GPC
Geiriadur Prifysgol Cymru: a Dictionary of the Welsh Language
Gom
Geiriadur Cymraeg Gomer
Gwe
Gweiadur
PMWD
The Pocket Modern Welsh Dictionary
TSD
University of Wales Trinity Saint David Dictionary
WLD
The Welsh Learner's Dictionary
16
Introduction
The inspiration for writing this doctoral thesis came from my own experience of attending
an intensive course of practical Welsh as a student at the Department of Celtic Languages
and Literatures (now Centre for Celtic Studies) at Adam Mickiewicz University in Poznań. After several years of studying the language, I became interested in idiomatic verbparticle constructions called phrasal verbs, which most advanced Polish speakers of English are familiar with and usually consider rather challenging to learn. Thinking of phrasal
verbs as something unique to the English language, I was surprised to encounter the constructions also in Welsh, but soon discovered that the majority of them appeared to be
word-for-word translations from English. However, what attracted my attention the most
were different approaches to these constructions on the part of my Welsh teachers. Whilst
native speakers of Welsh without formal education in linguistics used phrasal verbs extensively in their speech and considered them entirely natural, another teacher from outside Wales forbade the use of phrasal verbs as ‘incorrect’ calques from English. This view
seemed to be supported by the fact that verb-particle constructions were not included or
discussed in Welsh dictionaries, grammars and materials for learners available to me at
the time. Yet, the ubiquity of Welsh phrasal verbs appearing not only in spoken language,
but also in written texts, stood in sharp contrast with a universal prescriptive rule against
the use of these constructions and called for a more critical investigation.
This personal experience prompted me to ask questions on the status of phrasal
verbs in modern Welsh and investigate whether the linguistic norm I encountered as
a learner is indeed reflected in the standard language used today. These questions appear
valid both from a theoretical perspective of Welsh linguistics, and from a practical point
of view of language teaching.
17
Phrasal Verbs in Welsh – the present and past
Before outlining the aims and structure of the thesis, I believe it worthwhile to present the
reader with some illustrative examples of the use of phrasal verbs in Welsh, as well as
opinions of them voiced recently and in the past. This will help contextualise the investigated issues and delineate tensions between the linguistic norm and the usage of phrasal
verbs in modern Welsh.
To begin with the term ‘phrasal verb’ itself, it should be noted that it has been
defined in a multitude of ways, which will be discussed in detail in Chapter 2 of the thesis.
For the purpose of this introduction, phrasal verbs are defined as combinations of a verb
and a particle – an adverb, preposition, or both – which form a single unit in speakers’
minds. Consequently, speakers associate phrasal verbs primarily with idiomatic constructions rather than transparent phrases pertaining to movement in space (see 2.1.1.). Since
my research is going to focus on idiomatic phrasal verbs in Welsh as a phenomenon of
language contact, the examples in the present section also belong to the category of idiomatic constructions.
As phrasal verbs are extremely widespread in the spoken Welsh of today, they are
easily encountered outside of their natural occurrence, such as in scripted talk. Here, the
chosen example is from a single episode of a popular Welsh detective series Y Gwyll
(Evans 2013) set in Ceredigion, Mid-Wales, a region where just below half of the population is Welsh-speaking. Despite the bilingual sociolinguistic reality, the series portrays
a fictitious entirely Welsh-speaking world, given that it was produced in two language
versions, Welsh and English (under the title Hinterland), each of which is near monolingual. Thus, in the first episode of the Welsh version the characters speak only Welsh and
their language contains very few recent loanwords from English and no instances of codeswitching. Yet, dialogues in the hour and a half long episode contain 20 idiomatic phrasal
verbs, of which some examples are marked in bold, listed below:
(1)
Oeddwn i'n edrych ymlaen yn fawr i gwrdd â chi.
‘I was looking forward very much to meet you.’
(2)
(...) doedd o byth moyn tyfu lan.
‘He never wanted to grow up.’
(3)
Cwympon ni mas.
18
‘We fell out.’
(4)
‘Naethoch chi redeg allan o baent?
‘Did you run out of paint?’
(5)
Roedd hi'n credu bod angen bwrw'r cythraul mas o'r plant.
‘She believed that you had to drive the devil out of the children.’
(6)
Licen ni gario ymlaen 'da cwestiynu ac os ydy o'n torri i lawr…
‘I would like to carry on with interrogation and if he breaks down again...’
(7)
Oedd hi'n edrych ar fy ôl i.
‘She looked after me’.
(8)
Dim clem beth oedd yn mynd ymlaen yn y lle 'ma.
‘No idea what was going on in that house.’
(9)
(…) fel basai'n ffeindio ei hun neud lan am golli Awen.
‘As if she wanted to make up for losing Awen.’
(10) Dere mlaen, Mared.
‘Come on, Mared.’
Despite the fact that authors of the script strove to make the series almost monolingual,
they do not seem to avoid this type of idiomatic phrases, which appear to be word-forword translations from English. What is also worth noting is that even such a small and
random sample contains one example of a Welsh phrasal verb which is not directly translatable: bwrw mas ‘drive out’ (5), literally ‘strike out’. The example of the Y Gwyll episode leads us to two observations. Firstly, that the use of these constructions in speech is
unhindered even in forcedly monolingual contexts and, secondly, that the phenomenon
of phrasal verbs goes beyond the process of calquing.
The integration of idiomatic verb-particle constructions in Welsh is not a recent
phenomenon, as there is abundant evidence of them used extensively in both spoken and
written language well over a hundred years ago. A great number of idiomatic phrasal
verbs can be found for example in one of the earliest Welsh novels, Rhys Lewis by Daniel
Owen (1885). As noted by Hincks in his study of purism in 19th-century Wales, the language used by Daniel Owen was unhindered by ideologies of linguistic purism popular
of his time; the novelist would not limit his language to the formal literary register, but
reflect the speech of the Welsh countryside which was already strongly influenced by
English (Hincks 2007: 63–65). Consequently, the first 200 pages of this novel contain
19
over 100 instances of idiomatic phrasal verbs1 occurring primarily in dialogue, but also
in the narrative. Among the latter, one may find the same examples which have appeared
in the episode of Y Gwyll described above:
(11) Rhywbryd yn y cyfnod hwn deuthum i edrych yn mlaen at ddiwrnod y cyflog (...)
(Owen 1885:31).
‘Sometime at that period I began to look forward to the pay day.’
(12) Am oddeutu awr o amser, nid oedd yn ymddangos i mi fod dim gwaith yn myned
ymlaen yn yr ysgol (Owen 1885: 46).
‘For about an hour, it did not seem to me that any work was going on in the school.’
(13) Mae sŵn eu clogs ar y ffordd galed (...) yn ailennyn hiraeth yn nghalon ambell
weddw ieuanc, yr hon ddaw i’r drws a phlentyn ar ei braich, ac un arall yn cydio
yn ei ffedog, i edrych ar eu holau, fel pe byddai hi o hyd yn disgwyl i John ddyfod
yn ôl (Owen 1885:194).
‘The sound of their clogs on the hard road (...) rekindles yearning in the heart of a
young widow, the one who comes to the door with a child on her arm and another
one holding to her apron, to look after them, as if she were still waiting for John to
come back.’
While these examples provide evidence for the widespread use of phrasal verbs at the end
of the nineteenth century, prescriptive ideological stances against using phrasal verbs had
also existed at that time, as can be seen in the 1889 article “Plicio gwallt yr hanner Cymry”
[Plucking off the hair of the half Welsh] by Emrys ap Iwan (1848-1906). The famous
scholar criticises journalists who “publish in bad Welsh before making an effort to learn
better Welsh” and use an excessive amount of English words and idioms, which turns
their Welsh into “translated English” (ap Iwan 1939: 107–111)2. Ap Iwan proceeds to
give lists of “slovenly” and “un-Welsh-like phrases” (ymadroddion annestlus/anghymreigiaidd), which include ten sentences with phrasal verbs, for example:
“Aeth y tân allan”, yn lle Fe ddiffoddodd y tân. (…)
“Y maent bob amser yn ei rhedeg ef i lawr”, yn lle … yn lladd arno (…)
1
These estimations are derived from a preliminary corpus study conducted according to the methodology
presented in Chapter 3 of the thesis. The study of the novel was, however, excluded from the final project.
2
All the translations from Welsh sources are mine, ML.
20
“Os try rhywbeth i fyny, dywedwch,” yn lle Os digwydd rhywbeth. (…)
“Wedi eu llosgi i fyny,” yn lle Wedi eu llwyr losgi. (ap Iwan 1939: 120-122)
‘“The fire went out”, instead of The fire extinguished.’
“They always run him down”, instead of… criticise him [lit. kill on him]’.
“If something turns out, say”, instead If something happens.’
“Having burned out”, instead of Having burned completely.’
Ap Iwan’s remarks are clearly influenced by a puristic ideology, which condemns calqued
constructions as “un-Welsh”, stemming from the “laziness and carelessness” of writers
(1939:108). Perceiving phrasal verbs as a threat to natural Welsh idioms, the author went
as far as to deem those who use Anglicised language as “half Welsh”.
This kind of discourse can also be encountered today, for instance, in publicly
made comments regarding the language used in the media. As noted by Ball and Müller
the “complaint tradition” of writing to the Welsh press has been on the rise since the
establishment of Welsh broadcast media (Ball and Müller 1992: 264) as a way of expressing the ideology of purism. An example worth quoting is a reader’s letter published several years ago in a Welsh-language newspaper Y Cymro (18 Mar. 2009):
As one who worries about the obvious deterioration in the standard of Welsh among many
of our young people and in fact among some of the older ones on the media, may I have
some space to list some un-Welsh and unacceptable forms, and put the native idioms in
brackets, hoping that the guilty ones will pay some attention and improve! 3
The reader proceeds to give this list of ten commonly used expressions, six of which are
phrasal verbs, and their native equivalents in brackets:
(14) Colli allan (ar eu colled)
‘Miss out’
(15) Rhedeg allan (mynd yn brin o...)
‘Run out of’
(16) Marw allan (darfod, crebachu/edwino/dod i ben)
‘Die out’
(17) Cymryd drosodd (cymryd yr awenau/achub y blaen ar..)
“Fel un sy’n poeni am y dirywiad amlwg yn safon y Gymraeg ymysg llawer o’n pobl ifanc ac yn
wir ymysg rhai hŷn ar y cyfryngau, a gaf ychydig o ofod i restru rhai o’r ffurfiau anghymreig ac
annerbyniol, a rhoi’r priod-ddull cynhenid mewn cromfachau, yn y gobaith y bydd yr euog yn talu
sylw ac yn gwella!”
3
21
‘Take over’
(18) Ffeindio allan/mas (darganfod, dod i wybod am..)
‘Find out’
(19) Pethau’n edrych i fyny/lan (pethau’n argoeli’n dda)
‘Things are looking up’
The reader concludes:
I could add a lot more, but I will keep silent for the time being by noting the latest abomination I heard on S4C News, that is dringo i lawr [‘climb down’] instead of syrthio ar ei
fai [‘acknowledge his fault’, lit. ‘fall on his fault’] and that by an experienced journalist.
Sad.4
This approach highly resembles the opinions of Emrys ap Iwan, in that the reader
describes calqued phrasal verbs in strongly negative terms, such as “abomination”, “obvious deterioration”, “un-Welsh and unacceptable”, while people who use them are considered “guilty” of not using native Welsh idioms. Such voices are publicly raised by
a small number of people and although they might not have a tangible effect on speakers’
behaviours, they are important indicators of the changing linguistic norm and sensitivity
of some speakers to language contact phenomena (Ball and Müller 1992: 264). A major
point of concern for both ap Iwan and the contemporary reader is the endangered native
idiomaticity of Welsh in view of the rising influence of English. This has been aptly summarised some years ago by Hincks in his article for the magazine Barn (Hincks 1993):
By now we have to ask whether what is spoken by many Welshmen is Welsh or English
translated into Welsh. What guidelines can we follow while deciding what is acceptable?
It is obvious there is space for borrowed idioms which settled in the language a long time
ago, e.g. next door, of course, blow away, look forward, be worthy, but at the same time is
there place, even in spoken language for slavish translation (…)? There are other idioms
which are based on English but are accepted by many writers, e.g. take into consideration,
catch up with and even stand to reason. With many idioms, therefore, we cannot say they
are ‘correct’ or ‘incorrect’, but rather that they are ‘traditional’ or ‘untraditional’, ‘English-
“Mi fedrwn ychwanegu llawer mwy, ond tawaf am y tro trwy nodi’r erchylltra diweddaraf a glywais ar
Newyddion S4C yn ddiweddar sef ‘dringo i lawr’ yn lle ‘syrthio ar ei fai’, a hynny gan newyddiadurwr
profiadol. Trist.”
4
22
like’ or ‘Welsh-like’, ‘acceptable in literary language’ or unacceptable in literary language’,
etc.5
The largely anecdotal evidence presented above is intended to illustrate the major
issues regarding the nature of Welsh phrasal verbs in Welsh. On the one hand, it cannot
be doubted that the constructions are widely used and at least a number of them appear to
be well integrated in the language of native speakers. On the other hand, one may observe
a continuing discourse forbidding the use of phrasal verbs as calques from English, resulting in their absence from linguistic description of the language. These facts make
phrasal verbs in Welsh an extremely interesting case which has been remarkably little
researched thus far.
Phrasal verbs as a language contact feature
Bearing in mind the issue of the status of phrasal verbs as alleged borrowings, the present
research can be placed within studies of language contact between Welsh and English,
which is a relatively new area of study. The small number of pioneering publications on
the influence of English on Welsh has touched on a range of aspects, such as language
change and revitalisation (Jones 1998), code-switching and loanword integration (Deuchar 2005, 2006; Deuchar and Davies 2009; Stammers and Deuchar 2012; Parafita Couto
et al. 2015, Deuchar et al. 2016), the Welsh of heritage speakers (Boon 2014), as well as
cross-linguistic influence on phonetics and phonology (Morris 2013, Buczek-Zawiła
2014; Morris et al. 2016; Mayr et al. 2017), morphology (Phillips 2007), and structures
(Davies 2010; Hirata 2012; Nicoladis and Gavrila 2015) (see literature review 1.3.2.)6.
Phrasal verbs have attracted very little attention in Welsh linguistics thus far. They
are discussed in only one grammar of Welsh, Gramadeg y Gymraeg by P.W. Thomas
“Erbyn hyn rhaid gofyn ai Cymraeg a siaredir gan lawer o Gymru ynteu Saesneg wedi’i throsi i’r Gymraeg. Pa ganllawiau a allwn eu dilyn wrth benderfynu beth sy’n dderbyniol? Mae’n amlwg fod lle i briodddulliau benthyg sydd wedi hen ymgartrefu yn yr iaith, e.e. drws nesaf, wrth gwrs, chwythu i ffwrdd, edrych
ymlaen, bod yn werth, ond ar yr un pryd a oes lle, hyd yn oed yn yr iaith lafar, i drosi’n slafaidd(...)? Mae
priod-ddulliau eraill sy’n seiliedig ar y Saesneg, ond a dderbynnir gan lawer o ysgrifenwyr, e.e. cymryd i
ystyriaeth, dal i fyny â, a hyd yn oed sefyll i reswm. Gyda llawer o briod-ddulliau, felly, ni ellir dweud eu
bod yn ‘gywir’ neu ‘anghywir”, ond yn hytrach eu bod yn ‘draddodiadol’ neu ‘anhraddodiadol’, yn ‘Seisnigaidd” neu’n ‘Gymreigaidd”, yn ‘dderbyniol yn yr iaith lenyddol’ neu’n ‘annerbyniol yn yr iaith lenyddol’ etc.”
6
For an overview of studies of Welsh English see Durham and Morris (2016: 14–17).
5
23
(1996) in a long footnote, in which transparent and semi-idiomatic constructions are described as informal loanwords from English, while idiomatic phrasal verbs are almost
entirely omitted (for a review see 2.3.3). The only in-depth academic work on Welsh
phrasal verbs is an article by Rottet (2005). Having presented the key properties of Welsh
phrasal verbs, Rottet confirms their presence in contemporary written Welsh, illustrating
this by a number of examples from the press and fiction. This author also outlines directions for future diachronic studies on phrasal verbs in Welsh, comparing them with other
Brythonic languages. He presents evidence for the claim that the constructions are not
always simple calques from English, as they have been present in the Welsh language for
centuries. He also proposes a classification of Welsh phrasal verbs borrowed from English based on their type of transfer mechanism. The final part of his paper analyses how
phrasal verbs are represented in several selected dictionaries, a grammar and exercise
book and in other metalinguistic contexts, indicating a pedagogical norm prescribing the
use of some idiomatic constructions. Rottet’s short but comprehensive study approaches
the phenomenon of phrasal verbs in Welsh from a range of different perspectives and has
provided an excellent basis for the present extended research on the same subject (for a
review see 2.3.5).
Other than offering a purely linguistic description, the thesis will examine tensions
between the norm and usage regarding phrasal verbs in Welsh and investigate them in
view of linguistic ideologies manifested in the Wales of today. Linguistic ideologies are
another relatively little researched area of Welsh linguistics, although a small number of
studies have appeared in recent years, including Musk (2006, 2010, 2012) and Robert
(2011, 2013). The field of language ideologies is closely related to studies in language
attitudes. Studies on Welsh in this field in the last decade include Robert (2009), Williams
(2009), Davies et al. (2010), Morris (2014) and Owen (2018). As the present thesis is
oriented towards investigating a linguistic feature itself rather than purely social aspects
of language use, issues of ideology and attitudes will be considered as secondary in view
of the main aims of the research (see 1.2.1. for a detailed discussion).
24
Aims of the research and research questions.
Building on the previous work by Thomas (1996) and Rottet (2005) this study hopes to
provide fresh insights into the features and use of phrasal verbs in contemporary Welsh
and their acceptability within the linguistic norm. The short description of phrasal verbs
in Thomas is not devoid of shortcomings and rather unsystematic (2.3.3.). In turn, the
nature of Rottet’s (2005) pioneering study is largely exploratory. The author quotes examples of phrasal verbs from different sources and outlines major issues regarding acceptability and integration of these constructions. However, his observations are based on
selected examples encountered in a fairly random sample of texts rather than a rigours
corpus-based study which would supplement intuitive observations with some quantitative data. Moreover, Rottet’s observations on the acceptability of the constructions were
inferred from comments found in metalinguistic sources only and not supported by data
obtained directly from speakers. Finally, in the course of the last twenty years since his
article was written, especially with the advent of the Internet and online resources, major
changes have occurred in Welsh lexicography and teaching, providing new materials to
investigate.
In view of the above, the key aim of the thesis is to fill a major gap in Welsh
linguistics by providing a preliminary, yet comprehensive description of phrasal verbs
with focus on their acceptability within Welsh linguistic norm in view of contact with
English. The study applies mixed methodologies in order to explore different aspects of
phrasal verbs and provide the complementarity of findings, that is “to measure overlapping but also different facets of a phenomenon, yielding an enriched, elaborated understanding of that phenomenon” (Greene et al. 1989: 258). The main research questions are:
1. What are the major characteristics of phrasal verbs in Welsh?
Aiming to systematise basic information on phrasal verbs, which have not been thoroughly described in Welsh linguistics thus far, the thesis will examine their semantic
and syntactic properties and productivity on the basis of available literature and an
original corpus study. Although the study focuses on written Welsh, the analysis will
also include written representations of spoken language and the acceptability of
phrasal verbs therein. I will also propose a categorisation of Welsh phrasal verbs with
regard to contact with English and mechanisms of transfer, expanding on Rottet’s
25
(2005) typology, in particular by adding idiomatic prepositional constructions (Chapters 2 and 3).
2. Are Welsh phrasal verbs integrated in the standard written language of today?
The level of standardisation of phrasal verbs which is the focus of the thesis will be
investigated by examining their frequency and stylistic markedness in the written language (Chapter 3) as well as acceptability measured in a field study on professional
speakers of Welsh (Chapter 5), providing new quantitative and qualitative data.
3. How are Welsh phrasal verbs represented in contemporary grammars, teaching materials and dictionaries?
This question will be examined by a study of representation of phrasal verbs in contemporary grammars and linguistic works (Chapter 2) and normative sources, primarily teaching materials and dictionaries (Chapter 4), expanding the corpus investigated
by Rottet (2005), in particular with teaching and on-line materials, which are of great
importance today.
4. What ideologies towards phrasal verbs are manifested by Welsh scholars and proficient speakers of Welsh?
As a final issue, the thesis will address the question of possible motivations for accepting or rejecting phrasal verbs by authors of normative sources (Chapter 4) and
proficient speakers of Welsh (Chapter 5), thus placing the research within the wider
context of standardisation and providing insights for future studies in language ideologies and attitudes.
The study spans various areas of linguistics, such as semantics, stylistics, corpus studies,
lexicography, and language contact in order to conduct a multi-perspective synchronic
analysis of phrasal verbs in Welsh offering a basis for future studies on this subject. Unsurprisingly, the scope of the thesis will not allow to go in many potentially interesting
directions, such as diachronic analysis or studies of natural-occurring talk. It should be
noted in particular that due to the focus on linguistic norm, the analysis will be devoted
primarily, though not exclusively, to written language.
26
Structure of the thesis
This thesis contains five chapters. Chapter 1 begins with a review of the current situation
of the Welsh language in order to provide the background to the discussion which follows.
The second part of the chapter explores the concepts of language ideologies, attitudes,
standardisation processes and the notion of standard Welsh. The third part of the chapter
is concerned with issues of language contact between Welsh and English. It presents key
terms and definitions relevant to this research and discusses ideologies which influence
the Welsh linguistic norm, focusing on the changing acceptability of English borrowings.
Finally, implications of the theoretical background for the methodology applied in the
thesis are discussed.
Chapter 2 formulates the basis for the empirical studies in the subsequent chapters
by defining and describing Welsh phrasal verbs based on the existing sources. It begins
with a review of the literature regarding phrasal verbs in English and Welsh and places
the phenomenon of calquing in a wider context, providing comparative data with other
languages in intense contact with English. This chapter establishes an essential division
between transparent and idiomatic phrasal verbs and arrives at a definition and classification of idiomatic phrasal verbs used in subsequent chapters on the basis of studies by
Rottet (2000, 2005) and parallel studies on phrasal verbs in Irish.
Chapter 3, the first of the analytic chapters, presents a corpus of written Welsh
created for the purpose of this dissertation, comprised of works of fiction and press materials envisaged as a representative sample of semi-formal registers of Welsh. The quantitative results of the corpus analysis show the frequencies of phrasal verbs according to
text type, syntactic categories and contact-related classification. The grammatical description of Welsh phrasal verbs found in the literature is successively compared with
observations derived from corpus data. The second part of the chapter is a qualitative
analysis of corpus texts focusing on the examples of stylistic markedness which indicates
the acceptability of phrasal verbs within the linguistic norm. The third part is a semantic
analysis of particles found in the corpus. Based on Rudzka-Ostyn’s (2003) cognitive
model, this section compares corpus data with lexicographic materials and identifies extensions of particle meanings which are most likely to have emerged due to contact with
English.
27
Chapter 4 discusses the description of phrasal verbs in Welsh teaching materials
and dictionaries. The first part of this chapter reviews a number of pedagogical resources,
demonstrating the existence of a prescriptive norm which discourages the use of some
phrasal verbs as direct translations from English. The second part of the chapter examines
dictionary entries for the 25 most frequent phrasal verbs in the corpus, investigating the
consistency of their representation in lexicographic sources.
Finally, Chapter 5 presents the results of a field study conducted in 2016 and 2017
on 55 professional speakers of Welsh in seven groups: a) librarians, b) writers, c) school
teachers d) Welsh for Adults courses tutors, e) journalists f) staff of a research centre
responsible for preparing Welsh-language materials g) members of an organisation campaigning for the Welsh language. The study consisted of a questionnaire, which investigated the acceptability of phrasal verbs among speakers who are familiar with the linguistic norm, as well as semi-structured interviews which touched upon more general issues
related to borrowing and standardisation. The results of the field study show variation in
acceptability of phrasal verbs depending on the type of verb and the used register. They
also illustrate the complexity of opinions on the changing linguistic norm, while pointing
to potential factors which shape speakers’ beliefs on that matter.
It is hoped that the present investigation of the phenomenon of phrasal verbs will
contribute to the description of modern Welsh language and the changes it has undergone
due to extensive contact with English.
Practical implications of the research
Apart from its scholarly value, the study is expected to have practical implications. As
my experience has shown, the discrepancy between norm and usage can be a source of
confusion and difficulty for learners, who are usually eager to know whether they ‘are
allowed’ to use particular forms. A norm which is divorced from reality is bound to create
divisions between the language of native speakers and learners, which may be precarious
especially in view of the current vulnerable situation of Welsh. While making no claims
to judgments whether phrasal verbs in Welsh should be deemed ‘correct’ or ‘incorrect’,
I believe it is crucial to gather evidence for a realistic presentation of the contemporary
28
language, with the hope that by providing the basis for an accurate and systematic description of verb-particle constructions in modern Welsh, the present research will contribute to the preparation of future teaching resources, grammar books and dictionaries.
29
Chapter 1: Modern Welsh in contact with English –
perspectives, ideologies and issues of standardisation
Introduction
The purpose of this chapter is to present an overview of the sociolinguistic situation in
Wales, with a focus on the level of standardisation of the Welsh language and issues
connected with bilingualism and language contact. The first section describes the general
situation of the Welsh language and future perspectives for the language’s survival and
revitalisation. It outlines major factors influencing linguistic changes which have taken
place in Wales, providing the context for the discussion of standardisation to follow. The
second part of the chapter defines and discusses the notions of language ideologies, attitudes and other concepts related to language standardisation with a focus on minority
languages. Subsequently, by analysing various registers of Welsh and channels through
which linguistic norms are implemented, I attempt to establish the degree of standardisation of the Welsh language of today. This will provide the background for the discussion
of the status of phrasal verbs in standard varieties of Welsh in the later chapters. The third
part of the chapter concentrates on language contact and the phenomena of transfer between Welsh and English related to phrasal verbs. This is followed by an overview of the
literature on the influence of English on Welsh from a linguistic perspective, as well as
from the point of view of ideologies manifested by Welsh speakers throughout the centuries regarding the English element in Welsh. This provides a broader context in which the
investigation of the phenomenon of Welsh phrasal verbs can be placed. In the final section, I discuss some theoretical and methodological issues that have implications for the
nature and scope of subsequent analytical chapters.
30
1.1. The current situation of Welsh
1.1.1. Welsh between 1900 and present – an overview
Welsh is a Celtic language of the Indo-European family, currently spoken by over half
a million people, most of whom live in Wales and other parts of the United Kingdom.
There is also a small Welsh-speaking community in the Chubut province in Patagonia,
Argentina7, and a “broad but thin” Welsh diaspora in the United States (Coupland and
Aldridge 2009:10).
The sociolinguistic landscape of Wales has undergone immense changes in the
course of the last two hundred years with considerable population growth and a rapid
language decline. At the end of the 19th century, there were almost one million Welsh
speakers, who comprised 50% of the population, with over 50% of them being monolingual. The years 1871-1921 marked a dramatic decline in the number of speakers of Welsh.
The process was particularly rapid in the south, where some areas quickly became solely
English-speaking due to high numbers of migrants from England and the high prestige of
English (Löffler 2008: 352). The heartlands in the north and west of Wales did not remain
unaffected either, gradually changing into bilingual areas.
By 1981, the number of Welsh speakers had dropped to half a million with practically no monoglots left (Morris 2010b: 82; Jones 2012: 10). Currently, Wales has a
population of around 3.1 million people, 19% of whom (ca. 560,000 speakers) are Welshspeaking (Gwyn Lewis 2015: 149), a decrease of 1% in the number of speakers in the last
decade according to the National Census. Importantly, these figures express the declared
ability to speak the language, while the percentage of the population that speak Welsh
daily and can speak more than just a few words of Welsh has been estimated by the Welsh
Government at 10% (Welsh Government 2017: 11).
Accordingly, as a minority language with a decreasing numbers of users, Welsh
has been classified by the UNESCO Atlas of the World’s Languages in Danger as “vulnerable” (Moseley ed. 2010). However, it cannot go unnoticed that the pace of the demographic decline has been slowed down in recent decades due to multiple revitalisation
7
It originated in the establishment of Y Wladfa, a Welsh colony in 1895. Johnson (2013) estimates the
number of Welsh speakers in Chubut at no more than 4000.
31
efforts and the increasingly official status of the language. The critical period for this
reverse trend began in the 1960s when a series of movements, political initiatives and
campaigns were launched in order to maintain and secure the future of Welsh. Among
these ground-breaking events was “Tynged yr iaith” [The fate if the language], a radio
speech by Saunders Lewis, which inspired the establishment of Cymdeithas yr Iaith Gymraeg (Welsh Language Society), an activist group exercising pressure on the authorities
regarding such areas as access to Welsh-language media, education and services.
Indeed, the last fifty years saw major developments in these fields. Beginning
with the Welsh-language media, BBC Radio Cymru and the television channel S4C began
broadcasting in 1977 and 1982, respectively. In 2018, a second Welsh-language radio
channel BBC Radio Cymru 2 launched broadcasts for several hours a day as an alternative
morning programme for younger audience. Regional radio stations also offer programmes
and music in the Welsh language.
The development of Welsh-medium education began with the opening of the first
Welsh medium private primary school in Aberystwyth in 1939, followed by the first official Welsh medium school in Llanelli established in 1947. While these schools originally catered for the needs of native speakers, in subsequent years they became increasingly attractive for parents and children from English-speaking families (Redknap 2006:
4-5). Another landmark was the introduction of the National Curriculum in 1988 when
Welsh became a compulsory subject to be studied by all pupils in Wales aged 7-16, a core
subject in Welsh-medium schools and a foundation subject in others (Morris 2010b: 81)8.
In consequence, Welsh as the Second Language became a statutory subject in the National
Curriculum in 1990 (Welsh Government 2013:1). This created a potential for a substantial
group of new speakers of Welsh.
The official status of the language was gradually improving as well. In 1993, the
first Welsh Language Act gave Welsh and English equal status in the public sector and
“Core subjects are English, mathematics and science (and Welsh in Welsh-speaking schools) and are
mandatory; Foundation subjects are also compulsory but are given less lesson time in the curriculum than
Core subjects. Welsh became a compulsory subject for all pupils in Wales at Key Stages 1, 2 and 3 (i.e. up
to age 14) in 1992. In 1993 it became a compulsory subject at Key Stage 4; this meant that all pupils in
Wales studied Welsh (either as a first or a second language) for 11 years, from the ages of 5 to 16. From
2008, the National Curriculum for 3- to 7-year-olds has been superseded by the Foundation Phase programme. In the Foundation Phase, all schools and pre-school settings implement a Welsh-language educational programme for children in this age group” (Jones and Jones 2014: 11).
8
32
obliged public institutions to prepare appropriate statutory language schemes (cf. Williams 2010: 38; Aitchison and Carter 2000). It also created the Welsh Language Board to
“promote and facilitate the use of the language” (Parliament of the United Kingdom
1993). Following the 1997 referendum, a process of devolution began, leading to the establishment of the Welsh National Assembly in 1998. This paved the way for further
changes in legislation. In 2011, the Assembly passed the Welsh Language Measure, making Welsh the official language of Wales. By the same act, the Welsh Language Board
was replaced by the office of the Welsh Language Commissioner. The Welsh government
has continued to state their commitment to the revitalisation project by publishing national strategies, the most recent of which is entitled “Cymraeg 2050: A million Welsh speakers”. It aims to achieve two targets by 2050: a million Welsh speakers
and “the percentage of the population that speak Welsh daily, and can speak more than
just a few words of Welsh” to increase from 10% to 20% (Welsh Government 2017: 11).
The combined revitalisation efforts led to a small increase in the number of Welsh
speakers between the years 1991 and 2001, up to nearly 21% of the population of Wales
(Jones 2012: 21). Due to introducing Welsh into the national education system, the census
showed a rise in the number of young speakers (Morris 2010b: 81), which would seem to
offer good prospects for the future. Another important tendency was a reported positive
change of attitudes towards the language, both among Welsh and non-Welsh-speakers
(cf. Lyon and Ellis 1991; Davies et al. 2010: 149).
However, the results of the 2001 census proved to be “a false dawn” (Aitchinson
and Carter 2013), in that the positive trends appear to have reversed. Ten years on, the
2011 National Census once again showed a decline in the number of Welsh speakers from
20.8% to 19% within a decade (i.e. from 582, 000 to c. 562,000 speakers (Office for
National Statistics 2012). More detailed statistics suggest that the situation of the Welsh
language is, in fact, much more precarious. Only 14.6% of respondents (430,717) in the
2011 census declared being able to read, write and speak Welsh. The figure was slightly
higher in the National Welsh Language Use Survey for Wales 2013-14 (Welsh Government and Welsh Language Commissioner 2015) where the estimated number of adults
aged 16 and over who can write in Welsh was 457,000. The number of fluent speakers
was estimated by the same report at 310,000 (11%), while a 2012 Welsh Language Board
report stated that it does not exceed 300,000 (Jones 2012).
33
Another fact that causes concern is the decline in the number of habitual speakers
in traditionally stronghold areas in northern and western parts of Wales; these include
Ceredigion and Carmarthenshire, where the number of Welsh speakers in 2011 did not
exceed 50% for the first time in history. The only two counties where the number of
Welsh speakers is higher than 55% are Gwynedd and Isle of Anglesey (Gwyn Lewis
2015: 149–151). Changes in the distribution of speakers caused by the growing population mobility can also be seen in the almost 15% growth in the number of Welsh speakers
in the capital city Cardiff, which is now believed to have more Welsh speakers than the
whole of Ceredigion (Aitchinson and Carter 2013). This means that an increasing number
of Welsh speakers live in urban rather than rural settings. The census data do not, however, reflect actual language use. Studies have shown that Welsh is more likely to be used
in areas where it is spoken by more than 60% of the population and when the speakers
come from Welsh-speaking homes (Jones 2008: 552).
Other than geographical distribution, a change in the proportion between first language (L1) and second language (L2) speakers is another important issue. After introducing the language into the National Curriculum, the number of ‘new’ speakers 9 i.e.
L2 speakers who acquired Welsh through statutory education or other formal means, has
been on the rise and has contributed to the statistical increase in the total number of Welsh
speakers (Robert 2009: 94). According to the 2004 survey commissioned by the Welsh
Language Board, 73% of Welsh-speaking children learnt Welsh outside their home (Phillips 2007: 165–166). Concurrently, the same institution reported an annual net loss of
about 3,000 fluent speakers a year (Jones 2012), which means that the percentage of
learners and L2 users among the Welsh-speaking population is increasing.
The term new speakers denotes here individuals “with little or no home or community exposure to a
minority language but who instead acquire it through immersion or bilingual educational programs, revitalization projects or as adult language learners” (O’Rourke et al. 2014: 1, see also 1.2.1.).
9
34
1.1.2. Changes in the linguistic landscape
The reasons behind the ongoing decline of the Welsh language are multiple and complex.
They include civilizational changes, such as changes in the traditional family model10, as
well as the weakening importance of religion and the rural economy, areas which used to
be central for the maintenance of Welsh (Aitchison and Carter 2004: 23; Davies et al.
2010: 149). The phenomena of shrinking Welsh-speaking communities and the diminishing number of channels through which the language was previously transmitted are aggravated by high levels of out- and in- migration, particularly in the traditional Welshspeaking areas. The net in-migration from the rest of the UK has been in positive figures
since 1981; in 2001, over 20% of the inhabitants of Wales were born in England (Davies
et al. 2010: 148). Although studies have shown that in-migrants generally have a positive
attitude towards Welsh, “only about half of the adults had been prepared to take up the
challenge of learning Welsh, with relatively few succeeding at more than a basic level”
(Davies et al. 2010: 157–158). Low levels of linguistic integration of people born outside
Wales result in a process of Anglicisation (H. Jones 2010: 142).
The disintegration of traditional community ties weakens individuals’ attachment
to the language, as well as their ability and confidence. This finds its reflection in speakers’ behaviours and attitudes, such as insecurity, reluctance or lack of interest in using
Welsh. Several surveys suggest that the level of linguistic insecurity among native speakers of Welsh is very high. According to the Welsh Language Board, as much as 42% of
L1 Welsh speakers declare not feeling fully confident in using their mother tongue (Jones
2012: 9), while in the Beaufort Report commissioned by the S4C TV channel, 40% of
respondents who were fluent Welsh speakers declared that they “would like to speak better Welsh”11 (Beaufort Research 2013: 86).
Apart from the ability to speak and confidence in using Welsh, one should also
consider the actual usage of the language, which is not reflected in the figures cited above.
10
Involving, for example, growing numbers of mixed couple and single-parent households. Statistical data
suggest that these factors may influence language transmission. “Where the household contained a couple
(either married or cohabiting) and both adults could speak Welsh, 82% of children aged 3 to 4 could speak
Welsh. In the case of one-parent households, the percentage was 55%.(…) The transmission rates are also
lower in families where the couple are cohabiting, as compared with married couples, though we can perhaps explain this by their socio-economic status. Cohabiting couples tend to be younger than married couples and youth tends to be linked to lower socio-economic status” (Jones 2012: 59–60).
11
The survey measured linguistic confidence in general, though, and did not specify what “better Welsh”
meant.
35
In fact, one of the major problems as regards the future of Welsh is the reluctance of some
speakers to use the language in their everyday life (Morris 2010a: 6). This concerns, in
particular, members of the younger generation who, though able to speak the language,
often show little interest in using it when they are not obliged to do so in a classroom
environment (cf. Selleck 2016: 552). It has also been reported that young people are least
likely to speak Welsh at home or outside it, including participation in cultural, sport or
social events (Beaufort Research 2013: 87). Morris’s research on young people in twelve
different places across Wales showed that a decrease in the use of Welsh with friends and
peers both inside and outside school between subsequent school levels was observed regardless of the location (Morris 2010b: 89). Evans’s (2015) study of teenagers’ attitudes
towards Welsh and their use of it has shown that only less than 20% of the surveyed
declared using Welsh-language electronic media or reading Welsh books or magazines.
Similarly low figures are reported for regular use of Welsh on-line, although in this case
it is the young people who use this medium most (Beaufort Research 2013: 14; Welsh
Government and Welsh Language Commissioner 2015: 41–42). These data raise concern
for the future maintenance of the language. As Evans (2015) puts it: “when even fluent
speakers feel that they will not use Welsh in everyday life, this suggests that the language
infrastructure remains weak, and that Welsh remains far from a ‘living language’”. A
recent study by Owen (2018) suggests that one of the most important factors which discourages young people attending Welsh-medium schools from speaking Welsh is lack of
opportunities to use the language outside the classroom, especially in communities with
a low percentage of Welsh speakers (Owen 2018: 263). The same problem has been noted
in the Welsh Government report with regard to young people studying Welsh as Second
Language (2013: 39-40).
Taking the above issues into account, it seems that institutional domains remain
crucial for revitalisation efforts in securing the future of Welsh. Yet the functioning of
the public sector and education system does not necessarily provide speakers with sufficient motivation or confidence to speak Welsh, especially as regards L2 speakers. Although Welsh language schemes have been introduced extensively into the public sector,
full Welsh language services are not always available and there has been limited success
in promoting the language in the private sector (Williams 2010: 39). As a consequence,
36
opportunities for gaining better employment due to the knowledge of Welsh generate relatively little motivation to learn the language, particularly in the areas where English is
dominant (Robert 2009: 113).
Similar issues apply to Welsh-medium education12. Owen (2018:27-28) pays attention to the phenomenon of ‘disappearing’ young Welsh speakers emerging from census data: the figures for 5 to 15 year-old Welsh speakers in subsequent censuses constitute
the ‘peak’ across age groups with c. 40% speaking the language. However, the figures for
the 20-44 age group noted over decades remain the same, at c. 15%. This suggests that a
major percentage of people who are able to speak Welsh while attending school stop using
the language in their adult life.
Difficulties in the use of Welsh in educational environment stem from disproportions between L1 and L2 speakers in Welsh-medium schools. In fact, in many areas in
Wales the majority of children studying through the medium of Welsh come from nonWelsh speaking families (Gwyn Lewis 2015: 157). Researchers point at various results
of this imbalance. For example, Coupland and Aldridge (2009: 8) suggest that negative
attitudes towards the “new”, “school-learned” speakers, whose usage of Welsh may differ
a great deal from the “traditional” varieties, might discourage the L2 speaker and inhibit
revitalisation efforts. The existence of such negative attitudes has been confirmed by Robert (2009) whose MGT13 study investigated the perception of language competence, ethnicity, and evaluations of social attractiveness and prestige of L1 and L2 Welsh speakers.
She concludes that
[p]erceptions of social attractiveness are not dependent on identification as L1 or L2, but
on perceived language competence. The results point to a clear separation of the L2-low
group from all other three groups [L1-high, L1-mid, L2-mid – ML] in terms of social attractiveness. Members will, we can predict, encounter difficulties in integrating into social
networks and communities. (Robert 2009: 112)
12
Gwyn Lewis (2015: 157) notices that using the term Welsh-medium schools may be misleading as many
of them are in fact bilingual schools; in some cases Welsh may be available only as an option, which hardly
seems motivating for learners – for instance, students may be given bilingual examination papers and are
free to write answers in either of the two languages. There are four categories of secondary schools in Wales
according to Welsh medium provision: Welsh-medium Secondary School, Bilingual Secondary School
(with four subdivisions), Predominantly English-medium school with significant use of Welsh and Predominantly English-medium secondary school (Welsh Assembly Government 2007).
13
In a matched-guise test (MGT) participants are asked to evaluate a tape-recorded speaker according to
personal traits, such as ambition, leadership, sociability and sense of humour (O’Rourke 2011: 26).
37
Jones’s research (1998: 351) suggests, however, that the language of L1 speakers
might also be influenced by intensive contact with learners, creating a hybrid “interlanguage”, a simplified mixed variety. Baker (2010: 64) sees this type of language mixing
as a transitional stage from bilingualism in English and Welsh towards English monolingualism. Hickey (2007) notices “unofficial submersion” of minority-language speaking
children in the higher-status language when they are mixed with L2 speakers, while
Hickey et al. (2014: 230) also point out to the danger of the normalisation of code-switching and insufficient support for minority language L1 speakers in situations when teaching methods are being adjusted to the abilities of L2 speakers.
Similarly, Selleck’s studies (2012, 2013) in secondary schools in Wales have
demonstrated that students commonly used the so-called Wenglish to negotiate their position within different “hierarchies” of Welshness. “Wenglish” is here understood as
a mixture of Welsh and English, which, according to Selleck, goes beyond code-switching and can be described as heteroglossic bilingualism (2013: 30)14. The author links these
tendencies with unsuccessful implementation of government revitalisation policies,
which propagate equality of access to both languages, tolerance, openness and inclusion
with the aim of normalising bilingualism on these terms. In practice, however, in the
Welsh-medium school that she investigated the prevailing ideology was that of the strict
separation of Welsh and English within the ideology of protecting the minority language.
This, in the author’s opinion, creates language hierarchies and may lead to marginalisation and exclusion of groups. As a result, English is often seen as the universal language
of social inclusion, while Welsh might be seen as the language of exclusivity (Selleck
2012: 153-162).
The speed of Anglicisation processes varies across different regions. This was
confirmed e.g. by Morris (2010b), who established a typology of communities across
Wales based on relationships between members of language groups. She divided them
into: 1) assimilating communities, where there was pressure on L2 speakers to use the
language 2) distinctive language communities, where Welsh speakers belonged to quite
distinctive language groups and communities 3) assimilated communities, where speakers of Welsh were “rapidly becoming assimilated into the normative context where English was the predominant language” (Morris 2010b: 96–97). The last type of community
The term “Wenglish” is used also in another context, denoting the distinctive dialect of English spoken
in the South Wales Valleys (see 1.2.2.)
14
38
were those communities where the percentage of migrants was the highest. The author
does not mention the ethnic origin of the migrants; yet, it can be assumed that the growing
number of migrants from outside the UK as well as from England is going to further
increase the diversification of the linguistic landscape15. Brooks notes that due to the fact
that English is officially denoted as the common language for the whole multicultural
community, migrants to Wales lose any sense of obligation to learn Welsh (Brooks 2014).
Social and demographic changes in the Wales of today combined with the English-dominated global culture and the multitude of English-language media make the effective revitalisation of Welsh challenging and virtually impossible to maintain the language in its traditional form. Bilingual speakers of Welsh are a minority in a monolingual
English-speaking society and the intensity of English influence discussed above will inevitably bring profound changes to the Welsh language within the foreseeable future. As
stated by Durham and Morris (2016: 12) the main challenge of the present time is to
normalise the Welsh-language in Welsh society which, by the end of the twentieth century, had grown accustomed to seeing itself as bilingual. Efforts to introduce Welsh in all
domains of social life will, as a result, have to account for the concepts of standard language and a linguistic norm, which will be discussed in the following section.
1.2. Standard Welsh – linguistic norm(s)?
One of the major research questions of this thesis is how well phrasal verbs are integrated
in standard Welsh of today. This issue can be placed in a wider perspective as a case study
of changes in contemporary Welsh since acceptance of these constructions or lack of it
may to some degree reflect the prevailing ideologies connected with standardisation
among Welsh speakers. In order to address these issues it is necessary to describe the
level of standardisation of Welsh and identify major sources of the standard. This section
will discuss the historical evolution of standard varieties of Welsh as well as language
planning efforts in the twentieth and twenty-first century in the wider framework of the
standardisation of minority languages. The subject will be introduced in the following
15
However, as for 2011 National Census, 97% of inhabitants of Wales declared English or Welsh as their
main languages (Durham and Morris 2016: 6).
39
subsections: section 1.2.1 will present some theoretical concepts related to language ideologies and attitudes. Section 1.2.2. will discuss recent developments within considering
standardisation processes with focus on minority languages, while sections 1.2.3-1.2.7
will give a detailed discussion of the evolution of standard varieties of Welsh throughout
centuries and the current level of standardisation.
1.2.1. Language ideologies, attitudes and issues of standardisation
1.2.1.1. Ideologies and attitudes
As has been shown in the Introduction, in common discourse one can observe two contrasting perceptions of phrasal verbs in Welsh: as an unmarked feature of the language or
undesirable borrowings. These views are clearly linked to linguistic ideologies and attitudes of individual speakers. Language ideologies and language attitudes are two crucial
concepts used to investigate dynamic sociolinguistic processes such as language change,
which is of importance to the present dissertation. While attempting to determine the acceptability of phrasal verbs within the Welsh linguistic norm, the present study will inevitably touch on these issues. Although the two concepts are closely connected, as they
both deal with speakers’ feelings and notions about various forms of language, they differ
when it comes to their origin and methodologies applied by researchers in the two fields
(Kroskrity 2016:1). It is therefore worthwhile briefly describing the two areas in order to
highlight differences between them and relate relevant sections of the dissertation to these
concepts.
A classic definition of language ideology formulated by Silverstein states that it
comprises “sets of beliefs about language articulated by users as a rationalisation or justification of perceived language structure and use” (Silverstein 1979:173). Consequently,
research into language ideologies investigates relations between particular language situations and speakers’ expectations linked to their perceptions of various social identities.
Accordingly, Armstrong (2012: 152) provides a more detailed definition of ideology as
“a constellation of beliefs, attitudes and social norms, often concerned with the connec-
40
tion between language and identity or ethnicity, but also with reference to ontology, epistemology, history, politics, language structure and language use. (…) Language ideology
always implies a judgment of social correctness.” Studies in language ideologies are related to areas such as linguistic anthropology, interpretive sociology, and systemic functional linguistics. They rely mainly on qualitative methods such as ethnography, conversational analysis, and discourse analysis, inferring ideological stances from explanations
articulated by speakers as well as discursive practice (Kroskrity 2016). This kind of methodology will be also applied in the analytical chapters of this thesis, which make use of
text and discourse analysis.
In contrast, as the origins of the concept of language attitudes lie in social psychology, quantitative sociolinguistics, and educational linguistics, studies in this area typically focus on quantitative measurement of speakers’ reactions to language and its components (Kroskrity 2016). The notion of attitude has been present in sociolinguistics since
Labov’s classic study on social stratification of speech communities (1966) which
showed differences in language attitudes attributable to socioeconomic conditions and
social class. Derived from Sarnoff’s (1970:249) claim that an attitude is „a disposition to
respond favourably or unfavourably to a class of objects”, an attitude has been defined as
“an evaluative orientation to a social object of some sort” (Garrett 2010: 20)16. With regard to language, attitudes can be held not only towards a language as a whole but also at
all its levels, such as lexicon, grammar, accent, pronunciation and dialects (Garrett 2010:
2-15). Since attitudes are seen as dispositions, they are believed to be evaluative stances
which are sufficiently stable to be identified and in some sense measured (Garrett et al.
2003: 3).
The three main traditional approaches to researching language attitudes are: societal treatment of language varieties (content analysis of the treatment given to languages
and language varieties and to their speakers within society), direct measures (elicitation
of evaluations usually through questionnaires and interviews), and indirect measures (primarily matched-guised technique, MGT) (Garrett et al. 2003: 15-17). However, these
methods have been challenged in recent years by scholars who see attitudes as more context-bound and constructivist, arguing that they should always be studied in a discursive
16
It should be noted that in everyday usage and also in the discourse of some researchers the term “attitude”
is often synonymous with “opinion” which term lacks the affective component and is not latent but verbalised (Garrett 2010: 32).
41
context rather than measured in experiments. Such an approach gave rise to qualitative
“interaction research”, that is one that examines the emergence of attitudes in and through
interaction (for an overview see Liebscher and Dailey-O’Cain 2017). This approach uses
discursive psychology, conversational analysis, interactional sociolinguistics and theory
of motivated information management, which brings it closer to research in language ideologies. In fact, within this approach the two concepts are seen as related and interacting
(Liebscher and Dailey-O’Cain 2017:3).
This variety of research techniques in attitudes measurement stems from a number
of difficulties inherent to the field, some of which will be now addressed. Attitudes are
commonly believed to consist of three components: cognitive (thoughts and beliefs), affective (feelings, prejudices, anxiety) and behavioural (predisposing to act in a certain
way) (Garrett et al. 2003: 3). The issue of how these three are interconnected has been
widely discussed with regard to their congruity, in particular the relation between overt
and covert attitudes as well as relation between attitudes and actual behaviour (Garrett
2010: 23). The incompatibility between overt and covert attitudes has been aptly summarised by Baker (1992):
“Irrational prejudices, deep-seated anxieties and fears may occasionally be at variance with
formally stated beliefs. In attitude measurement, formal statements are made reflecting the
cognitive component of attitudes. These may only reflect surface evaluations. Doubt has to
be expressed whether deep-seated, private feelings, especially when incongruent with preferred public statements, are truly elicited in attitude measurement. Such measurement may
not always delve beneath the surface. Overtly stated attitudes may hide covert beliefs”.
(Baker 1992: 12-13)
The same point has been made in more recent studies, for example by Edwards (2010:
96), who argues that attitudinal measurements touch on peoples’ beliefs only without referring to the affective component. This twofold aspect of attitudes is a major challenge
to researchers and gave rise to the emergence of indirect measurements, such as MGT,
which hope to elicit attitudes of speakers by using a range of deceptive techniques (Garrett
et al. 2003:16).
Moreover, since attitudes are, at least to a certain degree, latent, it has been suggested that they should be inferred from external behaviour (Baker 1992: 11). In fact,
explaining behaviour or its patterns has been seen by many as one of the main purposes
of attitudinal studies with many practical implications (Garrett 2010: 25). Attitudes func-
42
tion in two ways, as input and output factors, a “cycle of influence between social cognition and language variation” (Garrett 2010: 21). In other words, they may play a role in
social processes while processes might create favourable or unfavourable language attitudes.
However, discrepancies between people’s declared attitudes and actual behaviour
have been noted for many decades, beginning with the widely-cited study by LaPiere
(1934). The same is true for attitudes regarding language; to give one example concerning
Welsh, experiments conducted by Parafita Couto et al. (2015) showed that Welsh speakers’ behaviour did not always match their acceptability judgments. On the theoretical
ground, it has been attempted to resolve this difficulty by introducing an intermediate
category of behavioural intentions (Garrett 2010: 25-26).
The validity of attitude measurements has been questioned by Baker, who argues
that “attitude measurement is rarely, if ever, totally valid” and lists three important reasons for this (Baker 1992: 18-19). Firstly, people may respond to an attitude test in a way
that makes them appear more prestigious and give socially desirable answers. Secondly,
the researcher (his or her ethnic identity, gender, status, age, language, social class), the
environment and the perceived purpose of the research may influence the speaker’s responses. The third argument is worth quoting in its entirety: “a good attitude test will
encompass the full range of issues and ideas involved in a topic. The initial item pool
must of necessity cover the fullest range of possible attitudes in terms of topic, complexity
and favourability and unfavourability, just as the item analysis on the item pool (to find
the most reliable statements and exclude unreliable items) must be executed on a representative and not atypical sample of people” (Baker 1992: 19). This implies that for the
attitude test to be valid it must be based on a sound pre-existing body of knowledge on
possible attitudes towards the investigated object. When these conditions are not met, i.e.
the object has not been well investigated, creating a good attitude test might present great
difficulties.
Concentrating on evaluative judgements, studies in language attitudes are typically linked to stereotyping in intergroup relations, which performs, among others, a social-explanatory function, creating and maintaining group ideologies (Garrett et al.
2003:3). Therefore, focus in such studies is placed on how a person's language variety is
socially meaningful to others i. e. what makes people associate the speaker with a particular social group or set of activities.
43
As can be seen, the two concepts – language ideologies and attitudes – are to some
extent interrelated: language ideologies are seen as constructs comprised of individual
language attitudes, while attitudes might be investigated as a factor in the emergence of
ideologies. In fact, one can interpret studies in ideologies as investigation of individual
attitudes from a socially derived and intellectualized, rather than psychological perspective (Schieffelin et al. 1998: 16). While attitudes focus on the individual, language ideology studies investigate competing representations of reality within the same community
and identities constructed in response to historical and social forces (Myers-Scotton
2006:110). What both areas have in common is the belief that “language is not only the
medium for expressing ideas, but also the medium for expressing evaluations across
groups and therefore for expressing power” (Myers-Scotton 2006:115). It has also been
argued by Maio et al. (2006: 284) that ideologies, attitudes and also values are three interrelated constructs operating on various levels of abstraction, and studying one in isolation from the others cannot fully explain people’s behaviours.
With its emphasis on the question of linguistic norms rather than feelings and behaviours of individual speakers, the present research on phrasal verbs does not focus on
language attitudes but is mostly concerned with standardisation-related ideologies prescribing the English element in Welsh, which certainly play a role in the representation
of these constructions, especially in metalinguistic discourse. Issues of language contact
have been an important area of language ideology studies as it has been observed that
language ideology is an essential phenomenon in language change having a potential to
influence the linguistic structure; however, the extent of this influence has been much
debated (Schieffelin et al. 1998: 12-13). Ideologies of standardisation which will be now
discussed are a focal point of such studies.
1.2.1.2. Ideologies of standardisation
Standardisation is a process which involves attempts to control the variability of language
by suppressing variation and can therefore be considered an ideology (Milroy 2005: 133).
It is closely related to two other language ideologies relevant to this thesis, namely prescriptivism and purism. Prescriptivism, as defined by Crystal (2008: 184), is an ideological approach “which attempts to lay down rules of correctness as to how language should
44
be used. Using criteria such as purity, logic, history or literary excellence, prescriptivism
aims to preserve imagined standards by insisting on norms of usage and criticizing departures from these norms.” Prescriptivism can be manifested in authoritative dictionaries
and grammars, which “identify conservative usages, linked with traditional literature and
established social values, and they discourage departure from the established norms”
(Bright 1998: 81). As will be shown, such prescriptive strategies can be also observed in
Welsh. Prescriptivism differs from standardisation in that the latter allows change and
innovation and therefore needs not necessarily be prescriptive (Walsh 2016: 9).
In the discussion of purism, I adopt here the definition of G. Thomas, who states
that purism is: “the manifestation of a desire on the part of a speech community (…) to
preserve a language from, or rid it of, putative foreign elements or other elements held to
be undesirable (…). It may be directed at all linguistic levels but primarily the lexicon”
(Thomas 1991:12). While all three ideologies of standardisation, prescriptivism and purism share the idea that there is only one correct form of the language, purism touches on
the themes of contamination, corruption, protection and preservation (Walsh 2016: 9).
According to Thomas, motivations behind purist attitudes are non-rational and motivated
by a need to perceive the world in terms of binary opposition, such as “foreign” versus
“native”, “pure” versus “corrupted” (1991: 37).
These three ideologies are firmly rooted in the notion of standard language, understood here as a prestige variety of language used within a speech community (Crystal
2008: 450). Standard forms are typically equated with a linguistic norm, i.e. “an accepted
set of rules among a group of people who may view themselves as belonging to a unified
language community” (Costa et al. 2017: 3). Standard varieties of a language function
predominantly in formal social situations, associated with fields such as politics, religion,
education and mass media (Baker and Jones 1998: 210). Standard languages evolve in
the process of standardisation, which, according to Ferguson’s definition (Ferguson
1996[1988]: 189), takes place when a variety of a language is acknowledged throughout
the speech community as a prevailing norm valued higher than regional and social dialects.
Standardisation is also a process of language planning, which, following Millar
(2005: 99-100), consists of three major elements: corpus planning, status planning and
acquisition planning. The first element, corpus planning, is the most relevant one for the
45
present research17. It involves producing new terms and identifying the ‘correct’ or ‘pure’
linguistic forms. Thus corpus planning is strongly connected with the ideologies of prescriptivism and purism. The basic difference between them is that while prescriptivism
focuses on the notion of correctness, purism refers to external or internal threat which is
believed to result in language decline. It is connected with a “complaint tradition” of publicly criticising non-standard usages, which has been observed for centuries in English
(Milroy and Milroy 1999: 24–44) and has been present also in Welsh (Ball and Müller
1992: 264). Purism can take different forms: while in minority languages it is usually
concerned with the purity of lexicon, in the case of a dominant language like English,
which generally embraces borrowings, more emphasis is placed on standards in pronunciation (Milroy 2005:105).
Introducing a recent collection of studies on prescriptivism, Percy and Ostade
(2016) note that prescriptive norms are more and more often challenged by socio-political
changes in modern societies, in particular new media and immigration, which increase
linguistic variation. The authors also observe that the growing domination of English affects traditional standards even in cultures with a strong monolingual norm, such as
France or Russia (Percy and Ostade 2016: 12).
Processes of standardisation may take a variety of forms. A commonly used division is between monocentric standards, which form a single set of universally accepted
norms and polycentric standards, where several norms, based on political, ethnic and religious allegiance, exist simultaneously (Millar 2005: 38). The range of language planning situations is also considerably wide. Standardisation may take place semi-consciously over centuries or be introduced through specific language planning strategies
over a short period of time (see Millar 2005: 114). All in all, as it is hardly possible to talk
of a universal standardisation process, the case of each language should be discussed individually (Percy and Ostade 2016: 18)
Standardisation occurs primarily in the domain of writing, since standard varieties
typically originate from the written language, which often successively contributes to the
emergence of a spoken standard. Implementing linguistic norms has been achieved by
creating canons of national literatures and normative dictionaries and grammars, which
17
Status planning attempts to elevate the status of a language variety in relation to other language varieties,
while acquisition planning is “an activity designed to encourage the greater knowledge, and (eventually)
native use, of a language variety” (Millar 2005:99-100).
46
facilitate a description of the standard variety and allow for building a comprehensive
educational system. Normative sources recognise conservative usages, associated with
traditional literary norms and established social values, and discourage departure from
them (Bright 1998: 186). Standardisation requires also the regulating and coining of specialist terminology in fields such as science and technology. Within a standardising strategy, abundant linguistic resources together with an effective education system are to ensure the use of a given language in every area of life (Baker and Jones 1998: 211). This
may be extremely challenging since even those languages which enjoy institutional support are bound to be disadvantaged by dominant languages in fields such as science (Millar 2005: 68).
As mentioned above, standard-related linguistic ideologies are tied to the question
of social power. This is expressed in the concept of legitimate language introduced by
Bourdieu (1991). He associates standard with social hierarchies, where linguistic practice
is measured against the universally recognised practices of the dominant groups enjoying
social and political power. Thus the term standard language does not refer only to the
internal characteristics of a language, i.e. its uniformity and invariance, but also the nonneutral, ideologically-based ideas of prestige and speakers’ “measure of achievement”; in
other words, prestige linked to specific language varieties is not necessarily related to the
linguistic forms used, but the social position of speakers (Milroy 2001: 531–532). This is
so because standard language is invariably connected with the notion of ‘correctness’
which can be mastered only through the process of formal education. According to Bourdieu,
'[c]orrect usage' is the product of a competence which is an incorporated grammar, the
word grammar being used explicitly (and not tacitly, as it is by the linguists) in its true
sense of a system of scholarly rules, derived ex post facto from expressed discourse and set
up as imperative norms for discourse yet to be expressed. (1991:61)
Consequently, the standard variety may be seen as an abstract stable and uniform
idea, which can never be perfectly and consistently realised in speech. In this sense the
ideology of the standard is related to the question of language ownership, based on an
assumption that language is not the possession of the native speakers, but those who have
been formally taught the canonical, externally defined forms (Milroy 537-543). Cooper
(1989: 88) divides groups responsible for implementing linguistic norms into formal
elites and influentials. Formal elites are people officially empowered to make policies,
47
while influentials enjoy authority without actual statutory powers. The latter include for
example scholars, journalists and members of middle class, such as teachers, clergy and
officials of corporate bodies. The standard is therefore strongly connected with social
class, as it is usually the middle-class elites who are responsible for its creation and
maintenance, although members of all social classes may contribute to language planning
indirectly by introducing change in the language (Millar 2005: 97–98).
On the level of language structure, Langer and Nesse (2012) note that standardisation necessarily involves exclusion of some linguistic elements: “codification cannot
really take place without stigmatizing those words and constructions which are not to
become part of the standard language” (Langer and Nesse 2012: 613). As shown in the
Introduction, phrasal verbs in Welsh might be sometimes placed outside the legitimate
language because of such ideological views. An overview of ideological stances towards
borrowings from English will be presented in 1.3.3.
1.2.2. Standard and minority languages
The strength of standards in the Western Europe of today is historically linked to the
ideologies of the modern state. Standards emerged in modern Europe as tools of integration and for the consolidation of nation states, where the drive towards uniformity was to
ensure greater efficiency in achieving social and economic goals: “progressive standardisation of monetary systems, weights, and measures and of factory-made goods generally,
has gone hand in hand with the rise of international trade and capitalism, and progressive
standardisation of language has developed alongside standardisation of these things”
(Milroy 2001: 534) ). The ideology of standardisation had a philosophical as well as a
socioeconomic dimension. The age of Enlightenment gave rise to the concept of language
as a channel of transmission for human reason. However, this role was given to state
languages only, while others, including minority languages, “were relegated to a position
outside of reason, as patois or of expressions of unreasonable and emotive thought” (William 1994: 81). In brief, the ideology of a monoglot standard has been based on two
founding concepts: unification and rationalisation, viewed as both the function of the language and the ultimate goal (Urla et al. 2017: 29).
48
Whilst widely used languages such as English or French have developed “standard
language cultures” (Milroy 2001:530), processes of standardisation are present also in
minority languages, constituting an important part of language planning, which is to ensure these languages’ survival. In the European context, the introduction of European
Charter for Regional or Minority Languages by the European Union has provided more
space for standardisation by recommending state support for language policy and planning initiatives for minority languages and thus raising their status (Percy and Ostade
2016: 18; Costa et al. 2017: 16).
Standardisation has a number of potential advantages for minority languages.
Firstly, being considered the most ‘correct’ and prestigious varieties associated with
a high social position, standards might elevate the language’s status. Secondly, in the
context of language planning and maintenance, they increase social cohesion by facilitating communication between speakers from different regions and between public institutions and the public (Lane 2014: 265). Last but not least, in the political dimension, they
may constitute a unifying factor for a minority community by emphasising its distinctiveness and thus constructing national identities (Sayers 2009: 207).
Yet although standardisation may strengthen the status of a minority language,
there is a number of difficulties inherent in the process. A well-established standard variety enjoys a degree of authority as long as it perceived by the speakers as authentic, i.e.
representative of their speech, and universal i.e. socially neutral and belonging to all citizens (Gal 2006: 166). Both criteria, however, give rise to some issues. The concept of
universality appears to be questionable in the case of a minority language, which by definition exists in the shadow of a “universal”, dominant language (Gal 2006: 171). Issues
of authenticity are bound to emerge in the situation of considerable dialectal variation
when a standard variety is rejected by the speakers. Such problems have been observed
in a number of minority languages in Europe, for instance in Galician, where “despite
over 30 years of institutional standardisation, half of all Galicians still see galego normativo as artificial” (O’Rourke 2017: 90), or in Scots, whose standard variety has been contested even by language advocates (Costa 2017: 52). Such phenomena often stem from
lack of central authority or power elite which would legitimise the standard language and
give it authenticity. Such seems to be the case of, for example, standard Breton, created
as something of a compromise between the four main dialects of the language (Baker and
Jones 1998: 214).
49
Standardisation might be also perceived as detrimental for linguistic diversity expressed in dialectal variation in that it results in dialect loss, disappearance or mixing.
Instead of strengthening the self-worth of speakers of dialectal varieties, it may lead to
stigmatisation of both the linguistic forms and the speakers (O’Rourke 2017: 89). For
example, it has been suggested that the weakening vitality of dialects in Irish might partially stem from lack of recognition in the public space dominated by the prescribed standard variety (Ó hIfearnáin and Ó Murchadha 2011: 100). Loss of dialectal diversity in
Welsh has been discussed by Jones (1998) and Sayers (2009: 311–322).
Other issues of standardisation emerge in the context of bilingualism, which is
especially relevant in the case of Wales. As mentioned above, the notion of a standard
variety of language, dating back to the Enlightenment period, is based on the idea of
a national state unified by a common tongue and territory. This concept gave rise to monolingual norms according to which if two or more languages exist next to each other they
should be kept apart by the speakers and each of them should remain ‘pure’, without
interference from another (Gal 2006: 170).
The monolingual norm has been challenged in recent years by a number of poststructuralist approaches which are used to explain current linguistic practices among marginal populations, and place practice rather than standards at the centre of research. “Linguistic and cultural practices are no longer examined against the background of abstract
standard languages, uniform views of speakers and stable group identities. Rather, such
practices are investigated with reference to the fragmented repertoires that people acquire,
construct and mobilise” (Pérez-Milans 2015: 88). Researchers have investigated new profiles of speakers, attempting to replace the traditional labels of “learners” or “L2 speakers” with the term “new speakers”, especially relevant to minority languages (O’Rourke
et al. 2014). As part of the critique of traditional ideologies of multilingualism, the poststructuralist approaches have used a variety of terms, such as, “heteroglossia”, “polylingualism” and “translanguaging”. The concept of heteroglossia derived from Bakhtin
(1981) views languages as a social and political mediums rather than distinct linguistic
systems, deconstructing the idea of a unitary language. Polylingualism (Jørgensen 2008)
differs from multilingualism in that it sees languages as constituents of a given speaker’s
language, rather than separate entities. Translanguaging (García 2009), in turn, involves
concurrent uses of two languages in a bilingual classroom, embracing the use of codeswitching and translation. It has been proposed as a method of retaining and developing
50
bilingualism in Wales (Lewis et al. 2012). Also in the context of bilingual education,
Blackledge and Creese (2010) have proposed a division between two alternative ideologies of bilingualism: “separate bilingualism”, which propagates the separation of language mentioned above and “flexible bilingualism”, where the school encourages fluid
movement from one language to another for effective communication.
Notwithstanding the potential of poststructuralist ideological frameworks, it
seems that more conservative views involving notions of distinct national languages, native speakers and mother tongue are still prevalent, underpinning an image of monolingualism in heteroglossic societies (O’Rourke 2017: 328). A popular idea worth mentioning is “additive bilingualism”, which views bilingualism as a positive phenomenon, yet
holds that a bilingual person should have an equal command of two languages and be able
to clearly separate one from another (Landry and Allard 1991: 198–199). Consequently,
when bilingualism is seen as “double monolingualism in two distinct, standardized national languages” (Flores and Schissel 2014: 457), linguistic phenomena such as extensive borrowing or code-switching are seen as nonstandard and unwelcome. This notion
might be challenging both for native and new speakers living in a fully bilingual environment, as is often the case in Wales (cf. Selleck’s 2013 study on bilingual education in
Wales).
In view of the above, if the standard is too much divorced from the varieties used
by the speakers of a minority language on everyday basis, imposing restrictive norms may
have negative results on the users in that they start to depreciate their own linguistic ability (Gal 2006: 165; Baker and Jones 1998: 213). On the other hand, as pointed above,
linguistic norm and the existence of a prestigious variety of a language can be a source of
pride and strong cultural and political self-identity, ensuring a connection with tradition
and the heritage of the past. A standardised language able to describe the latest developments in technology or science is likely to enjoy high prestige and be associated with
modern life. Neglecting standards may create a notion that the minority language is
a thing of the past, “outside of the realms of reason” (Williams 1994: 82), and lead to its
further decline at the cost of the dominant language. For that reason the effectiveness of
language revitalization may significantly depend on widespread literacy and proficiency
in the prestigious variety.
51
This overview points out to some of the issues connected with standardisation of
minority languages and illustrates the many facets of the problem. In Wales, with its varied linguistic landscape, the situation is no less complex. First and foremost, in contrast
to other Celtic languages such as Irish or Breton, an official Standard Welsh has not been
created neither for educational nor official purposes. Standard varieties of Welsh have
evolved throughout centuries, their most important source being the language of literature
stemming from the practices of the elites supporting a traditional culture and bearing relatively little relationship to speech. However, the last hundred years of sociolinguistic
changes in Wales have brought an immense change also in the perception of the linguistic
norm. In these circumstances, one can hardly talk about a single standard, but a number
of high (H) varieties, or registers, of the language which are seen as standard ones.
1.2.3. Registers of Welsh – typologies
Following Crystal’s definition (2008: 409), the term register will be used here to denote,
varieties of language used according to their use in social situations, as opposed to those
defined according to the user, for instance dialects. Register, therefore, will be viewed
here as a functional and contextual category reflecting the social order in the variety of
social processes (Malmkjaer 2004: 169). With the multitude of situations and domains,
registers “comprise an open-ended set of varieties (or styles) of language typical of occupational fields, such as the language of religion, the language of legal documents, the
language of newspaper reporting, medical language, technical language, etc.” (Trosborg
1997: 5). In consequence, setting clear-cut boundaries between various registers of a language is hardly possible.
It should be emphasised at this point that the present discussion of standard varieties of Welsh will focus on the contextual dimension of registers rather than possible
geographical variance. The reason for that is that Welsh H registers are comparatively
universal in that they do not seem to favour a particular dialect. For example, considering
the lexicon, they are more inclined towards the Northern dialects, e.g. preferring gan over
gyda (for ‘with’ used to express ‘to have’), but in some cases prefer the southern variants
e.g. gyda over efo, (for the preposition ‘with’). Unsurprisingly, the lower one moves along
the register continuum, the more pronounced dialectal differences become.
52
The divisions between registers in Welsh are mostly based on the formality criterion, which might intuitively seem clear-cut due to the diglossia18 between the formal,
literary language and its spoken varieties, which is a characteristic feature of modern
Welsh (cf. e.g. Fife 1986). This substantial variation emerges in nearly all aspects of the
language, including phonology, morphology, syntactic patterns, lexicon and discourse
(for detailed description of the differences see e.g. National Language Unit 1978;
D. G. Jones 1988; Morris Jones 1993). Morphology of the verb is where the greatest discrepancy occurs. For example, certain conjugated forms of the verb, such as the present/future 3rd person irregular forms tense are confined to H varieties. Adjective inflections are heavily reduced in the informal language and the equative degree is expressed
in a different way. There is also diversity in the rules governing the use of particles and
pronouns. As regards syntax, periphrastic forms of the verb are preferred in informal varieties. In terms of lexicon, a characteristic feature of the more formal varieties is the
tendency to avoid Anglicisms (see 1.3.3.). Some researchers believe that the differences
span also the non-borrowed lexicon: “anyone who tries to read Welsh poetry from
a grounding in colloquial forms is in for a nasty shock, much more so (on a purely intuitive basis) than with the comparable transition in, say, French or Russian” (Fife 1986:
146).
As a result of this diversity, in today’s popular discourse a division is often made
between ‘literary’ and ‘colloquial’ Welsh, the first being associated with written and formal, the other with the spoken and informal language. This differentiation, supported by
the popular grammar by King (1st ed. 1993), who claims that “Literary Welsh” and “Colloquial Welsh” are so different that they might be called two distinct languages, found its
way into the English Wikipedia in the form of two separate articles on Welsh morphology
("Colloquial Welsh morphology" 2018; "Literary Welsh morphology" 2018).
Many researchers, however, view this as an oversimplification and exaggeration.
It is a general view in modern linguistics that diglossia is a continuum, “a gradient cline,
with one variant shading into another” (Schiffman 1998: 210). Roberts and Jones (1974)
Following Ferguson’s classic definition, diglossia is understood here as “a relatively stable language
situation in which, in addition to the primary dialects of the language (which may include a standard or
regional standards), there is a very divergent, highly codified (often grammatically more complex) superposed variety, the vehicle of a large and respected body of written literature, either of an earlier period or
in another speech community, which is learned largely by formal education and is used for most written
and formal spoken purposes but is not used by any sector of the community for ordinary conversation”
(Ferguson 1972[1959]: 245).
18
53
proposed three main “levels” of Welsh: literary, standard speech and local dialect, each
with two “modes of expression” (written vs. spoken) and each with a variety of “registers”, while Jones and Thomas (1977) classified varieties of Welsh into formal written,
formal spoken and informal/spontaneous spoken (after D. G. Jones 1988: 137). Morris
Jones (1993: 2–4) distinguishes between written and spoken Welsh in two dimensions:
channel, i.e. the physical representation of language (written/spoken), and medium i.e.
grammatical, phonetic and discursive features characteristic for written/spoken language.
In this model the registers are placed along the formal-informal continuum. Lewis (1995)
differentiates between written formal, written informal and spoken vernacular. Focusing
on the written language, he identifies the "formal, traditional" as characteristic of poetry,
legal language, and religious writing, while "informal, contemporary" Welsh is expected
to be used in the press and popular works of fiction and on everyday basis by an educated
Welshman (Lewis 1995: 10)19. Jones and Chapman (2000: xi–xii) draw attention to the
multitude of registers in Welsh, dividing them into the following:
a. literary Welsh to be found in the Bible, and "most forms of highbrow literature";
b. official Welsh identified with Lewis's "contemporary written informal";
c. colloquial Welsh used mostly in informal speech, but also accepted in formal
spoken situations and used in fiction;
d. bratiaith – a "debased Welsh" very heavily influenced by English, rarely to be
found in writing (see Błąd! Nie można odnaleźć źródła odwołania.);
e. Wenglish, which can be seen as Welsh "spoken through the medium of English" or a variety of English heavily influenced by Welsh in the industrial valleys of South Wales20.
The models are summarised in Table 1.
19
This division is by necessity simplified to serve the practical purpose of Lewis's book.
This dialect combines the intonation and accent of South West Welsh (Gwenhwyseg dialect) with the
speech rhythms of spoken English. It is considered a variety of English with a high number of lexical and
structural borrowings from Welsh and is estimated to be spoken by over a million people (Lewis 2016: 11–
34).
20
54
Table 1. Varieties of Welsh - typologies
H variety
L variety
Robert
and Jones
1974
Literary
Standard
speech
Local dialect
Jones and
Thomas
1977
Formal
written
Formal
spoken
Informal/
spontaneous speech
Morris
Jones 1993
King 1993
Lewis 1995
Literary
Written formal
Written/spoken informal
Vernacular
Formal written/spoken
Colloquial
Informal
written/spoken
Jones and
Chapman
2000
Literary
Official
Colloquial
Bratiaith
(Wenglish)
The key divisions to be seen in these typologies are between written and spoken
mode and along the formality continuum. While written formal/literary language remains
the H variety in all models, a change can be observed in the middle of the register continuum: the semi-formal register, which used to be mostly associated with spoken language,
is now regarded as the official variety used also in writing. This points to the growing
discrepancy between traditional and contemporary semi-formal language, as the latter
contains more features of informal language than before.
In an attempt to explain the tendency to eliminate traditional diglossia, the next
sections will look in more detail at the H varieties of Welsh and seek to establish the
current functioning of the linguistic norm in Wales by addressing the question which varieties of the written and spoken language are considered standard and to what degree
they are accepted and universal.
1.2.4. Written standard
The discussion of written standards will focus on two varieties of Welsh described most
often as H varieties (Table 1): literary and official Welsh. Out of the two, literary Welsh
has a much longer tradition behind it, dating back to the earliest written accounts of the
Welsh language (D. G. Jones 1988: 126). Legal Welsh, or as B.P. Jones (1988: 173) calls
it, proto-official Welsh, emerged as early as the 10th century with the laws of Hywel Dda.
However, with the Act of Union, Welsh lost its position as the language of politics and
law for hundreds of years. Thus, it was the language of literature, and not of the court,
that became the major standard in the centuries to come.
55
1.2.4.1. Literary Welsh
The earliest written accounts of literary Welsh, dated to the 6th century, are the works of
the so-called Y Cynfeirdd (“The Early Poets”), whose poetry is believed to have had important political, magical and ritual functions. By the 12th and 13th centuries poets established themselves as a class of professionals, who played a vital role in the social system
of medieval Wales, ensuring the survival of kingdoms by composing eulogies and elegies
for the ruler, legitimising his power. This generation of poets, known as Beirdd y Tywysogion (“Poets of the Princes” also called Gogynfeirdd lit. “not so early poets”) performed
their poetry accompanied by music. It should be borne in mind, therefore, that literary
Welsh was originally a spoken, not a written variety. At the same time, the preserved
manuscripts, such as the four branches of Mabinogi, provide some evidence of what informal spoken language of that time might have been like and point to the fact that next
to the elevated, formal language of the court poets, there functioned a more “relaxed”
standard, confined to prose (D. G. Jones 1988: 126).
The language of the poets underwent extensive codification in the 15th century.
The new generation of bards, known as Beirdd yr Uchelwyr (“Bards of the Gentry”) continued to create intricate verse, following the complex rules of poetic art, cerdd dafod.
Professional poets were obliged to study the grammatical rules of the language, yet at the
same time, they experimented with the lexicon and were not reluctant to borrow extensively from other languages (see 1.3.3.) Although from the 16th century onward the poetic
tradition began to decline, the language of the poets became the basis for the literary
language used until today. One of the major reasons was that the bardic diction served as
a model for the 1588 translation of the Bible by Bishop William Morgan (1545-1604)
commissioned by Elizabeth I. Morgan’s Bible was a major landmark in the history
of Welsh, partly regaining its prestige as the language of the church. The revised edition
of the Welsh Bible by Bishop Richard Parry (1560-1623) and Dr John Davies of Mallwyd
(c.1967-1644) was published in 1620 and its influence on the standard language extended
well into the late 20th century (Hunter 2006: 210; Robert 2011: 135). As Lewis (1987: 13)
puts it: “the authority of the text – a sacred book – supported the authority of the ‘standard’
language into which it was translated”.
With the advent of the Renaissance and the New Learning, attempts were made
by humanists to adapt the Welsh language to the new intellectual climate. It was the time
56
of creating first dictionaries and first attempts to standardise the orthography
(D. G. Jones 1988: 129). Another key point in history was the emergence of Protestant
denominations: literary language became known to the common people not only through
the Bible but also abundant religious prose, which followed somewhat less rigid linguistic
norm. This trend increased in the two centuries to follow with gradual spread of literacy
due
to
emerging
Welsh-medium
circulating
schools
and
Sunday
schools
(Hunter 2006: 210). In the 19th and in the first half of the 20th century the chapel was
closely linked to the Welsh identity and Welsh language. As well as catering for local
Welsh-speaking communities and providing basic education, chapels supported the revival of the idea of poetry competitions, eisteddfods, on the local and national level, which
brought back focus on poetry in Wales after the decline of the bardic tradition (Cohen
2007: 92–97).
The late 18th century was the beginning of a period which has been called the
times of “xenophobic purism” (Hincks 2000: 17) or “strange and sad corruption” (D. G.
Jones 1988: 130), when a number of major writers decided to ‘elevate’ Welsh by introducing artificial changes to the language, for instance by arranging words to imitate the
patterns of English. This made the idiom extremely unnatural and pretentious. The most
influential figure of that time was William Owen Pughe (1759-1835), author of a dictionary and grammar notorious for its unnatural and strange syntax, orthography and coinages. His influence extended over the following decades (Heinz 2003: 170).
By the end of the 19th century there emerged a new generation of scholars willing
to purge Welsh of its peculiar mannerisms. The two most prominent figures here were
Emrys ap Iwan (1848-1906), author of essays, who based his model of good Welsh on
prose writers of the Reformation and Methodist revival period, and John Morris-Jones
(1864-1929), an Oxford scholar and poet who advocated the authority of the poetic tradition (D. G. Jones 1988: 132). They were concerned with the relationship between the
written and the spoken language and encouraged narrowing the gap between the two.
Morris-Jones believed, however, that literary Welsh is something to be learned and mastered and that therefore absolute unification was impossible:
The value of the tradition is that it represents the language in a form which was everywhere
recognised as pure, and of which various dialects represent different corruptions. (MorrisJones 1921: v)
57
‘Pure’ language was thus understood as the language of educated people devoid
of dialectal elements. In consequence, in Morris-Jones’s view, the main criterion considered as a benchmark for standard Welsh was the correctness and naturalness perceived
by an educated Welsh speaker.
The great authority of Morris-Jones lasted for many decades. He was successful
in establishing a new standard nicknamed “Oxford Welsh” (D. G. Jones 1988: 134–135),
much closer to the spoken language than the elevated Victorian idiom. Morris-Jones also
played a pivotal role in standardising Welsh spelling. He created the Dafydd ap Gwilym
Society which helped develop a new consistent orthography, leading to the publication of
Orgraff yr Iaith Gymraeg [Orthography of the Welsh language] in 1928 by the University
of Wales. This is the standard spelling valid to this day.
As can be seen from this historical overview, by the early 20th century the Welsh
language had a developed and distinctive H variety maintained through three main channels: the church, traditional literature and high educational institutions, which ensured its
stability. The last hundred years saw deep-going changes in that respect.
To begin with the religious domain, the 1930s marked the beginning of a gradual
decline in church adherence and attendance in Wales, which accelerated dramatically in
the 1960s and has continued until today (Brown 2014: 26–27). According to Morgan
(2000), the decline of Welsh language and Christianity were closely intertwined:
[u]ntil the mid-20th century and after that in the areas where Welsh continued to be the
main language, religion was the only field where Welsh had a strong position. English was
the medium of everything else – government, trade, education (except for primary schools),
public media such as radio, television, the cinema and the major newspapers, and the modern Anglo-American culture. All of this was absolutely devastating for both religion and
Welsh. Any dualism between the holy and secular is injurious to religion as it restricts the
sovereignty of God (…). In Welsh-language areas there was a tendency to narrow down
"religion" to specific religious practices, such as going to chapel, worshipping, praying,
holding Sunday schools and practicing some aspects on individual morality. Just as religion
could not predominate secular matters, be it politics, economy, aesthetics or social ethics,
neither could Welsh embrace life in its entirety since English was the language of the
"world". (Morgan 2000: 378)
The 2011 census has shown a 14.3 percentage point decline in the number of
Christians since 2001 and a parallel growth in the number of citizens stating no religion.
Moreover, according to a 2007 survey, the level of regular church attendance in Wales is
12%, the lowest in the UK (Williamson 2014). These numbers indicate that the number
of Welsh people who have access to biblical language is low and constantly decreasing.
58
What is more, the classic Morgan’s Bible ceased to be the default version of the Scriptures
when a new translation, Beibl Cymraeg Newydd [New Welsh Bible] was published in
1988 (Hunter 2006: 210). Its grammar was modernised and the language simplified to be
more comprehensible. Beibl Cymraeg Newydd is the version currently used in services
but it is questionable whether it has ever reached the same prestigious position as Morgan’s Bible. For example, the authors of Geiriadur yr Academi (1995) chose to quote the
traditional translation when including biblical idioms in their dictionary (D. G. Jones
1988: 152).
It seems, however, that even the more comprehensible language of Beibl Cymraeg
Newydd may be inaccessible to the younger generation of speakers. A symptomatic illustration of this is the launch of yet another translation of the Bible in 2015, beibl.net, aimed
at “young people, learners and those who wish to have a more straightforward explanation
of the original text” (Pritchard 2015). The new translation is written in spoken, colloquial
Welsh with simplified morphology and a reduced lexicon. All in all, the importance of
the Bible as a source of the standard language appears to be diminishing alongside the
decline of traditional Christianity in Wales.
The language of literature also underwent deep changes in the 20th century. One
of the major reasons for that was the advent of the modern novel and short story. These
forms emerged comparatively late in Wales, at the turn of 19th and 20th centuries
(Hunter 2006; Tomos 2006). Possible reasons for that are the hostility of Nonconformist
churches towards fiction and the lack of a rich urban middle-class in Wales, who at that
time would be the main target audience of these genres (Morgan 1981: 239; Snell
1998: 38). The Welsh novel was thus not associated with the elites; Welsh authors who
aimed at a realistic portrayal of their communities did not wish to preserve the classical
elevated standard. Welsh translations of English-language fiction might also have played
a role in that their authors attempted to convey different registers of English through the
use of Welsh vernacular (Willis 2016). As a consequence, in the course of time, dialect
and colloquial spoken language began to be increasingly incorporated into prose. The
poetic idiom has also moved much closer to the spoken language, although it still contains
numerous forms characteristic of the high register (cf. D. G. Jones 1988: 152).
An institution that should be mentioned at this point is the National Eisteddfod,
a Welsh-language national festival of culture held annually since 1861. The key elements
of the Eisteddfod are literature competitions, most notably competition for the Chair for
59
strict-metre verse, the Crown for free-metre in poetry and the Daniel Owen Memorial
Prize (awarded since 1978) for the best novel (see 3.2.1). The importance of the Eisteddfod as a platform of maintaining literary standards and traditions is partly due to the influence of John Morris-Jones, who, a poet himself, advocated his ideas about the language
as an Eisteddfod adjudicator (Hunter 2006: 1776). To this day the linguistic element plays
an important role at the Eisteddfod in adjudicating both poetry and prose. Although there
are no official standards for literary competitions, judges frequently pay attention to authors’ excellence in the language and to the ‘purity’ of their Welsh (see 1.3.3.).
Despite the unbroken literary tradition, contemporary Welsh fiction has fully embraced a wide variety of registers in Welsh, including the least prestigious varieties such
as bratiaith to accurately reflect the linguistic diversity in Wales. Writers do not seem to
feel constrained by a necessity to use traditional literary Welsh, although many of them
draw from its rich resources. For example, Jones and Chapman notice that the language
of contemporary fiction might be more conservative than the language of the new Bible
in its use of idioms (2000: 53).
Another channel of transmission of the literary language are universities. Since
the time of John Morris-Jones university scholars have been responsible for linguistic
publications describing and teaching literary Welsh, such as Cyflwyno’r iaith lenyddol
[Introducing the literary language] (National Language Unit 1978). Special courses designed to help students recognise forms of the literary language and acknowledge differences between literary and colloquial Welsh have also been established. As for 20162018, such courses are offered by all major university in Wales e.g. the module Cymraeg
Llyfr a Chymraeg Llafar [Literary Welsh and Spoken Welsh] in Aberystwyth, Cyflwyniad
i’r Gymraeg [Introduction to Welsh] and Iaith ac Ystyr [Language and Meaning] in Cardiff, Defnyddio Cymraeg [Using Welsh] and Ymarfer Ysgrifennu [Writing Practice] in
Bangor and Ymarfer Ysgrifennu: Uwch [Advanced Writing Practice] at University of
Wales Trinity Saint David.
The existence of these courses suggests that compulsory education provides students of Welsh with relatively little access to forms of literary Welsh. For example, according to the current school curriculum for the GCE AS and A level in Welsh as L1,
students are obliged to read less than five longer works of fiction in the course of their
schooling and the majority of the books on the compulsory reading lists are contemporary
60
novels written in the spoken, dialectal idiom, such as Un nos ola leuad by Caradog Prichard or Martha, Siac a Sianco by Caryl Lewis (Welsh Government 2015).
As can be seen, literary Welsh continues to serve as a standard variety in certain
contexts, such as religious writing and high-brow literature, in particular poetry associated with the National Eisteddfod. It is also part of the higher education system, taught to
students of Welsh and a number of other courses in humanities. This means, however,
that access to literary Welsh is confined to a relatively narrow groups of individuals. The
affiliation of literary Welsh with high-status institutions might have a twofold effect of
making it a prestigious variety, but also an elitist and hermetic one.
In 1988, D.G. Jones wrote that with the death of last monolingual speakers, literary Welsh can serve as the only standard for correctness.
As recently as 60 years ago it was possible to appeal to the authority of the spoken language
in order to eradicate some of the corruption which affected the literary medium. Because
of the pervading influence of English this is no longer possible. The literary language has
to serve as the only standard for correctness. (D. G. Jones 1988: 135)
Now, 30 years later, the influence of English seems ever so strong, while the exposure to
the monolingual norm embedded in the traditional literary language appears to be decreasing. It seems that a significant number of Welsh speakers may be little familiar with
literary Welsh in its traditional form as two of the main channels of transmission of the
literary diction, namely religious writing and fiction have undergone dramatic changes.
Question remains whether the weight of tradition and the prestige of academia or literary
circles will be enough to maintain this register as the H variety or whether conservative
grammatical forms will gradually disappear, giving way to less formal varieties.
Indeed, another crucial change that took place in the last 30 years is the strengthening of the official status of Welsh and developments in semi-formal varieties of the
language.
1.2.4.2. Official Welsh and semi-formal varieties
Semi-formal varieties of Welsh encompass a wide range of registers and thus evade simple description. In the past, the semi-formal varieties have been paid little attention by
61
linguists, who focused on the presumably clear-cut distinction between literary and colloquial Welsh, with traditional grammar books focusing on the H variety. At present,
a third, in-between register of Welsh is acknowledged by most researchers. One of the
most recent typologies by Jones and Chapman (2000) distinguishes a variety called Official Welsh (OW) situated between Literary (LW) and Colloquial Welsh (CW) The term
“official” may be somewhat misleading as the authors identify it with Lewis’s “contemporary written informal” (Lewis 1995) used not only in strictly administrative contexts
but also cultural domains such as literature and press. In such an understanding the term
official does not refer only to the political status of a language but denotes “a legally
appropriate language for all politically and culturally representative purposes on a nationwide basis” (Stewart 1968: 540).
From a linguistic point of view, Official Welsh is similar to Literary Welsh in
most aspects, the main difference being that Official Welsh avoids a number of more
conservative grammatical forms. This is visible mostly in the morphology of the verb and
the use of pronouns and particles (C. Jones 2013: 179–212).
The emergence of Official Welsh coincides historically with Welsh regaining
a certain level of political status and prestige (see 1.1.1). After Wales lost its independence, the development of medieval official Welsh was hindered by the gradual dominance
of English in law and administration. The culmination of this process were the so-called
Acts of Union of 1536 and 1542 when Wales was formally annexed into England and
English was imposed as the only official language necessary for social advancement. The
prestige of Welsh was further lowered after the industrial revolution, and in 1870 the
Primary Education Act banned Welsh from the compulsory education system, pushing
the language to the domains of the chapel and the home while English remained the language of commerce and law (Loffler 2008: 352). Thus “two standard languages were used
in Welsh-speaking Wales, endogenous (standard Welsh, used in literature, religion, etc.),
and exogenous (standard English, used in public administration)” (Robert 2013: 94).
The Welsh Language Act of 1967 was the first major step in re-introducing Welsh
as the official language of Wales and a need emerged to create a register to be used in
legal documents. To that purpose a panel of translators within the newly-established
Welsh Office was appointed to translate all the necessary documents from English. It was
decided at that time that forms of the language used would be based on standard literary
62
Welsh as the default H variety (B. P. Jones 1988: 174). Thomas (1982: 92) provides examples of extremely conservative grammatical forms used in bank forms and other official documents at that time.
According to B. P. Jones (1988) in the major article written thus far on official
Welsh, the new variety generated problems from the beginning of its existence. Welsh
speakers had difficulties in acknowledging the official status of Welsh, since Englishlanguage documents seemed more authoritative on the one hand, and more authentic and
natural than their Welsh translations on the other. What is more, many people found the
classical variety of Welsh too demanding. Difficulties stemmed from the use of traditional
literary forms and the fact that the texts were an outcome of translating not only English
words, but also concepts characteristic of English legislation. All in all the language
sounded complicated, unnatural and intimidating to some speakers:
The classical nature of the Welsh used formed another barrier: it was ''proper" Welsh, "perfect" Welsh, even "textbook" Welsh, the kind of "pulpit" Welsh many people could understand but were not proficient in its use. The complaint "my Welsh isn't good enough" echoed throughout the land. (B. P. Jones 1988: 174)
Although little research has been conducted on the perception of official Welsh in
the 21st century, there is ample evidence to suggest that problems with comprehensibility
continue. Robert (2013: 1) claims that the convention of using Welsh in public administration is not very strong and quotes examples of incomprehensible documents.
Gruffudd (2015) describes difficulties similar to those listed by B. P. Jones (1988), claiming that the language found in official documents is too complicated, loaded with difficult
terms and grammatically considerably different from the spoken varieties. He also draws
attention to the fact that official texts are translated from English and the quality of translations is insufficient in some cases, since legal phrases acceptable in English might sound
highly unnatural and thus unacceptable in Welsh. However, what exactly is natural and
acceptable might be difficult to decide as there are many fields that are discussed in Welsh
only through the medium of translation from English (Gruffudd 2015: 118). For that reason the author states that few people choose to fill official documents in Welsh and turn
to English, being aware that the original text is more likely to be correct and comprehensible than a translation. Gruffudd’s largely intuitive observations were confirmed in
a small qualitative study (Bazoli 2015) on the use of Welsh in official service. The Welsh
speakers surveyed complained of the fact that documents are translated too literally from
63
English, producing incomprehensible and unnatural texts. Of the Welsh speakers surveyed in Beaufort’s report only 31% admit ever filling a form in the Welsh language
(Beaufort Research 2013: 34). Evas and Cunliffe, investigating the use of Welsh e-services, provide evidence on considerably low use of Welsh-language services (Evas and
Cunliffe 2016: 63-64).
In light of these problems, tendencies to change the standard official Welsh in a
way which would facilitate understanding have been observed in the last 20 years. A notable initiative in that respect is a guidebook Cymraeg Clir [“Clear Welsh”] by Cen Williams (1999) which aims at helping users to adopt an appropriate register to their writing,
so that the text is understandable, “keeps the honour of the language by maintaining
Welsh forms” and avoids Anglicisms (Williams 1999: 22). Courses in Cymraeg Clir are
offered in some workplaces in Wales (Bryer 2016: 61), and it is described as the Welsh
counterpart of plain English in the national standards for adult literacy and numeracy
(Basic Skills Agency 2003). However, Cymraeg Clir has also met with some criticism
for propagating a simplified model of Welsh (e.g. D. G. Jones 2013; Asmus and Williams
2014: 51). In a handbook for Welsh translators (Prys and Trefor eds. 2015), Jones argues
that the model requires further development since due to the “unfortunate” use of words
such as ”simple”, the guidelines may indeed be associated with simplifying the language,
rather than making it sound natural (Jones 2015: 102). Incidentally, the book in question
touches on many similar issues concerning the use of registers in Welsh and illustrates
the ongoing discussion within the translators’ community on establishing the right balance between maintaining the traditional formal register and ensuring comprehension in
the current bilingual society.
At present, the Welsh authorities seem to acknowledge the need to differentiate
between different registers depending on the need and receiver. In 2006 the Welsh Government published Yr Arddulliadur, translation style guides for translators of Welsh and
English. The guidelines are aimed to help achieve consistency in texts produced for the
government, based on formality and text type. In the section “cywair” (‘register’) the
authors distinguish between four registers, providing examples of texts in which these
should be used:
1. Classic register (semi-legal, very formal materials);
2. Formal register (committee and cabinet proceedings, announcements, job advertise-
ments, reports, letters, questionnaires, etc.);
64
3. Informal register (press statement, other advertisements, public campaigns, materials
for schools, leaflets, posters, stickers);
4. Spoken/very informal register (scripts of television campaigns, subtitles for television
campaigns, some posters, pamphlets, games and cartoons). (Welsh Government 2006
15-19).
Other than legal documents and public service, another domain where semi-formal registers of Welsh can be found is the press. No daily Welsh newspaper is currently
published. However, two national daily newspapers, The Western Mail and The Daily
Post and a number of local ones, contain columns in Welsh. There is also a weekly fully
Welsh-language magazine Golwg. A weekly newspaper Y Cymro was published until
June 2017 and relaunched in 2018. A more high-brow opinion magazine, Barn, is published once a month. There is also a considerable number of local newspapers, magazines
and specialist publications available in Welsh, including the so-called papurau bro,
monthly community newsletters. Welsh magazines are now available also in digital form
for iPhone users via an application Ap Cylchgronau Cymru. Golwg also runs online news
portals Golwg360. Another popular Welsh-language news portal is the BBC Cymru Fyw.
No linguistic analysis of Welsh press published these days has been found. However, it cannot be doubted that the registers of Welsh used in the press can be placed in
the middle of the formality continuum. Conservative literary forms are evidently avoided
by journalists and the language used is much closer to the colloquial variety than it was
in the past. This “relaxing” of standards was noticed by Davies (1988: 209) who attributes
it to the growth of papurau bro which made readers more accustomed to seeing colloquial
language in print. Obviously, differences in register and style can be observed across various newspapers and magazines. For example, one finds the literary pronoun ef [he] in
Barn, while more popular Golwg and Y Cymro will prefer the colloquial e and o, depending on the journalist’s dialect21. Similarly, particles with affixed pronouns and impersonal
and present/future conjugated forms of the verb, characteristic of literary language are
avoided in magazines and newspapers intended for wider audience. In terms of lexicon,
the differences are less tangible. They will be discussed in Chapter 3 of the present work
(3.2.4.2).
21
Presumably with a preference for e considered more “standard” being more similar to ef.
65
The third domain of semi-formal Welsh is popular literature, which, as mentioned
in the previous section, uses the whole range of available registers in Welsh. The differences here are largely idiosyncratic and particular to a given writer. Some remarks on the
lexical norms in selected works of contemporary Welsh authors will be included in Chapter 3 (3.2.4.1).
To summarise, one can observe the development of a semi-formal variety of
Welsh which differs from traditional literary Welsh mostly in grammatical aspects but
also on the lexical level. This variety is to be found in press and literature and there is
a growing tendency to use it in administrative and other official contexts at the cost of
‘conservative’ or ‘traditional’ forms of literary Welsh. Groups of professionals responsible for creating this variety – translators, journalists, writers – are constantly striving to
produce a form of language which would be easily accessible to the average speaker, by
striking the balance between traditional literary norm and the spoken colloquial of today.
1.2.5. Spoken Standard
In contrast to the written standard of literary Welsh, the spoken standard is a fairly modern
concept. Before the 19th century, relatively little attention was paid to the spoken idiom
as it was the literary language that was considered the prestigious variety. Consequently,
prior to the advent of modern novel and press, spoken varieties were little recorded in
writing. Furthermore, as Wales did not have a powerful political centre for centuries, no
standard accent similar to English RP emerged (Morris Jones 1993: 177). By the early
20th century, however, a prestigious H register, the so-called “pulpit” Welsh, was used
in contexts mainly associated with religion and cultural events. In the second half of the
century, with the growing number of L2 speakers, the Cymraeg Byw (‘living Welsh’)
initiative attempted to establish a spoken standard for educational purposes. Finally, the
advent of Welsh-language media and the possibility to use Welsh in jurisdiction and political institutions created a new context for a nationally recognised spoken norm.
With this polycentric notion of the standard language, it is hardly surprising that
the spoken linguistic norm in Welsh is not easily definable. In the typologies presented
in Table 1, standard spoken varieties were recognised in the middle of the formality continuum (Roberts and Jones 1974; Jones and Thomas 1977). Similarly, in the discussion
66
of Welsh for Adults courses, Jones and Chapman (2000: 54–55) identified certain forms
introduced in classrooms as standard, placing them at the same level as the semi-formal
Official Welsh.
R.M. Jones, one of the creators of Cymraeg Byw, lists three main issues inherent
in any attempts to create a standard spoken variety of Welsh (2011: 85):
the relation of the standard to the literary written language i.e. degree of formality,
the relation to dialect i.e. including or rejecting particular elements of dialects,
the relation to the sub-standard i.e. rejecting elements of learner’s language.
The sections to follow will discuss the conservative H spoken variety (“pulpit” Welsh),
Cymraeg Byw and mass-media language in relation to these three areas.
1.2.5.1. Conservative spoken standard – “pulpit” Welsh
The first variety to be considered a modern spoken standard in Welsh was the so-called
“pulpit” Welsh, the language used by Non-conformist preachers. As can be expected, this
register emerged out of literary Welsh – biblical language in particular – and in consequence was characterized by literary conservative grammatical forms and careful, orthographic pronunciation devoid of dialectal features. With the 18th-century religious revival
and the unprecedented level of literacy in Wales, the language of the pulpit became the H
spoken variety for ordinary people, used in formal discussions or public addresses
(Thomas 1987: 102). According to Williams (1994: 86), when the first Welsh-language
media were created, this variety was also adopted as the language of formal broadcasts
e.g. news bulletins. However, when Welsh became widely used in a less formal context,
such as education and popular programmes, the conservative variety did not have enough
appeal to ordinary users. As a descendant of literary Welsh, pulpit Welsh remained confined to the domains of religion and literary and musical events, such as the eisteddfod,
as many people found it too detached from the colloquial idiom to use it in everyday life
(Morris Jones 1993: 193). Obviously, this concerned also the growing number of learners
who did not have sufficient social background or confidence to use this variety (Thomas
67
1987: 102; Williams 1994: 86). On the other hand, it can be argued that pulpit Welsh
fulfilled the criteria of authenticity and universality, in that it evolved naturally out of
religious writings which were once familiar to the majority of the society, and in that it
did not favour any regional variety of the language (cf. Thomas 1982). It carried a weight
of tradition and prestige, difficult to achieve by artificially created standard varieties, such
as Cymraeg Byw.
1.2.5.2. Cymraeg Byw
Creating a new standard out of vernacular varieties presented considerable difficulties.
Firstly, the well-established, conservative register derived from the written language had
to be unified with a wide range of its spoken realisations; secondly, a general high regard
of Welsh speakers for dialects as forms of ‘pure’, traditional Welsh had to be taken into
account (Thomas 1987: 100). Hence there emerged an idea of creating a new variety
which would combine elements of different dialects and at the same time bridge the gap
between the literary and colloquial. The most influential attempt to achieve this was the
Cymraeg Byw [Living Welsh] project. It was funded by the government-sponsored Welsh
Joint Education Committee and meant to provide guidelines for teaching a standardised
form of Welsh in classroom environment. Three booklets of Cymraeg Byw were published between 1964 and1970, although the idea itself emerged earlier (Davies 1988:
200–201).
The Welsh taught to L2 speakers in the post-war period was usually the literary
language, quite unlike the colloquial idiom learners heard around them (Davies 1988:
202). For that reason, the new Cymraeg Byw approach rejected those grammatical forms
and patterns which were claimed to be absent from the spoken language. This led to simplification of certain syntactical and morphological features which were characteristic of
literary Welsh, yet often not entirely absent from the colloquial idiom.
This strategy turned out to be the main reason for which Cymraeg Byw was widely
criticised by teachers (cf. Davies 1988: 204) and linguists (e.g. Thomas 1967, 1982; Fife
1986: 133; D. G. Jones 1988: 148). One objection raised towards the model was that it
was too selective and presented a misleading picture of the language spoken. Secondly,
68
it was criticised for suggesting a clear-cut division between the spoken ‘living’ and literary, ‘dead’ language, although in its creators’ view, it was supposed to serve as a starting
point for getting familiar with both registers (Davies 1988: 204).
The envisioned unification of dialect also presented difficulties. Morris Jones
(1993: 242) argues that Cymraeg Byw in fact allowed for a larger degree of dialectal variation than would have been expected from the original assumptions and thus condoned
different dialectal forms, rather than unified them. Nevertheless, it seems that the final
product was rejected by many native speakers as artificial, as claimed by Jones, who explains it with the fact that while traditional standard Welsh represented a dialect “common
core”, Cymraeg Byw is a dialectal fusion, which makes it strange to everyone (Jones 1998:
271–272). Similar opinion was expressed by Fife, who calls Cymraeg Byw an artificial
blend of dialects with a strong provenance to literary Welsh, which makes it an unnatural
“second language” (1986: 144).
All in all, although Cymraeg Byw certainly brought a change in classroom strategies of teaching Welsh to L2 speakers, in particular in Welsh courses for adults (Davies
1988: 206), its influence does not appear to be too pervasive in other domains, such as
the language of the media (cf. Thomas 1987: 111).
1.2.5.3. Broadcast Welsh
The Welsh-language media BBC Radio Cymru and the television channel S4C, described
as “a success story in the world of lesser-used languages” (Löffler 2006: 1102) are one of
the most powerful channels through which linguistic norms can be introduced. They constitute a national forum for discussion of a wide variety of subjects, making the audience
familiar with modern technical vocabulary as well as a range of dialectal forms and different registers. This may lead to creating a nationally recognised spoken standard:
The language policies and linguistic practices of a television channel can have a far-reaching effect within a minority language community. The fundamental televisual convention
of juxtaposing the audio and the visual within the mise-en-scène lends itself to the practice
of using lesser-known or new forms of speech in a context that is reinforced by paralinguistic elements and the use of images. This can extend the dissemination and comprehension
of new or unfamiliar linguistic elements within the language community that in turn may
well lead to further dissemination and use by third parties. Similarly, it can promote mutual
understanding of different dialects, if it chooses to allow or encourage its presenters and
69
scriptwriters to use dialect forms of speech in addition to (or instead of) contributing to the
development of a ‘standard oral language’. (Jones 2007: 194)
Although it cannot be doubted that Welsh language broadcast reaches a significant
percentage of Welsh speakers, the viewing figures of recent years show a decrease in the
popularity of both the radio and television. In 2014-2015, BBC Radio Cymru reached an
average of 119,000 adult listeners per week, a drop by 25,000 in comparison to the previous year (British Broadcasting Corporation 2015: 3)22. S4C had an average of 360,000
viewers a week in Wales in 2015, out of whom 173,000 were Welsh-speaking23. The
decrease is attributed mainly to changing demographics and the falling number of households where everyone speaks Welsh (S4C 2015: 7–13). Another factor might be the number and quality of programmes produced, bearing in mind that funding for the Welshlanguage media is constantly being reduced – for instance the cost per hour of S4C broadcast hours has been cut by 35% since 2009 (S4C 2015: 17).
As regards the linguistic norm and standards of the language spoken, the Welsh
language media face a number of challenges connected with the three dimensions crucial
for a spoken standard mentioned before: the degree of formality, presence of dialects and
learners’ language. An additional extremely important dimension are the strategies to be
employed in a bilingual environment.
As can be expected, the level of formality of broadcast Welsh varies depending
on the type of programme. H varieties are bound to be used in factual programmes, such
as news or documentaries, while L varieties might be expected in entertainment programmes. This was shown by Ball et al. (1988: 187–190) in their study of BBC Radio
Cymru broadcasts. Although originally, as mentioned above, some of the broadcast language was based on the “pulpit” Welsh variety and used the most conservative forms, it
was already noticed by C. Thomas (Thomas 1982: 102) that media language was generally less formal than the traditional H variety. This trend continued` with the development
of Welsh-language television, which introduced a policy of avoiding high-brow language
and allowing for the capabilities of the average speaker (Robert 2013: 119). It can be
22
The second Welsh-language radio channel, Radio Cymru 2, has been opened in January 2018 and is
planned to broadcast two hours a day. The aim of establishing the channel is to attract younger audience
(Golwg360 2018).
23
These figures do not include viewers who used online viewing platforms iPlayer and Clic, which are
becoming increasingly popular (S4C 2015: 43).
70
easily observed that at present the most conservative grammatical forms are generally
avoided in media language.
Dialectal diversity is another issue that programme-makers need to cope with.
Jones conducted an analysis of dialectal features in the media, concluding that there is an
emerging “media dialect” in the speech of media presenters (Jones 1998: 1282). This
variety often contains phonological and lexical features of dialects, yet over-use of local
vocabulary might be discouraged. On the other hand, it is emphasised that the reporters
should use vocabulary which is natural to them. Jones also observed that programmes for
L2 speakers presented a simplified, de-dialectalized variety of Welsh, devoid of more
complex constructions and idioms (Jones 1998: 275).
Investigating the usage of dialects in fictional programmes, Sayers conducted interviews with the producer of the most popular Welsh soap opera Pobl y Cwm (2009:
330–333). The answers suggest that although a number of local dialects are represented
in the show, dialectal features which would be incomprehensible to a wider audience are
restricted. This illustrates a situation when a balance needs to be found between representing linguistic diversity and accessibility of the speech presented.
These are not the only tensions that S4C has to face being the only Welsh language
channel. Yet another challenge are diversified linguistic abilities of bilingual listeners and
viewers. This is acknowledged for example in the S4C mission statement:
[w]e will need to continue to deliver content for fluent speakers, for those who are less
fluent, for those who are learning as well as those who may not be able to speak Welsh.
(S4C 2014: 26–27).
As can be seen, the channel aims to cater for the needs of viewers with little command of Welsh and also those who do not speak it at all. Currently the main strategy
applied by S4C is to providing English-langue subtitling. In 2015, subtitles in English
were available on 78.05% of Welsh language programmes, whilst Welsh subtitles were
available for only 13 hours a week (S4C 2015: 35). This illustrates the imbalance between
serving different groups of speakers (in this case non-Welsh speaking viewers, who need
English and learners who may prefer Welsh subtitles) and difficult in implementing, presumably due to budget restrictions.
In the public debate over Welsh-language media, the use of English is generally
the most controversial issue. To give a recent illustration of this, in 2016, S4C made an
71
attempt to broaden its audience by further increasing the number of subtitles in English
without the option of turning them off. This experiment has caused some public uproar,
including an official complaint by Cymdeithas yr Iaith, as many Welsh-speakers believed
that an excessive presence of English in the Welsh-language media is threatening for the
language’s future (Williams 2016). A similar controversy regarding radio arose when
BBC Radio Cymru planned to increase the number of English-language songs played
(Golwg360: 2014).
It has been observed that, as far as linguistic standards are concerned, programmemakers in Welsh-language media make an effort to maintain the monolingual norm and
adhere to standards of correctness and purity. Robert mentions that the S4C guidelines
for presenters discourage “needless literary translations from English” and oblige the socalled professional contributors to use “’high standard of correctness and articulation”,
including correct mutations and noun gender (Robert 2013: 122). No such limitations
however, can be put on guests invited to the programmes and the presenters might accommodate to the language of their interlocutors.
All in all, as Robert (2013: 122) claims, S4C guidelines are rather ambiguous.
They refer to the notions of standardness and correctness but the meaning of these notions
is not specified. On the other hand, they need to face a number of difficulties connected
with diglossia, bilingualism and a small number of highly proficient Welsh-speaking contributors. The tensions have been aptly summarised by Uribe-Jongbloed (2015: 7):
Finding Welsh-speaking contributors is more difficult, when not impossible, compared to
finding those who can speak English. Also, because they actively try to represent the whole
of Wales through their production, they accept varying levels of language ability in Welsh.
They find themselves in a quandary. On the one hand, they have to follow strict guidelines
defined formally or informally in the production process based on the S4C (2008) language
policy guidelines. On the other hand, they strive to produce programmes where people
speak the way they normally do, including dialects and English borrowings.
The above evidence illustrates the fact that the Welsh used in the Welsh-language media
(in particular television) is far from standardised. The variety of programmes available
opens the possibility of portraying Welsh in all its diversity, both as regards dialects and
registers and thus, as Ball et al. (1988: 90) suggested, may teach about the appropriateness
of a given variety to a given situation. Similarly to the language of the press, broadcast
language is in the process of developing a variety that would maintain a degree of ‘correctness’ and ‘standardness’ inherent in the monolingual linguistic norms, at the same
72
time ensuring inclusion of audience with a limited command of Welsh. With the decline
in the number of viewers and listeners of Welsh-language media, pressures to make the
language more accessible to learners and attract non-Welsh speaking audience are bound
to increase and broadcast Welsh may become increasingly informal, if not simplified.
1.2.6. Lexical planning – dictionaries and terminology
The last area to be described as regards the standardness of Welsh is corpus planning in
the context of dictionary-making and terminology planning. As can be seen from previous
sections, there is relatively little conscious language-planning in Wales, both as regards
grammar and lexicon. As far as grammar is concerned, some norms might be introduced
via projects such as Cymraeg Byw and Cymraeg Clir or the media. Dictionary making is
another area where some degree of lexical corpus planning occurs, especially in relation
to the expansion of vocabulary through both terminology planning and lexicography
(Baker 2010: 108; Robert 2013: 16–17).
General monolingual and bilingual dictionaries play an important role in the process of language standardisation in several respects. They serve as an “arbiter of linguistic
standards” (Hartmann and James 2002: 10) and, whether written within a more descriptive or prescriptive approach, are a reliable source of authority for the average user. In
consequence, they are an essential reference work for students and professionals dealing
with language on everyday basis, such as translators or journalists (see 5.4.1.1). Dictionaries standardise the orthography and, what is of particular significance to the present
thesis, recognise certain lexical items as part of the lexicon (Gouws 2009: 266), thus legitimising the usage of a particular word or phrase. This section will present an overview
of Welsh dictionaries focusing on the degree of their normativity and descriptiveness.
Welsh lexicography, although lagging somewhat unsurprisingly behind mainstream languages, is at a relatively advanced stage of development, especially in comparison with other Celtic languages (cf. Heinz 2003: 473–486). Despite the fact that some
types of dictionaries are still missing, most notably a pronunciation dictionary, an immense progress in Welsh dictionary-making and above all in dictionaries’ accessibility
can be observed in recent years, largely due to technological developments.
73
The current state of Welsh lexicography is of particular relevance for this thesis,
which focuses on the lexical aspects of the linguistic norm and analyses the representation
of phrasal verbs in Welsh dictionaries (Chapter 4). For this reason, the following sections
will describe major lexicographic works currently available. The overview will begin
with two largest and most important general-purpose dictionaries of Welsh, Geiriadur
Prifysgol Cymru and Geiriadur yr Academi in their online versions. Although originally
printed and still available in paper editions, both works have been digitalised and their
online versions updated.
1.2.6.1. Geiriadur Prifysgol Cymru
Geiriadur Prifysgol Cymru: a Dictionary of the Welsh Language (GPC) (Thomas et al.
1967-2002) is a semi-bilingual Welsh-Welsh-English dictionary. This monumental project began in the early 1920s as the first comprehensive historical dictionary of Welsh,
a Welsh counterpart of the Oxford English Dictionary. The first part was published in
1950, but it was not until 2002 that the letter ‘z’ was reached. By that time, however, the
first volumes required a thorough revision and the work on the 2nd edition is currently in
progress. Since 2014 the dictionary has been available online (University of Wales 2014).
Being the most comprehensive Welsh dictionary, GPC is recommended as serving
as the main spelling reference for specialists such as translators (Prys 2015: 42). Other
than that, however, as a historical dictionary GPC is of limited use to the average user.
GPC is highly descriptive in nature, including obsolete items and coinages, as well as
well-established items. The dictionary also contains a great number of colloquial and dialectal items, usually placed at the end of an entry and marked by a label Ar lafar (‘spoken’) which suggests their non-standard nature. It can be noticed that the last volumes of
the dictionary contain many more of such forms than the first ones, which reflects
a change in attitudes towards spoken Welsh which had taken place during the decades of
compilation. For the same reason, some modern terms may be absent from the wordlist,
which is, however, being updated on a regular basis24. All in all, GPC is an immensely
important dictionary intended to serve as a standard for orthography, a token of prestige
24
For example, the word e-bost ‘e-mail’ was added only in 2016.
74
and an invaluable source on the history of the language. On the other hand, due to the fact
that it was practically unavailable to the average user until recently, its popularity might
be confined to a narrow circle of specialists. No quantitative studies of GPC’s popularity
among users have been located, however.
1.2.6.2. Geiriadur yr Academi
Geiriadur yr Academi, (GA) (Griffiths and Jones 1995) is probably one of the most influential and popular Welsh dictionaries, which is illustrated by the fact it has the colloquial name of “Bruce” after one of its editors. Geiriadur yr Academi is a large EnglishWelsh dictionary based on the Shorter English and French Dictionary of 1975. It was
published by a society of writers called The Welsh Academy, with the purpose of presenting the Welsh language in all its richness, but also of filling the gaps in Welsh vocabulary (Griffiths and Jones 1995: ix–xi). It offers an extensive coverage of phraseology
and comprehensive grammatical information, all of which makes it highly useful in composing texts in Welsh. Since 2012, the dictionary has been available online, but the digitalisation is still in progress, with many incomplete entries marked in pink colour.
The issue of GA’s normativity is quite complex. On the one hand the prescriptive
approach is reflected in a number of items coined by the authors when an obvious Welsh
equivalent was missing. As a consequence, it is not uncommon for the dictionary’s user
to encounter words which have not been accepted into general use and may appear artificial and obsolete. On the other hand, Prys (2015: 43), describes GA as a descriptive dictionary, in that it does not impose specific terminology, instead recording a range of items
that occurred at the time the dictionary was compiled. What is more, the Welsh equivalents encompass a wide variety of registers, including a plethora of dialectal forms. Items
belonging to various registers are labelled accordingly, creating a sharp division between
standard literary Welsh and local varieties.
As regards standard Welsh, in particular grammar, GA’s approach is often
strongly prescriptive. This concerns especially patterns that seem to be calqued from English. For example in the entry for if, introducing the sense ‘whether’, one can find the
following judgemental view:
75
C. if = whether, introducing indirect questions, after verbs of asking, doubting, wondering,
etc.: (the use of os in this sense is an Anglicism and a grotesque solecism) 25
In summary, Geiriadur yr Academi attempts to set a strict standard of correctness
as regards the “purity” of Welsh grammar and vocabulary, at the same time demonstrating
the richness of Welsh in all its dialectal variations. The dictionary has been observed to
function as the main authoritative reference work for educated Welsh speakers. For instance, Robert quotes her interviewees, creators of a S4C programme, asking rhetorically
“Ble bydden ni heb Bruce?” (‘Where would we be without Bruce?’) (Robert 2013: 179).
It was the most popular dictionary included in the quantitative study by Prys et al. with
an average of 117,415 searches per month (2015: 356), far outnumbering the other dictionaries investigated. The same was found among speakers interviewed for the present
study (5.4.1.1), However, as the editors of the dictionary state in a press interview, the
process of updating the online version of the dictionary is at a standstill and it may become
outdated in the near future (BBC Cymru Fyw 2015).
1.2.6.3. Other online dictionaries
Other online dictionaries that need to be mentioned are Trinity Saint David’s dictionary
geiriadur.net (TSD) (University of Wales Trinity Saint David 2018), Geiriadur Bangor
(GB) (Bangor University 2016) and Gweiadur (GW).
TSD is an extensive bi-directional bilingual wordlist of over 250,000 headwords
based on a Camu language course developed at the University of Lampeter, now part of
Trinity Saint David University (Campbell 2005). Offering the option of a phrase, as well
as word search, the dictionary contains a large number of phrases and example sentences.
These are mostly related to official contexts, for example quotations from official documents. The dictionary contains many colloquial items, but few dialectal forms. No labels
distinguishing between registers are used. No figures on the popularity of this dictionary
are available.
25
Since os should be used in the sense of conditional if only. However, calquing the meaning of this word
to the English sense ‘whether’ seems to be very common among the younger generation of Welsh speakers.
76
Geiriadur Bangor (GB) (Bangor University 2016), which replaced the popular
BBC Welsh Dictionary in 2014, was developed by Canolfan Bedwyr of Bangor University. It is a bi-directional bilingual aggregate of two dictionaries: a general Cysgair dictionary and a terminology dictionary Y Termiadur Addysg (see section 1.2.6.5). It is also
available as a mobile application ApGeiriaduron. This dictionary is an extensive wordlist
containing a number of technical terms, targeted primarily at learners of Welsh. It contains few phrases and no example sentences or register labels. As for 2014, the dictionary
had an average of over 4,000 searches per month (Prys et al. 2015: 356)
One of the most recent and innovative projects in Welsh lexicography is Gweiadur
(GWE) by D. Geraint Lewis launched in 2013 (Lewis 2018). It is based on the author’s
previous works aimed at learners of Welsh, such as Geiriadur Gomer i’r Ifanc (1994)
dictionary, and a grammar guidebooks Y Llyfr Berfau (1995). The core of the project is
a large semi-bilingual Welsh-Welsh-English dictionary, accompanied with an EnglishWelsh index. The entries include definitions in Welsh, English equivalents, examples and
usage notes, phrases and idioms, audio pronunciation and links to other dictionaries. The
greatest novelty of the dictionary, however, is that it offers full inflection patterns of
Welsh verbs, preposition and adjectives in two different varieties: formal and informal.
This is an immense step forward in providing learners with easy access to irregular grammatical forms and informs them about differences across registers. In terms of lexicon,
GWE generally avoids prescriptive labels, although it does include normative guidelines
in usage notes, such as the one in the entry for os ‘if’ related to the calquing phenomenon
mentioned before:
peidiwch â defnyddio os i gyflwyno cwestiwn anuniongyrchol: 'Gofynnais iddo a oedd yn
dod' nid 'os oedd yn dod';
‘Do not use os to introduce indirect question: “I asked him if he was coming”’
Other features indicating the normative character of the dictionary are a small number of
dialectal words and occasional deliberate avoidance of commonly used borrowings from
English in favour of native Welsh words. For example, although the word mwstash
[‘moustache’] is to be found in the Welsh-English part, the only equivalent for moustache
provided by the English-Welsh part is a largely obscure item trawswch.
77
Using GWE requires registration. No figures on the average number of searches
of this dictionary have been found; however, as of June 2016, it had over 3,400 registered
users.
1.2.6.4. General printed dictionaries
This section will now briefly describe several smaller printed dictionaries of Welsh. The
two major volumes, GPC and GA have been described in the section above as they are
both available online.
A curious fact about smaller printed dictionaries of Welsh is that, notwithstanding
recent developments, the market continues to be dominated by works compiled more than
half a century ago. The most popular of those are Geiriadur Mawr (Evans and Thomas
eds. 1958) and Geiriadur Bach (Evans and Thomas eds. 1959), which date back to the
1950s, and Collins Spurrell (Amiot-Cadey and Seaton eds. 2009), which is in fact a 1960
revision of Spurrell’s dictionary first published in 191626. These repeatedly reprinted volumes, whose quality is much below current international standards, offer no more than
lists of decontextualized and frequently outdated headwords and equivalents. They tend
to misrepresent contemporary Welsh, and are of little use to learners (Campbell 2005;
Heinz 2003: 475; Klonowska 2013). Serious deficiencies in currency and good
knowledge of Welsh required of the user in order to discriminate between strings of decontextualized equivalents are features that disqualify these dictionaries as useful reference works for purposes other than checking spelling. The deficiencies of Collins Spurrell
have been at least partly recognised, however, as the latest edition of the dictionary (Beattie ed. 2017) has been extensively updated.
Another, learner-oriented approach to dictionary-compiling is represented by
pocket dictionaries written in the 1990s after the introduction of Welsh as a compulsory
subject in schools in Wales. The Welsh Learner's Dictionary (WLD) (Gruffudd 1998) and
The Pocket Modern Welsh Dictionary (PMWD) (King 2000), written by experienced
teachers of Welsh, were innovative not only in their attempt to include grammatical tips
for learners, but also in their focus on a chosen variety of spoken, colloquial Welsh. This
26
That is the 1916 J.B. Anwyl’s revision of original 19-th century dictionaries by William Spurrell.
78
approach was much in line with Cymraeg Byw and thus is subject to the same critical
arguments. A sharp separation of the colloquial from the literary resulted in an underrepresentation of dialectal variations and occasionally inadequate, simplified presentation of the Welsh morphology.
The most recently published dictionary of Welsh is Geiriadur Cymraeg Gomer by
Geraint and Nudd Lewis (Lewis and Lewis eds. 2016). It is a printed version of Gweiadur
and as such has the modern features mentioned above, such as comprehensive wordlist
and grammatical appendixes acknowledging different registers of Welsh.
1.2.6.5. Terminology dictionaries
The first initiatives of terminology corpus-planning for the Welsh language began in the
1940s (Lewis 1987:20). Over subsequent decades glossaries of specialist terms in various
fields were prepared by academic institutions and some basic principles for creating new
terms were established:
(…) direct borrowing where the words represented few orthographic problems, the creation
of new words using well-established roots and terminations and the resurrection of old
words giving them a more technological significance. This procedure was primarily for use
in education with Welsh medium education expanding considerably at the time. [following
the Welsh Language Act of 1967 - ML] (Williams 1994: 89)
Currently, there is a set of well-developed dictionaries of terminology, the socalled termiaduron whose aim is to unify the technical vocabulary (for a list see Prys
2015: 43–45). All major terminology dictionaries, most of which were printed, have been
available online since 2010 aggregated on a single website Porth Termau Cenedlaethol
Cymru, The Welsh National Terminology Portal (Jones et al. 2011). Other popular terminology dictionaries include Y Termiadur Addysg (Prys and Prys 2011) and Geiriadur Termau’r Coleg Cymraeg Cenedlaethol (Andrews and Prys 2015), intended for pre-university and university students, respectively.
Although terminology planning is an important part of language standardisation,
it refers primarily to the concept of unification, rather than prestige or legitimacy (Milroy
2001). However, terminology planning should not be seen as a strictly technical endeav-
79
our, since one cannot draw a distinct border between general and technical usage. Technical terms are bound to emerge also in general contexts and consequently terminology
planning involves a certain degree of linguistic purism, favouring native terms over borrowings (Robert 2013: 23-24). It is difficult to establish what impact terminology planning has on the average speaker. Robert believes that this may even lead to stigmatisation
of speakers who do not conform to the norm (2011: 149), while Baker (2003) claims that
[m]odern Welsh terminological dictionaries, (…) are valuable successes in Welsh corpus
(…) language planning. Yet these successes have relatively little impact on the daily language lives of people. None of these are foundational in affecting the everyday language of
the vast majority. (Baker 2003: 99)
Such observations are not based on empirical research, though, as no studies of the use of
Welsh dictionaries have been conducted, except from some quantitative data on the number of dictionary searches in the study by Prys et al. (2015). This study only concerned
online dictionaries developed by Canolfan Bedwyr in Bangor. With so little tangible evidence on the popularity of various dictionaries of Welsh and their actual influence on the
Welsh speakers, it can only be supposed that they do play a role in standardising the
vocabulary and setting the standards of ‘correct’ Welsh based solely on anecdotal evidence.
1.2.7. Standard and non-standard – further remarks
In comparison to many other minority languages Welsh is in a privileged situation as
regards the potential for standardisation. First and foremost, the official status of the language guarantees its functioning in the domain of law, public service and education, creating a range of institutions that may establish a linguistic norm. Welsh-language press
and broadcasting media provide every speaker of Welsh with access to the living language
used in a wide range of domains, and are potentially a good forum for negotiating linguistic standards between different social groups.
Secondly, Welsh has a long and unbroken literary tradition which formed a basis
for a highly standardised and prestigious variety of the language. The centuries-old heritage may be a source of pride to the speakers of Welsh and a socially unifying factor,
with cultural events such as eisteddfods ensuring maintenance of literary standards.
80
Thirdly, modern technology provides ample linguistic resources and favourable
conditions for efforts of corpus planners. With the digital revolution, resources which
used to be available to narrow circles of specialists, are now to be found online in a matter
of seconds. Welsh lexicography, no longer limited by space constraints, enters an era of
user-friendly, pedagogically-oriented dictionaries capable of describing the language in
all its variety and richness, thus making users familiar with variations across formal and
informal registers. The constantly developing corpus of terminology dictionaries also has
a standardising potential.
Yet, as has been demonstrated in this overview, multiple and diverse environments that create a linguistic norm translate into several channels of standardisation existing simultaneously and to some extent incompatible with one another. Morris Jones
(1993: 171) distinguishes three types of standardisation which are at play in the context
of Welsh: imposed, institutional and evolutionary. Characteristically, they are strongly
related to formal, semi-formal and informal registers of Welsh, respectively.
The first current of standardisation relates to the imposed norms that are introduced by an authority. These norms are to be found in the H varieties of Welsh which
have been labelled here “literary”, “conservative” or “traditional”. Since no language
planning institutions like the Académie Française or the Real Academia Española exist
in Wales, a country that until very recently has been devoid of a political centre, the norms
are not political in character but associated more with religious and cultural ideas, such
as the authority of the Welsh Bible and the centuries-old bardic tradition. Two other manifestations of imposed codification took place in the 20th century. The first is a normative
standardisation associated with “Oxford Welsh” and John Morris-Jones’s school of
thought, embedded for example in Orgraff yr Iaith Gymraeg. The second is Cymraeg
Byw, which undoubtedly exercised influence on education. Both of these standards function within a strictly monolingual norm, highlighting the notion of correctness and purity
of the language, although Cymraeg Byw might not always be consistent with it as it allows
for certain degree of dialectal and stylistic variation (Morris Jones 1993: 172). Some dictionaries of Welsh, which are fairly conservative and prescriptive, such as Geiriadur
Mawr or, to some extent, Geiriadur yr Academi might also be seen as belonging to that
current.
81
The second type of standardisation is connected with official institutions, such as
press, broadcast, cultural movements, education, administration and business. The varieties of Welsh used there have been hitherto named “semi-formal” or “official”. According
to Morris Jones (1993: 172), standards of this variety tend to be based on the notions of
correctness and monolingualism along the lines of Morris-Jones’s tradition; the tendency
to rely on normative dictionaries has also been indicated in the previous sections. On the
other hand, the semi-formal varieties appear to recede from using conservative grammatical patterns or elevated, ‘pure’ lexicon. This is driven by the need to remain comprehensible to contemporary Welsh speakers, the majority of whom will not be familiar with
high-brow literary language, viewed by some as elitist or old-fashioned. Indeed, in the
Welsh-language Wikipedia article on written Welsh the anonymous author describes this
emerging variety using the term “Contemporary Welsh”:
Inspired by Cymraeg Byw, a style more informal than traditional literary Welsh was developed for Welsh writing, using grammatical forms of spoken language and standard Welsh
vocabulary. It is a compromise between spoken language and formal written Welsh. It was
a reaction to the new need for written communication with the public not all of whom were
used to reading literary Welsh outside the chapel, following the request for Welsh-language
service. This style is called Contemporary Welsh or less formal Welsh. Contemporary
Welsh is to be seen in information leaflets, newspaper articles and novels27. (“Cymraeg
ysgrifenedig” 2015)
As can be seen, the new semi-formal standard is expected to draw from resources
of both written and spoken language. Gruffudd in fact proposes “careful spoken” as an
alternative to conservative H variety used in public communication (2015: 124):
I should claim that what is called “informal” or “careful spoken” is the closest to what
might be considered a standard language and also to an acceptable written form. This is the
register we should use in materials that will be read by the general public.28
“Yn sgil Cymraeg Byw datblygwyd dull mwy anffurfiol na Chymraeg lenyddol draddodiadol o ysgrifennu'r Gymraeg, mewn arddull a ddefnyddiai ffurfiau gramadegol yr iaith lafar a geirfa Cymraeg safonol.
Cyfaddawd rhwng iaith lafar a Chymraeg ysgrifenedig ffurfiol ydyw. Gelwir yr arddull hon yn Gymraeg
Cyfoes neu yn Gymraeg llai ffurfiol. Ymateb ydoedd i'r angen newydd am gyfathrebu ar bapur â'r cyhoedd
nad oeddynt i gyd wedi arfer â darllen Cymraeg llenyddol tu allan i'r capel, yn sgil y galw am wasanaeth
trwy gyfrwng y Gymraeg. Gwelir Cymraeg Cyfoes mewn taflenni gwybodaeth, erthyglau papur newydd
ac mewn nofelau.”
28
“Byddwn i'n honni mai'r hyn sydd yma'n 'anffurfiol' neu'n 'llafar gofalus' sydd agosaf at yr hyn y mae
modd ei ystyried yn iaith lafar safonol a hefyd at ffurf ysgrifenedig dderbyniol. Dyma'r cywair y byddai'n
dda i ni ei ddefnyddio mewn deunydd fydd yn cael ei ddarllen gan y cyhoedd yn gyffredinol.”
27
82
Similarly, a recent report commissioned by S4C concludes that media language
remains too formal and there is more of a need for “everyday” language that would attract
a wider audience, “normalise less perfect Welsh” and “better reflect the way that people
often speak”29. Such examples show that institutional standards are in the process of development, with a tendency to ever stronger adherence to norms of the colloquial, rather
than literary. The phrase “normalise less perfect Welsh” is a good description of this tendency, in view of the growing number of L2 speakers and seemingly decreasing familiarity with literary registers.
Finally, the third type of standardisation distinguished by Morris Jones is evolutional standardisation, which takes place through the naturally changing habits of speakers, who are not guided by authorities and institutions and who might reject imposed
norms. While the previous two types of standardisation rely primarily on the monolingual
norm and the language of influential speakers/writers, the same cannot be said of norms
emerging among speakers living in bilingual (if not mostly English-speaking) environments. One may refer to the notion of covert prestige (Chambers and Trudgill 1998: 85)
where non-standard forms may be favoured as a sign of one’s identity as a member of a
group. According to this concept, it is the colloquial forms that could be seen as more
‘prestigious’ for some Welsh speakers.
The influence of the colloquial language of bilingual speakers, a significant percent of whom are new speakers, is the least tangible and at the same time the most interesting element of the standardisation of Welsh. John Morris-Jones’s criterion of standardness based on intuitions of an educated Welsh monoglot ceased to be valid roughly
half a century ago; it is only the bilingual speaker whose intuitions can be taken into
account while discussing modern standards. Otherwise one might expect discrepancies
between imposed and institutional monolingual norms, which now have only written
standards as their basis, and the habits of ordinary speakers. At the same time, it should
not be forgotten that evolutional standardisation cannot be seen as completely natural and
“bottom-up”, since speakers’ practices are to some degree affected by institutional linguistic norms introduced for instance by media or educational institutions (Morris Jones
1993: 173).
“Portreadu Cymraeg anffurfiol mwy 'bob dydd' ar y radio a'r teledu a allai helpu i normaleiddio Cymraeg
llai perffaith, ac adlewyrchu'n well y ffordd mae pobl yn siarad yn aml (...)”.
29
83
The three types of standardisation form an intricate whole and influence one another. Tensions between the formal norm and the process of creating a “contemporary”,
semi-formal register have been indicated by a number of researchers (Fife 1986;
B. P. Jones 1988; Gruffudd 1998; Robert 2011), particularly in relation to the written language, for example in the evidence for reluctance in using official documents in Welsh
presented above. Similar issues have been observed in the spoken mode. Müller and Ball
describe the phenomenon of interlinguistic diglossia (Müller and Ball 2005: 12) where
speakers of Welsh who do not have much access to the H variety prefer to switch to
English. The same has been observed by Kaufman in her study of public meetings in city
councils in North Wales (Kaufmann 2009: 134), where some Welsh speakers preferred
using English in situations requiring formal registers.
The present time is one of dynamic changes within the Welsh-speaking community, not only in view of the internal factors, such as mixed abilities of the speakers, low
confidence levels and various access to planned language (Robert 2013: 107), but also
external ones, such as the Internet culture which has deeply changed the character of
written language and the relation between spoken and colloquial. Difference between the
written and spoken have been waning due to the variety of social activities pursued with
new forms of communication such as e-mail, text messaging and social networking sites
(Robert 2013: 26). On the other hand, one should be careful in discarding the most formal
H varieties of Welsh and labelling them as ‘dead’, ‘traditional’ and ‘conservative’, as they
constitute a vital part of the linguistic landscape. However, there is the need to reconsider
a notion of standard based on classical written language, which, as Millar puts it, may
“exert such an influence upon the new standard as to stunt its development, or at least
divert its development in otherwise unexpected directions, since they do not encourage
a close connection between the spoken and written contemporary language” (Millar
2005: 92).
In the context of the present investigation of the standardness of phrasal verbs in
Welsh, the focus will therefore be on the developing semi-formal registers of Welsh and
the integration and acceptability of phrasal verbs therein. It is notable that in comparing
different registers the above discussion of standards has concentrated on the grammatical
features rather than lexicon. Indeed, as Jones and Chapman (2000: 53) comment:
The difference between the literary, official and colloquial language is not obvious as regards the vocabulary, but rather in syntax, grammar and morphology. Therefore register is
84
not particularly important while discussing the subjects of teaching vocabulary to learners30.
However, it cannot be doubted that ideologies of lexical purism exist in the minds
of many speakers, which is expressed, for example, in using negatively loaded words such
as annerbyniol (unacceptable), ansafonol (non-standard), anghywir (incorrect), amhur
(impure), llygredig (corrupted), dirywiad (decline) and bratiaith (‘slang’), while talking
about non-standard varieties of Welsh (Morris Jones 1993: 126). Negative valuing of certain lexical items borrowed from English is undoubtedly an element of this ideology. The
English elements in Welsh are one of the tokens of the sub-standard, defined here “a
manner of discourse which deviates in pronunciation, grammar and vocabulary from a
common standard. As opposed to the neutral term non-standard, sub-standard dialects
may carry a social stigma.” (Hartmann and James 2002: 225). As the definition indicates,
the notion of sub-standardness is inherently connected with value-judgements, measuring
speakers’ achievement in mastering the language.
It is precisely in the standard written language that language mixing is deemed poor and
crude, and at worst proof of a lack of linguistic competence. (…) The written standard thus
gains a taken-for-granted status to such an extent that it even regularly serves as a benchmark for “good” spoken language. (Musk 2012: 664)
The negative valuing of borrowing and language-mixing is thus connected with
linguistic prescriptivism or purism which stem from cultural and political factors, as well
as linguistic discourse within the academia. Such ideas naturally change over time. The
next part of the chapter will present an overview of the influence of English on Welsh
and changing views on borrowings from English in order to trace the evolution of attitudes towards them observed today.
1.3. Welsh and English in contact – the influence of English on Welsh
Although the contact between Welsh and English began centuries ago, English did not
become the major language in Wales until the 20th century (Löffler 2008: 350). While
“Nid yw’r gwahaniaeth rhwng yr iaith lenyddol, yr iaith swyddogol a’r iaith lafar yn amlygu ei hun yng
ngeirfa’r iaith, ond yn hytrach mae’n amlwg o ran cystrawen, gramadeg a morffoleg. Felly, nid yw’r cywair
hwn yn arbennig o berthnasol wrth drafod y pwnc o ddysgu geirfa i ddysgwyr.”
30
85
some parts of Wales were English-speaking already in the early Middle Ages, the two
languages remained clearly separated and little bilingualism existed until the 18th century. Anglo-Welsh literature developed only in the 1920s. Filppula et al. (2008: 137) call
this “diglossia without bilingualism” as English and Welsh had different social functions
and the contact between the two speakers’ communities was limited. The situation
changed after the industrial revolution, which marked the beginning of the language shift.
The number of Welsh monoglots dropped rapidly; the 1981 National Census was the last
one in which they were included. Nonetheless, although English is spoken by nearly all
the inhabitants of Wales and is the dominant language in the country, Welsh appears to
be a stronger identity marker than English (Löffler 2008: 350).
Changes in Welsh brought by contact with English are easily observable in the
presence of numerous lexical borrowings, changes in the language’s morphology and
syntax and other mechanisms such as widespread code-switching. Paradoxically, the
strong influence of English on Welsh has been given relatively little attention among
scholars until recently, be it for ideological reasons or presumed obviousness of the influence. Although in the last two decades the number of publications concerning various
aspects of contact-induced language change in Welsh has grown, research on lexical borrowing, particularly multiword items, has been scarce. The following subsections will
define and discuss concepts related to language contact and lexical transfer relevant to
the thesis, and present an overview of the available literature on the influence of English
on Welsh, with focus on phraseology and idioms. The final part will reflect on the standardisation-related ideologies among Welsh speakers regarding the English element in
Welsh.
1.3.1. Language contact, language change and transfer
1.3.1.1. Language contact
Language contact is defined here as the use of more than one language in the same place
at the same time (Thomason 2001: 1). The present study falls within the field of contact
linguistics, which is a relatively new area of research, embracing various academic fields
86
such as linguistics, psycholinguistics, sociolinguistics, anthropology, and education
(Heine and Kuteva 2005: 5). Systematic studies of language contact began in mid-twentieth century with the ground-breaking works by Weinreich (1953) and Haugen (1950,
1953). After the period when relatively little research into contact situations was done
due to the rise of generative linguistics and its focus on Universal Grammar, interest in
the discipline was re-sparked by an influential publication by Thomason and Kaufman
Language contact, creolization and genetic linguistics (1988), which set the basis for
theoretical and empirical framework for language contact studies, further developed in
works such as van Coetsem (2000), Grosjean (2001), Thomason (2001), Myers-Scotton
(2002) Winford (2003), Heine and Kuteva (2005) and Hickey (2010). Currently, contact
linguistics is concerned with all processes related to language contact, be it socio-political
conflicts or language change in terms of lexicon, morphosyntactic aspects of the grammar
and phonology (Myers-Scotton 2002: 4).
The present study discusses verb-particle constructions as a case of potential lexical transfer between Welsh and English. The major process involved in lexical transfer
between languages is borrowing, which in the case of minority languages as a rule occurs
in one direction, that is borrowing words from dominant into minority language (Sankoff
2001: 649). Based on literature in contact linguistics this section will briefly delineate
issues and concepts related to contact-induced change valid for this thesis and establish
terminology used in the subsequent chapters.
Prior to the publication of Thomason and Kaufman (1988), studies on borrowing
within contact linguistics were fairly rare as language change had been most often attributed to internal causes. However, in the last thirty years much research has been done
with regard to grammatical constraints on borrowing, phonological, syntactic and morphological integration of borrowed words, the classification of borrowings and distinguishing them from other language contact phenomena e.g. code-switching (TreffersDaller 2010: 17-18, Winford 2010). While discussing contact-induced mechanisms, it
should be noted, however, that the emergence of idiomatic phrasal verbs in Welsh might
not be attributable solely to language contact, but also to internal language development
or a combination of the two. There are cases when classifying directly translatable items,
such as Welsh phrasal verbs, as borrowings or native constructions proves to be an extremely difficult task, not the least because of difficulties in ascertaining what ‘native’
means. Chronological classification can be made only on the basis of written attestations
87
(with historical dictionaries as the easiest accessible source of data), thus ignoring the fact
that a word might have entered the spoken language a long time before (cf. Flohr
2013:69). This question will be further addressed in Chapters 2 and 3 considering some
diachronic aspects of phrasal verbs in Welsh and the potential of the language to create
its own idiomatic phrasal verbs without the influence of English. Despite these reservations it cannot be doubted that with regard to the majority of Welsh idiomatic phrasal
verbs contact-induced processes are at play.
1.3.1.2. Contact-induced change and lexical transfer
Contact-induced change is defined as “the adoption of a structural feature into a language
as a result of some level of bilingualism in the history of the relevant speech community”
(Treffers-Daller 2012: 56). A framework for analysing language change has been established by Thomason and Kaufman (1988) who challenged the traditional structuralist
view that contact-induced language depends on purely linguistic factors, i.e. the structure
of the languages in contact. In these researchers’ view the primary determinant of contactinduced change are social factors, such as prestige, intensity of contact and language attitudes (Thomason and Kaufman 1988: 35). According to Backus (2004: 179), the causes
of contact-induced change can be further assigned to ultimate causes – social factors, such
as dominance and prestige – which find their reflection in proximal causal mechanisms,
such as code-switching.
The effects of contact-induced change are commonly referred to as transfer of
linguistic material (Heine and Kuteva 2005: 4). The term “transfer” has been used interchangeably with “interference”, especially in early studies of language contact, such as
Weinreich (1953:1), who defined interference phenomena broadly as “those instances of
deviation from the norms of either language which occur in the speech of bilinguals as
a result of language contact”. However, the present study follows Grosjean (2012), in
whose model transfer is used in a narrower sense of “static phenomena which reflect
permanent traces of one language (La) on the other (Lb) in the bilingual”, distinguished
from interference, which denotes “the dynamic phenomena which are elements of the
other language which slip into the output of the language being spoken (or written) and
88
hence interfere with it” (Grosjean 2012: 15). Thus in the present context transfer is identified with the process of borrowing. Following Haspelmath (2009: 37) the language from
which a borrowing has been transferred is termed the donor language, and the language
into which it has been transferred is the recipient language31.
While discussing potential transfer one should also consider the situation of the
recipient language, taking into account Thomason and Kaufman’s (1988) distinction between two contact situations: language shift and maintenance. Language shift occurs
when speakers abandon their native language in favour of the language of the people they
are in contact with (target language). The linguistic result has been termed substratum
interference and involves changes to the target language due to imperfect learning (Thomason and Kaufman 1988:41-42). As described before, language shift from Welsh to English occurred in Wales in the 19th and 20th century (cf. Kandler et al. 2010: 3855) and
led to the emergence of Welsh varieties of English. However, nowadays, as described in
1.1., due to revitalisation efforts the situation of Welsh is generally that of stable bilingualism, i.e. language maintenance, the situation when “the native language is maintained
but is changed by the addition of the incorporated feature” (Thomason and Kaufman
1988: 37). It is in this situation that borrowing occurs. Its nature depends on the intensity
of contact: when it is casual, changes involve only the lexicon, but as the intensity increases grammatical structures might be affected as well (Thomason and Kaufman 1988:
50). According to Sankoff (2001: 642), the situation of stable bilingualism is the one most
likely to lead to the acceptance of borrowed structures within the recipient language.
One more concept to be mentioned here is metatypy, also called extreme structural
borrowing which may take place in long-term bilingualism. This term, introduced by Ross
(1996) pertains to a maintenance situation and denotes a thorough restructuralisation of
a language due to influence of another, up to the point when every sentence in the recipient language becomes a calque (Trask 1993: 212). In his revised definition, Ross
(2007: 116) distinguishes metatypy from calquing, in that the latter precedes metatypy
chronologically. A review of literature on English borrowings in Welsh featuring these
concepts will be presented in 1.3.2.
31
Several alternative terms are used in the literature, notably those introduced by Weinreich, who distinguished between source language and recipient language for the transfer of morphemes and model language and replica language when no morphemes are transferred (Weinreich 1963: 30-31).
89
It should be also borne in mind while discussing mechanisms of transfer and the
extent to which they occur that bilingual societies such as Welsh are characterised by
greater inter-individual variation, varying degrees of proficiency and lack of uniformed
structural base in comparison with monolingual and majority-language communities
(Sankoff 2001: 642). Moreover, individual speakers may switch between different linguistic modes. This term has been introduced by Grosjean and is defined as “the state of
activation of the bilingual’s languages and language processing mechanisms at a given
point in time” (Grosjean 2012: 12). In this view bilingualism is a continuum, where speakers are able to adapt the level of interference in their language to their interlocutor and
switch between ‘monolingual’ and ‘bilingual’ modes. The mode modulates the degree of
transfer and interference in the speech of an individual, with code switching and borrowing occurring in the bilingual situation and being supressed in monolingual contexts
(Grosjean 2012: 13). This distinction is topical for the thesis which investigates the linguistic norm in Welsh and therefore focuses on the monolingual mode embedded mostly
in writing.
The influence of one language on another may manifest itself in the transfer of
meaning, form, structure, or various combinations thereof (Heine and Kuteva 2005: 2).
Moreover, other than creating new items and structures, contact influence may also lead
to changes in the frequency of use of constructions already existing in the language (Pakendorf 2013: 200). As will be discussed in the subsequent chapters, the transfer of
phrasal verbs involves several transfer mechanisms, not only on the level of lexicon but
also structure and frequency of use. For this reason, further discussion requires the use of
some concepts related to borrowing based on structuralist frameworks, which will be now
described.
1.3.1.3. Borrowing and its integration
The term borrowing has been defined in many different ways. Debates among scholars
over the definition concern the question whether and how it should be differentiated from
interference as well as transfer of meaning and syntactic relations (cf. Winford 2010). For
the purpose of this thesis and concerning the situation of Welsh, borrowing is understood
as a term for “all kinds of transfer or copying processes, whether they are due to native
90
speakers adopting elements from other languages into the recipient language, or whether
they result from non-native speakers imposing properties of their native language onto a
recipient language” (Haspelmath 2009: 36).
A distinction useful for this thesis has been made by Myers-Scotton (2002), who
distinguishes between cultural and core borrowings. Core borrowings are words “that
more or less duplicate already existing words in the L1”, while cultural borrowings denote
objects and concepts new to the culture. Whilst core borrowings enter the recipient language gradually, cultural borrowings are likely to be quickly integrated (Myers-Scotton
2002: 41). As will be discussed in the analytical chapters of the thesis, although many
phrasal verbs in Welsh belong to the first category, competing with native items, cases of
cultural borrowings may also be found.
Regarding the competition with native words, Pulcini et al. distinguish the following three effects of a borrowing, based on Haspelmath (2009):
(a) insertion, when the loanword is adopted by the recipient language as a new
lexical item;
(b) coexistence, when the loanword is adopted by the recipient language in spite
of the existence of a native equivalent;
(c) replacement, when the loanword is adopted by the recipient language replacing an already-existent native equivalent which, as a consequence, falls out of use. (Pulcini et al. 2012: 16).
These possible effects of the emergence of borrowed phrasal verbs on native
Welsh vocabulary will be explored in the analytical chapters.
Integration is another concept crucial for the discussion of verb-particle constructions in Welsh. It is a process by which bilingual speakers adapt items in the donor language to the phonological, morphological and syntactic patterns of the recipient language
(Poplack 2017: 8). In fact, borrowing has been described as the integration of one language into another (Grosjean 2010: 58). Levels of integration of borrowings depend on a
great number of both socio-linguistic and structural aspects. Among social factors Winford (2010: 177-178) enumerates the need to express a concept that does not exist in the
recipient language (i.e. cultural borrowings), prestige of the dominant language, the degree of bilingualism, the demographics and power relationships, attitudes, language loyalty and language ideologies. Linguistic factors that may also put constraints on borrowing include the word class, degree of typological distance and congruence in structure
91
across languages (Winford 2010: 177-178). As will be discussed further in the thesis
many of these factors might play a role in the integration of borrowed phrasal verbs in
Welsh.
A notion closely related to integration is entrenchment, i.e. the degree to which a
loanword is known to a speaker, determined by the frequency of use (Backhus 2004:179).
Backhus notes that when a borrowing competes with native constructions the degree of
entrenchment depends on use of it in both the donor and recipient language: “Therefore,
it is not just the importation of L2 patterns into L1 speech that affects language change;
fluctuations in language choice, such as the encroachment of the L2 in domains previously reserved for L1, also determine the differential use of these constructi” (Backhus
2004:179). Frequency of use of phrasal verbs in semi-formal written Welsh will be examined in Chapter 3 of the thesis.
Borrowings are often placed in opposition to native words, that is words which
can be identified in the earliest known stages of a language (Haspelmath 2009: 38). Although this distinction is often recognised by speakers and researchers, from a logical
standpoint its validity is somewhat questionable, given that one can hardly ever be certain
whether or not a word had been borrowed from another language at some point in history;
therefore it may be more valid to talk about the level of integration or entrenchment of
borrowed items. According to Grosjean (2001: 335) a borrowing “is finally accepted
when it is no longer treated differently from other words in the language and when dictionaries, national academies and influential writers accept it. It is then a loanword only
in the historical sense.” Similar criteria in establishing the integration of phrasal verbs in
Welsh will be used in the thesis when examining their stylistic markedness (Chapter 3)
representation in dictionaries and teaching materials (Chapter 4) and acceptability among
professional speakers (Chapter 5).
1.3.1.4. Classification of borrowings
Borrowings can be further subdivided into several categories depending on the mechanism of the process. The first and most basic type of borrowing is loanword, where an
item is transferred from one language into another, functioning in the usual grammatical
processes of the recipient language (Hoffer 2005: 53).
92
The Welsh language is abundant in loanwords from English, particularly in its
spoken varieties (cf. Parina 2010). In fact, there is a considerable difference in the frequency of use of English loanwords between spoken and written language. According to
Thomas (1987: 107), in the second half of the 20th century Welsh underwent massive
relexification with a great number of Welsh words introduced into formal language to
replace Anglicisms. As a result, many loanwords from English exist along with native
terms (see also Morris Jones 1993: 86). Some borrowings seem to be preferred in the
spoken register, due to the fact that the ‘proper’ Welsh words, often recent coinages,
evoke formality and ‘big Welsh’ (cf. Fife 1986: 147; D. G. Jones 1988: 156). In some
cases there seems to be a slight semantic difference when the English item is associated
with modern, while the Welsh one with more traditional life. Hincks (2007: 14) provides
the example of cawl for ‘soup’ vs. sŵp for ‘instant soup’. A. R. Thomas (1987) claims
that that native forms are usually associated with standard and formal language, while
borrowed items pertain to the casual use, e.g. cerddoriaeth and miwsig for ‘music’. However, Robert (2013: 81) notes cases when loanwords appear to be less accepted in terms
of frequency e.g. doctor appears to be less popular than meddyg ‘doctor’.
The second type of borrowing, which is of most relevance to the present study is
calquing. In fact idiomatic phrasal verbs are one of the most recognised cases of calquing
from English into Welsh (cf. Thomas 1987: 107). A calque, also called loan translation,
is “a complex lexical unit (either a single word or a fixed phrasal expression) that was
created by an item-by-item translation of the (complex) source unit” (Haspelmath 2009:
39). In contrast to loanwords, calques are often a barely noticeable phenomenon: “it is a
form of externally motivated change that can affect both the lexis and the syntax of a
language but that appears largely undetectable at first glance due to the fact that it is the
meaning rather than the form of the foreign item which is borrowed” (Jones and Singh
2013: 40). With regard to phraseology in particular, establishing whether a unit formally
similar to another language is a calque or whether it is an outcome of internal language
development is not always easy to make (Fiedler 2012: 248). On the other hand, calques
may be also intentional creations motivated by ideologies of purism, where loan translations are introduced as part of official language planning to prevent the usage of loanwords (Görlach 2003: 97).
In Welsh calques from English occur in both single items and phrases. For example, the Welsh sglefrfwrdd ‘skateboard’ is a replication of skateboard (sglefrio ‘to
93
skate’ + bwrdd ‘board’), while y Rhyngrwyd is the equivalent of the Internet (rhwng ‘between’ + rhywd ‘net’). It should be noted that a great number of loan translations in Welsh
have been coined by language planners (Musk 2006: 406) to fill the gaps in technological
and scientific vocabulary, e.g. meddalwedd ‘software’ (meddal ‘soft’ + gwedd ‘form’),
plaleiddiad ‘pesticide’ (pla ‘pest’ + lleiddiad ‘killer’). Others are coined by speakers on
an everyday basis; a good example is Gweplyfr, a recently created Welsh name for Facebook (gwep ‘face’ + llyfr ‘book’)32. Welsh has also a great number of idioms and idiomatic phrases which appear to be calqued from English, such as cwympo mewn cariad
‘fall in love’ (for other examples see Fowkes 1945). As the majority of phrasal verbs can
be translated word-for-word into English there is a case to consider them to be calques.
This matter will be examined in detail in Chapters 2 and 3.
Other two types of borrowing to be defined are loanblends and semantic loans.
A loanblend is a form in which one element is a loanword and the other a native element
(Hoffer 2005: 54). A common type of loanblends in Welsh are English verbs to which the
derivational ending -o or -io is added, for instance cicio ‘kick”, paentio ‘paint’. As will
be shown in Chapter 2, a considerable number of phrasal verbs in Welsh, such as ffeindio
allan ‘find out’ belong to this category.
The last type of borrowing in the classification used here is semantic loan, or
loanshift, which occurs when a meaning of a given item in the donor language has been
transferred into the recipient language (Hoffer 2002: 5). An example in Welsh is llygoden
‘mouse’ used for the computer mouse. Cases of the extension of meanings of Welsh transparent phrasal verbs into idiomatic ones will be discussed in the analytical chapters of the
thesis.
1.3.1.5. Borrowing vs code-switching
A phenomenon closely related to borrowing is code-switching. It is worth mentioning in
this dissertation in the context of a type of phrasal verbs, called here, after Rottet (2005)
“wholesale borrowings” (see 2.4.4.) which are highly integrated loanwords from English
32
The name is not widely used, though, and rejected by some speakers (see interviews, 5.4.4.1).
94
and as such might be difficult to differentiate from code-switching. The term code-switching may be ambiguous due to a multitude of definitions. In the broader sense, it may
denote “all phenomena where elements from at least two linguistic systems (separate languages, or distinguishable varieties of one language) are used in the same speech situation” (Müller and Ball 2005: 49). This definition encompasses lexical borrowings, intraand intersentential code-switching and switching languages between speakers. The present study will follow a narrower definition after Thomason (2001: 132), who understands
code-switching as “the use of material from two (or more) languages by a single speaker
in the same conversation”. One may distinguish then between intrasentential and intersentential code-switching, i.e. one that occurs within one sentence or between sentences. Yet
another distinction is made between code-switching and code-mixing: the first refers to
“the choice of a language on a given occasion, depending on the circumstances or the
subject matter”, while code-mixing refers to switching within one sentence or conversation (Trask 1993: 61). This distinction will not be used in this thesis.
As the border between borrowing and code-switching may be blurred, attempts
have been made to distinguish between the two using the criteria of long-term establishment, social significance, acceptance into dictionaries and frequency; some researchers
distinguish an in-between category of nonce-borrowings (for detailed discussion see
Hoffer 2002; Stammers and Deuchar 2012; Poplack and Dion 2012) which will also not
be used in this study. The type of phrasal verbs which are not calqued but borrowed directly from English will be distinguished from code-switching in the corpus study in
Chapter 3 on the basis of their Welsh spelling in the text, which suggests phonological
integration.
Having introduced the key terms and concepts necessary for the discussion of
transfer phenomena between Welsh and English, the next section will present an overview of literature on the subject, with focus on studies on lexical borrowing.
95
1.3.2. English borrowings in Welsh – literature review
In the field of investigating lexical borrowings from English into Welsh, T.H. Parry-Williams’s (1923) ground-breaking Ph.D. thesis33 still remains the most comprehensive work
on the subject. The study is concerned with single-item loanwords rather than phrases. In
the introduction, the author presents an overview of the presence of English loanwords
throughout the history of Welsh literature and attitudes of Welsh writers and scholars
towards them. This is followed by an extensive analysis of Anglicisms extracted from a
corpus of literary texts, focusing on sound-changes and the rules that govern them (see
1.3.3).
Later works have little expanded on Parry-William’s thesis. The first modern synchronic study of the English element in Welsh was a series of articles by Fowkes (1945,
1948, 1949, 1954), each of them focusing on a different aspect of the phenomenon: idioms, prosody, initial mutations and gender of the noun. The first of the four articles
(Fowkes 1945) devoted to idioms deserves special attention in the context of this thesis.
It presents varying degrees of patency of English borrowings, from obvious loanwords to
syntactic calques. The article contains a long list of examples of calquing in both written
and spoken language, including a considerable number of phrasal verbs (see 2.3.1). Finally, Fowkes points out that similar calques can be found in the literature of past centuries, but hardly any are to be seen in Middle Welsh texts. The conclusion is that transfer
processes had begun centuries ago and accelerated in contemporary times (Fowkes
194:247-248).
The subsequent decades of the 20th century witnessed few studies on lexical borrowings, as researches seemed to focus primarily on dialectology (Thomas 1973; Thomas
and Thomas 1989; Jones 1998; Thomas ed. 2000). In the last twenty years, the majority
of studies on the English influence on Welsh were conducted in areas other than lexicon,
such as morphology, syntax and phonology. Jones (1998) investigated two diverse linguistic communities in south-east and north-east Wales in the context of language shift,
dialect loss, attitudes to Welsh and the standardising effect of education. She observed
a strong tendency among young speakers of Welsh from both communities to use grammatical forms which would be historically considered inappropriate. Phillips (2007) has
33
It was preceded by several 19th century publications (see Parina 2010: 183–184).
96
analysed changes of subsystems in the Welsh language and the loss or reduction of certain
characteristic, “Celtic” features of Welsh, such as mutations, vigesimal numerals, voiceless nasals and inflected prepositions. Pointing at intergenerational differences, he concludes that these features are “receding or already extinct” in the language of younger
speakers (Phillips 2007: 154). Other papers describing changes in Welsh under the English influence are Willis (2008) on new patterns of negation, Davies (2010) on word-order
convergence and auxiliary deletion in bilinguals and Nicoladis and Gavrila (2015) on the
production of adjectival constructions in Welsh-English bilingual children. Phonetic aspects of English borrowings in Welsh have been investigated by Buczek-Zawiła (2014)
in her study of accent accommodation and Asmus and Anderson (2015) on vowel length.
Cross-linguistic variation in the production of consonants in bilingual speech was investigated by Morris (2013) and Morris et al. (2016). Mayr et al. (2017) researched the effects
of individual bilingualism and long-term language contact on monophthongal vowel production.
Another language contact phenomenon which has attracted some attention is
code-switching (e.g. Deuchar 2006; Deuchar and Davies 2009; Musk 2010; Stammers
and Deuchar 2012; Deuchar et al. 2016). Notably, Deuchar and Davies (2009) have researched patterns of code-switching among speakers of Welsh in order to establish the
stage of advancement of the language shift. Code-switching occurred mostly in the informal language. The dominant pattern was insertion, a typical case of code-switching, indicating the first stage of bilingualism, characteristic for stable bilingualism. Stammers
and Deuchar (2012) analysed patterns of soft mutation in English-origin verbs in Welsh
and demonstrated that their integration is related to frequency, thus establishing the difference between borrowing and code-switching in Welsh. Deuchar et al.’s corpus study
(2016) found that code-switching was more widespread among young people and those
who acquired Welsh and English at the same time.
A small number of publications related to lexical borrowings have appeared in the
last ten years. Among them is ‘The English Element of Welsh’ by Grey (2008), which
remains unpublished (cited by Parina 2010: 184). Parina’s paper (2010) is one of the first
attempts at quantitative analysis of English borrowings in Welsh using the electronic CEG
corpus. In the analysis of 1000 most frequent words in the corpus, there were 87 Latin
borrowings and 40 English borrowings. Parina sets her research in the context of language
97
ideology and purism regarding negative stances towards borrowing from English and
postulates acceptance of the diversity of Welsh vocabulary.
A notable recent publication concerned with the borrowing of idioms is Flohr’s
(2013) comparative study of translations of Harry Potter and the Philosopher’s Stone
into Welsh and Irish. The author of the study calculated the number of borrowings from
English in a sample taken from these translations and compared them with translations
into other European languages. He concludes that the percentage of loanwords and
calques in the Welsh translation is exceptionally high and the English influence is much
more profound and extensive than in any other European language (Flohr 2013: 103).
Naturally, these results need to be treated with caution, as the conclusions are derived
from a single text sample which may reflect the translator’s idiolect or her individual
translation strategy.
Another study demonstrating the extent of English influence on Welsh phraseology is Woolridge (2011), who investigated the use of the spell-checker tool Cysill to recognise English interference in Welsh texts. The author has conducted a study on a parallel
corpus in which two professional translators evaluated texts translated from English into
Welsh. It emerged that 62% of cases of interference occurred in the field of phraseology,
30% in syntax and 8% in vocabulary (2011: 129).
Some valuable observations regarding borrowings were also made by Robert
(2013) in her investigation of the implementation of language planning initiatives in the
Welsh-language media and in workplaces. The author places her research in the context
of ideologies of purism, concluding that they were rather prevalent in both examined dataset, in that borrowings from English appeared to be avoided in standard language.
The above literature overview demonstrates that lexical borrowings from English
into Welsh have not been given much attention until very recently. It has been suggested
by Parina (2010: 184) that there were ideological reasons behind this gap and the lack of
publications demonstrates deliberate avoidance of the subject as one politically incorrect.
It is characteristic that since the times of T.H. Parry-Williams nearly all the scholars publishing on the subject have come from outside Wales. Indeed, discourse on the English
element in Welsh does seem to be ideologically biased, as borrowings are frequently seen
as a sign of language deterioration. On the other hand, as shown above, more and more
arguments proposing the acceptance of the current language change can be heard in recent
years.
98
1.3.3. Ideological stances towards borrowing from English – an overview
Ideological stances towards the English element in Welsh have fluctuated over the course
of history, from broad acceptance of loanwords to linguistic purism. As T.H. Parry-Williams noted:
Welsh has no doubt been subject to the vagaries of linguistic fashion. There are times when
borrowing is popular and is being encouraged, and times when such a practice meets with
the condemnation of the literary leaders and scholars of the day. (Parry-Williams 1923: 15)
Several such stages may be distinguished in the history of Welsh. Naturally, before modern times, little is known about the spoken language, therefore one can only
discuss the linguistic norm inferred from the available written corpus. According to ParryWilliams (1923: 12 ff.), the first known borrowings from English are to be found in the
earliest examples of Welsh literature from the 8th and 9th century, although they might
be easily confused with borrowings from Old Norse. A substantial number of Anglicism
can be found in the medieval prose of the 13th - 15th centuries and in the late poetry of
Gogynfeirdd (court poets of the 13th century). The borrowings found at that time appear
to be stylistically marked and used primarily for satirical purposes. From the 14th century
onwards, the new generation of poets known as Cywyddwyr, the most famous of whom
was Dafydd ap Gwilym, began to diverge from the diction of their predecessors by seeking inspiration in the satirical language of the lower order of bards (Clerwyr), whose language was less formal and probably closer to the spoken idiom. Cywyddwyr’s poetry is
abundant in Anglicisms, which were, however, still used largely to the effect of satire or
caricature. There are even examples of writing cynghanedd in English dating to the 15th
century (Heinz 2010: 26).
These trends belong to the period when Welsh flourished and the threat of English
dominance was not significant. With diminishing prestige and position of Welsh after the
Acts of Union, one may observe the beginning of resistance to borrowings caused by the
strength and conservatism of the literary tradition (Hincks 2007: 16). Bards were supposed to keep the language pure from foreign influence and some authors were critical of
using too many loanwords in poetry (Parry-Williams 1923: 6). A fashionable theory
among scholars and writers such as Edward Lhwyd (1660-1790) and Theophilus Evans
(1693-1767) was that English borrowings are in fact ‘native’ Welsh words borrowed by
99
the English before the conquest of Wales. At the same time, the number of loanwords
from English continued to grow, especially in free metre poetry and prose, which suggests
that the spoken language was already permeated with borrowings. The difference between
the standards in spoken and written Welsh became more prominent in the 18th and especially the 19th century, when the climax of linguistic purism occurred. Scholars of the
time introduced a plethora of new, often bizarre Welsh coinages into literature and dictionaries in order to avoid Anglicisms at all cost. Strangely enough, this trend was accompanied by artificial Anglicisation of syntax and style. The gap between the highly sophisticated language of a learned purist and the speech of an ordinary man was widening
along with growing bilingualism and the decline in the number of Welsh speakers (Hincks
2007: 9). On the other hand, many of the coinages were accepted in everyday language
and are used until today (cf. Robert 2013: 91).
A change came at the beginning of the 20th century when a group of prominent
scholars, most notably John Morris-Jones and Emrys ap Iwan fiercely opposed the purist
tendencies of the previous century, aiming to unite the spoken and written Welsh, while
preserving its richness and variety (see 1.2.4.1). However, these scholars were not primarily concerned with lexical borrowings from English as they perceived the influence
on syntax as a much more serious threat. Older and established borrowings from English
were seen as acceptable as long as they were easily integrated into the Welsh phonetic
system and sounded natural (Hincks 2007: 22, 76).
Yet the efforts of the two generations of scholars were confronted with increasing
bilingualism, the growing prestige of English and negative attitudes towards Welsh. The
19th century saw a flood of English vocabulary into spoken Welsh; intellectuals of that
time complained about people littering their speech with English words they did not even
understand (Hincks 2007: 27). The word bratiaith, a derogatory term for the ‘debased,
adulterated’ Welsh abounding in Anglicisms was first noted in 1858. Yet, despite negative opinions on this kind of language, examples of books written in bratiaith can be
found towards the end of the 19th century34.
For example in Thomas Thomas’s Llyfr Coginio a Chadw Tŷ [The Book of Cooking and Housekeeping]
(Wrexham c. 1880) intended for uneducated women. (Hincks 2007: 61) This may suggest a gender difference in the usage of borrowings. Hincks mentions evidence for the tendency among Welsh women in the
19th century to reject the Welsh language perceiving it as inferior to English. It might have stemmed from
the fact that many Welsh women worked as maids for the anglicised gentry (Hincks 2007: 26).
34
100
By the second half of the 20th century when the Welsh language became truly
endangered, purist tendencies among scholars continued. As an example, one may mention publications such as Cywiriadur Cymraeg (Jones 1965), A guide to correct Welsh
(Jones 1976) or Cymraeg Idiomatig (Cule 1971) containing sections on avoiding “erroneous idioms” calqued from English.
At the same time, initiatives such as Cymraeg Byw were launched in an attempt
to bridge the gap between different registers of Welsh. It is worth pointing out, however,
that its creators viewed calquing from English negatively as well (cf. e.g. Jones 2011: 82–
84). Moreover, vernacular varieties of the spoken language rich in loanwords from English were embraced by novelists, such as Caradog Prichard in Un nos ola leuad (1961),
distancing the literary language further from the purism and formality of the previous
decades. This initiated the process of informalisation of the written and spoken standard.
However, with regard to the linguistic description of Welsh, an influential grammar by Stephen J. Williams (1980) was based on the traditional monolingual norm, modelled on the Welsh Bible and the standards proposed by John Morris-Jones. A major linguistic work of the 1990s, P.W. Thomas’s Gramadeg Y Gymraeg (1996), is also written
within the prescriptive approach. Thomas clearly states his principles in the introduction,
stressing the need to maintain “correct” Welsh, by which he means a language that develops in a “natural” way, i.e. without the influence of another language. He considers
structures which developed due to contact with English to be erroneous. Similarly to Williams, Thomas’s stance clearly derives from the heritage of John Morris-Jones, who is
often referred to in the introduction. Consequently, Thomas uses the same criteria that
were once applied by Morris-Jones: the authority of scholars and the set of conventions
that developed among ordinary speakers of the language. Yet the author has not ignored
the fact that reality had changed:
It is relatively safe in a monolingual society to appeal to speakers’ and users' habits while
interpreting the significance of linguistic variance. But in a bilingual society with no monolingual speakers, the number of varying features of the language may increase due to interference from the majority language 35. (Thomas 1996: 9)
“Cymharol ddiogel mewn cymdeithas uniaith yw apelio at arferion siaradwyr ac ysgrifenwyr yr iaith
wrth ddehongli arwyddocâd nodweddion ieithyddol amrywiol. Ond mewn cymdeithas ddwyieithog nad oes
ynddi siaradwyr uniaith, gall nodweddion ieithyddol amrywiol godi yn sgîl ymyrraeth o du’r iaith fwyafrifol.”
35
101
It can be seen that Thomas sees bilingualism as a danger for Welsh in that it introduces undesirable linguistic variations, yet he does not oppose lexical borrowings from
English as such; like Morris-Jones before, he is concerned about keeping the structure
pure:
The influence of English in Welsh is a fact. Not only today, but for centuries (…) However,
vocabulary is just one layer of the language; its structure is a completely different matter.
A language may borrow thousands of words from other languages and flourish; you need
not look further than English to see an extreme and very successful example of this. However, when one language influences the structure of another heavily, and the one affected
is an endangered minority language, you need consideration before giving a blessing to any
new borrowed structural feature.36 (Thomas 1996: 11–12)
Another development which triggered discussion on the English element in Welsh
was the advent of the Welsh media. Since its beginning, broadcast language was a subject
of complaints from purists among the public who felt English vocabulary was overused
(Ball et al. 1988: 184). The fiercest arguments appealed to nationalistic sentiment:
A recent development is the sick tendency to use English in news bulletins. (…) One needs
to ensure that the Welsh language used in Welsh news bulletins has the same status as
English in English news bulletins – it is a matter of national and linguistic pride and honour.
(Y Faner Newydd 1997 as cited in Ball et al. 1988: 184)
Such attitudes seem to have prevailed in Wales until the end of the 20th century. Also
nowadays the use of English on the media gives rise to hot debates. For instance, browsing through the articles on the Welsh media from the last 5 years on the online news portal
Golwg360, one can see that the majority of discussions on the media revolve around issues connected with the excessive presence of English. Presenters are often accused of
using bratiaith and overusing English terms. Robert summarises these issues as follows:
In its most intense articulations (…) Welsh standard language ideology is fervently monolingual and protectionist. And despite the difficulties of locating any particular form of
standard Welsh (…) the issue of contact phenomena seems to be one where there is little
“Mae dylanwad y Saesneg ar y Gymraeg yn ffaith. Nid heddiw’n unig ond ers canrifoedd (...) Ond un
wedd ar iaith yw ei eirfa; cwbl wahanol yw ei chystrawen. Gall iaith fenthyca miloedd o eiriau oddi wrth
ieithoedd eraill a ffynnu; does dim rhaid edrych ymhellach na’r Saesneg ei hun i weld enghraifft eithafol
a thra llwyddiannus o hynny. Eithr pan fydd un iaith yn dylanwadu’n drwm ar gystrawen iaith arall, a’r
iaith y dylanwedir arni yn un lleiafrifol dan fygythiad, yna mae’n rhaid ymbwyllo cyn rhoi sêl bendith ar
unrhyw nodwedd fenthyg newydd yn ei chystrawen”.
36
102
disagreement amongst those who concern themselves with language standards. (Robert
2011: 140)
Quite the opposite ideology of accepting language change has emerged as well, both in
academic and public discourse in view of poststructuralist approaches to multilingualism
(see 1.2.1). Crystal (1999) in his article “Is Welsh safe?” quite fiercely opposes linguistic
purism in Welsh as one that discourages speakers from using the language. He quotes an
anecdotal piece of evidence of public condemnation of the band Manic Street Preachers
who used an ‘incorrect’ i.e. calqued Welsh phrase in their advertising poster. Crystal believes that this kind of purism may have a detrimental effect on the less proficient speakers and “generate (…) an inferiority complex which further harms their motivation to
continue with the language. Purists, accordingly, are an endangered language’s worst enemy. (…) Instead of condemning people for failing to use the language more, they [the
users, ML] should be praised for whatever language they use” (Crystal 1999: 10).
Similar views are expressed by Robert (2009), who criticises purists’ attitudes for
inhibiting the accommodation of L2 speakers into Welsh speech communities, and who
postulates accepting the presence of English borrowings into the Welsh language, as purist tendencies might reduce the use of Welsh due to the low confidence of speakers. Some
academics have presented similar opinions in the media, for example Margaret Deuchar
in an interview for Golwg:
Criticising people for using English may discourage them from using Welsh entirely. (...)
It harms people when they are corrected all the time. They lose their confidence then and
we very quickly fall into the ‘my-Welsh-is-not-good-enough’ attitude. (....).37 (Gibbard
2010: 14)
The above stances refer more to the spoken than written language. However, changes in
attitudes towards ‘pure’ Welsh can be observed also with regard to writing even in the
most conservative environments, such as the National Eisteddfod. Traditionally, criteria
of language correctness and purity played an important role in evaluations of the works
of authors competing in various literary competitions of the festival. It suffices to quote
some comments from Emyr Llywelyn’s 2011 adjudication for Daniel Owen Memorial
„Mae beirniadu pobol am ddefnyddio Saesneg yn gallu eu troi nhw oddi wrth y Gymraeg yn gyfangwbl.
(..) Mae yn gwneud niwed i bobol i gael eu cywiro drwy’r amser. Mae pobol yn colli hyder wedyn a rydyn
ni’n syrthio’n gyflym iawn i’r agwedd ‘dy’n Nghymraeg i ddim yn ddigon da’ (...).”
37
103
Prize for the best novel. The following quotes justify the rejection of two of the novels
submitted for the competition:
One of the main weaknesses of this novel is a belief that using English dialogue in a Welsh
novel is acceptable. If a Frenchman or German appears in an English novel, he speaks
English. Although a character from another country speaks another language, it should be
Welsh. English dialogue turns a novel into a bilingual one and it ceases to be a Welsh
novel38. (Hughes ed. 2011: 115–116)
The novel is full of English phrases and idioms. Actually, it feels as if it was originally
written in English and then a literal translation was applied!39 (Hughes ed. 2011: 116)
These comments clearly demonstrate the traditional norm, according to which a bilingual
text “ceases to be Welsh”. It also criticises excessive translation from English. But this
standard seems to be at least partly contested by an increasing number of authors playing
with bilingualism in their works, including those writing in traditional strict metre. In
2014, a debate arose over the use of English in the winning poems by Ceri Wyn Jones
and Guto Dafydd during the National Eisteddfod (in strict metre and free metre respectively) due to the fact that they contained a large number of Anglicisms. The decision of
one of the judges to refuse to accept Wyn Jones as the winner of the Chair was criticised
in the press (Siencyn 2014) and the general reception of the poetry was positive. Although
the extent of changes in perception of borrowings from English in recent years is difficult
to establish, the above evidence seems to support the observation made by Robert (2013:
108) that there is a shift towards more tolerance to variable language practices.
1.4. Concluding remarks
This chapter has aimed to outline the current state of Welsh with a focus on the level of
standardisation and the extent of the English influence and establish theoretical background for the analytical chapters to follow. It has been shown that the future of Welsh is
“Un o brif wendidau’r nofel hon yw credu bod defnyddio deialog Saesneg mewn nofel Gymraeg yn
dderbyniol. Pan fydd Ffrancwr neu Almaenwyr yn ymddangos mewn nofel Saesneg, y mae’n siarad Saesneg. Er bod cymeriad o wlad arall ac yn siarad iaith arall, Cymraeg ddylai fod iaith deialog cymeriadau
mewn nofel Gymraeg. Mae deialog Saesneg yn troi nofel yn ddwyieithog ac mae’n peidio â bod yn nofel
Gymraeg.”
39
“Mae’r nofel yn llawn ymadroddion ac idiomau Saesneg. Yn wir, teimlwn wrth ei darllen ei bod wedi ei
hysgrifennu’n wreiddiol yn Saesneg, a bod cyfieithu llythrennol o’r Saesneg yn digwydd!”
38
104
insecure as revitalisation efforts are faced with a number of challenges related to demographic and social changes, not the least of which is the growing proportion of non-native
speakers. Bilingualism and intense contact with English inevitably affect the traditional
linguistic norm and standard language. This is accompanied by the contraction of domains responsible for the transmission of standard Welsh in the 20th century, such as
literature and religion. In view of the problems in implementing Welsh as the language
of administration and law, the main channel of spreading the linguistic norm in today’s
Wales are the media, who strive to appeal to the largest possible audience by seeking
a balance between formal and informal registers. Hence the emergence of new semi-formal varieties of Welsh, which appear to exhibit fewer and fewer features of traditional
H varieties.
On the other hand, ideologies of prescriptivism and purism are noticeable both in
corpus planning initiatives and in some academic works in the field of Welsh linguistics.
The general lack of publications on the influence of English on Welsh is an illustration of
a tendency to exclude transfer from English from the description of the language. The
weight of literary tradition and also the rising official status of Welsh play an important
role in maintaining the traditional monolingual norm, which is supported not only by
lexical planning, such as the production of dictionaries, but also by the educational system
and cultural initiatives, such as eisteddfods. Altogether, one can agree with Robert
(2013: 98) that although standard Welsh as a social construct appears to be much less
consolidated than standard English the idea of standardness seems to be quite strongly
embedded in speakers’ perception.
Studies in language contact between Welsh and English and other related disciplines have provided evidence that in view of the bilingualism of speakers the English
influence on Welsh is becoming ever more prominent not only in the field of lexicon, but
also morphology and structure. This phenomenon can be viewed from various perspectives and there is a disagreement as to how far the contact-induced changes in contemporary Welsh reach. According to a more extreme view, expressed e.g. by Asmus and Williams (2014: 54), the profoundness of change, including not only lexical borrowing but
also syntactic calquing, is a symptom of pidginisation and a sign of oncoming language
death. Other researchers, such as Phillips (2007) believe that although the modelling of
structures after English is considerable, it is not strong enough to speak of metatypy:
105
Welsh has not yet reached the stage of word-for-word intertranslatability with English
which could be called metatypy, but there is certainly widespread calquing, resulting in
ever-increasing similarity to English. (Phillips 2007: 178)
Similarly, Flohr (2013) concludes that:
[t]he fact is that the Celtic languages have changed in their essence over the centuries and
mainly in the direction of a steady assimilation to English. (…) Regarding phraseology, it
has to be said that Welsh (in some other texts more so than in the translation analysed here)
sometimes gives the impression of being a sort of hybrid.
The author proposes to talk about “structural interference based on either shift or continued and widespread bilingualism“, rather than metatypy (Flohr 2013: 108).
It cannot be denied that in the current sociolinguistic situation, new generations of
bilingual speakers will find it more and more challenging to speak ‘pure’ and ‘idiomatic’
Welsh as long as it means a language devoid of loanwords and calques from English. The
question remains how they will respond to the still existing gap between the language
they speak and the one presented to them as standard Welsh. Musk’s study of young
people’s attitudes seems to confirm that many young people perceive the tension between
the externally imposed monolingual norm characteristic for the written language and their
conviction that inserting English vocabulary while speaking Welsh is a natural thing
(Musk 2012: 664).
Presently, half a century will have passed since the disappearance of Welsh monoglots and this might be the time for monolingual norm to slowly but inevitably retreat.
A linguistic change, however, does not need to be a sign of deterioration. Perhaps it is too
early to establish the exact character of the change that the Welsh language is undergoing
by using labels such as ‘pidginisation’ or ‘hybridisation’. But it is certainly worth investigating and discussing the developments in view of the vulnerable status of Welsh. This
thesis, which analyses a single linguistic phenomenon related to language contact hopes
to contribute to this debate.
106
1.5. Methodological considerations for the present study
The background knowledge presented in this chapter has implications on the choice of
the focus areas of the dissertation and methodology applied in the subsequent analytical
chapters. Some of the major issues will be now shortly discussed.
The first issue concerns the investigation of written and spoken language. Bearing
in mind the discussion on the standardisation of Welsh presented in this chapter, priority
was given to written representations of the language, in particular semi-formal varieties
in the press and fiction; these varieties would also, however, include samples of spoken
language, be it in press interviews or fictional dialogues. A corpus of c. 1 million words
created for the analysis is presented and analysed in Chapter 3. Supplementing it with a
corpus of spoken Welsh proved to be a challenging task. The main corpus of spoken
Welsh currently available is the Bangor Siarad corpus (Deuchar et al. 2018,
http://www.bangortalk.org.uk) consisting of c. 460,000 word tokens. This corpus, however, consists of samples in only one register i.e. informal conversations between friends,
work colleagues, family members, etc., and is therefore inappropriate for examining the
linguistic norm and semi-formal varieties of Welsh. Suitable data could be obtained by
recording formal speech situations or broadcast language. The latter would be particularly
valuable as the broadcast media undoubtedly play a role in standardisation processes.
However, time and funding constraints (especially considering that Welsh-language television is unavailable outside of Wales) made it impossible to include broadcast language
in the present dissertation. Creating a second corpus of spoken language comparable in
size with the first one with transcription and annotations would require an immense
amount of time and was unfeasible within the time available for competing this thesis.
Other than these technical difficulties, problems with defining standard spoken language
and the level of its standardisation described in 1.2.5. were also considered, in particular
the presence of dialects and learners’ speech, classifying scripted talk etc. Media language
analysed in the thesis will be therefore narrowed to the press. It is hoped that future research in broadcast language in Welsh and the completion of Corpws Cenedlaethol Cymraeg Cyfoes, the National Corpus of Contemporary Welsh (http://www.corcencc.cymru),
which is currently being created, will provide a sounder basis for corpus studies on spoken
norms.
107
Nevertheless, a sample of literature and press language analysed in Chapter 3 is
hoped to provide sufficient amount of data to complete the basic linguistic description of
phrasal verbs, which is the main aim of the thesis, and examine their stylistic markedness
at least in the written language considering a largely monolingual mode in which such
texts are produced as a kind of litmus test for the standardisation of phrasal verbs.
Another question discussed in the thesis are different opinions on phrasal verbs
held by Welsh speakers, which are undoubtedly related to both their ideologies and attitudes (1.2.1). While an attitudinal study of phrasal verbs would be another interesting
direction of research it was not conducted due to a number of potential difficulties. Firstly,
as discussed above, the decision to centre the research around the linguistic norm and
narrow corpus study to prose and fiction necessarily shifted the focus to written language
which is untypical of studies in language attitudes. More difficulties would arise from the
nature of the investigated constructions. First of all, phrasal verbs are generally hard to
identify by speakers without specialist linguistic knowledge as a separate object of a positive or negative attitude, due to the multitude of forms and the fact that they are not
always identified as calques from English. According to Haspelmath (2009:47) “speakers
are not likely to be aware of their attitudes to borrowing, because they rarely have extensive knowledge about other sociolinguistic situations and other possible attitudes”. Therefore a typical attitudinal test, asking the subjects for an evaluation of a given feature on
an evaluation scale might be confusing for the participants and yield ambiguous results.
Other than that, major problems in measuring language attitudes which have been outlined in 1.2.1.1. had to be considered too. These include differences between covert and
overt attitudes, discrepancies between declared attitudes and behaviour and factors influencing speakers’’ responses which are difficult to control in direct measurements.
Bearing this in mind, rather than investigating thoughts and emotions of individuals, the present research is concerned with more general, “more constructed” (MyersScotton 2006:110) assessments embedded in standardisation-related ideologies and attempts to examine whether and how they might influence the status of phrasal verbs in
Welsh. This will be done through an analysis of metalinguistic discourse on phrasal verbs
in academic literature (Chapter 2), qualitative analysis of stylistic markedness in a written
corpus of press and fiction (Chapter 3), an analysis of metalinguistic discourse in teaching
materials and dictionaries (Chapter 4) and a study on acceptability of phrasal verbs among
professional speakers of Welsh (Chapter 5).The latter contains an acceptability judgement
108
test and semi-structured interviews touching on the perceived acceptability of borrowings
in general and phrasal verbs in particular.
It should be emphasised that the empirical part of the dissertation is largely exploratory due to the lack of previous studies on the subject. The limited scope of a single
thesis does not allow for an in-depth exploration of some important aspects of the issues
discussed above. However, the research is hoped to provide evidence which could be used
in the future studies on these subjects.
Having presented the sociolinguistic and theoretical background to the analysed
issues, the next chapter is going to define and describe phrasal verbs in Welsh on the basis
of the available literature as well as establish the scope of research in the chapters to
follow.
109
Chapter 2: Phrasal verbs in Welsh – cross-linguistic
background and synchronic description
Introduction
Verb-particle combinations, commonly known as phrasal verbs, appear to be one of the
most elusive notions in English linguistics. No single definition of a phrasal verb is universally agreed upon, and nearly all aspects of these multi-word items, such as their status,
classification, syntactic properties, and historical origins have been the subject of debates
among scholars. In Welsh linguistics, on the other hand, considerably little attention has
been paid to phrasal verbs due to the fact they have often been seen as a simple case of
calquing from English and consequently excluded from standard language (e.g. Thomas
1996, Asmus and Williams 2014).
This chapter provides the cross-linguistic background for researching the level of
standardisation of phrasal verbs in contemporary Welsh. After presenting the current state
of research on phrasal verbs in English and Welsh, a working definition of a phrasal verb
applicable to the present study is established and selected features of these constructions
relevant to the thesis are described. The next section places the phenomenon of calquing
phrasal verbs, which is central to the thesis, in a wider context of language contact, presenting evidence for this mechanism occurring between English and other languages, in
particular Celtic languages. This is followed by a review of literature on phrasal verbs in
Welsh, which identifies gaps in the linguistic description of these constructions and presents different ideological stances of scholars towards them. The final part of the chapter
sets the scope of the present study together with my own classification of phrasal verbs
110
used in the analytical chapters to follow, based on the reviewed literature, including parallel studies on Irish.
Although the study focuses on the synchronic aspect, it also contains some observations on historical development of phrasal verbs in Welsh. This subject, extremely interesting in itself, has not been investigated thus far, except for a short section in Rottet
(2005). This chapter represents, therefore, a foray into previously unchartered linguistic
territory and seeks to establish a base point from which further discussion among linguists
might ensue.
2.1. Phrasal verbs in English linguistics – an overview
The interest in phrasal verbs in English linguistics has been on the rise over the last one
hundred years, beginning with a pioneering monograph by Kennedy (1920) on the syntactic behaviour of what he named verb-adverb combinations and their history in the English language. Successive major works on phrasal verbs with focus on syntactic aspects
include Bolinger (1971), Sroka (1972) and Fraser (1976). Semantics of phrasal verbs were
studied by Lipka (1972) and Pelli (1976), both of whom use the term verb-particle constructions. Phrasal verbs have also been analysed within various grammar models: generative (mentioned by Chomsky 1965; Aarts 1989; Dehé 2002), cognitive (Lindner 1981;
Gries 1999; Rudzka-Ostyn 2003) and functional discourse grammar (O’Dowd 1998;
Keizer 2009).
Recent works on the history of phrasal verbs include Denison (1981), Hiltunen
(1983) and a study of multi-verb items in Early Modern English by Claridge (2000). Another comprehensive monograph on phrasal verbs with a focus on the historical aspects
of their development was written by Thim (2012), who also proposed a novel approach
of regarding these constructions as a phenomenon of periphrastic word formation.
Last but not least, since phrasal verbs are commonly viewed as one of the most
challenging elements of English vocabulary to be acquired by non-native speakers, they
have been increasingly studied in the area of language acquisition, teaching and pedagogical lexicography (e.g. Televnaja et al. 2004; Perdek 2010; Wierszycka 2014).
111
2.1.1. Definitions
Phrasal verb is an umbrella term for a range of constructions which, as stated in Keizer
(2009: 1187), “though superficially quite similar, form a heterogeneous set, exhibiting
subtle differences in meaning and formal behaviour”. As a complex phenomenon on the
verge of syntax, phraseology, morphology and word formation they present notorious
difficulties for defining and analysing. The ambiguity of the term phrasal verb is acknowledged for instance by dictionaries of linguistics by Crystal (2008) and Trask (1993). Nevertheless, both authors accept the term, defining phrasal verbs as lexical verbs consisting
of a sequence of a lexical element and one or more particles. A similar definition is provided by Quirk et al. (1985: 1150–1152) in their Comprehensive Grammar of the English
Language; however, they attribute the important notion of idiomaticity to it: “The meaning of the combination manifestly cannot be predicted from the meaning of verb and particle in isolation” (1985: 1152). In Longman Grammar of Spoken and Written English
Biber et al. (1999: 403) as well acknowledge the “relative idiomaticity” of phrasal verbs,
which in their approach belong to one of several subcategories of multi-word items (see
2.4.2).
Idiomaticity is also the main functional criterion for inclusion in general dictionaries of English and specific dictionaries of phrasal verbs (e.g. Sinclair ed. 1991; Courtney
ed. 2000; Rundell ed. 2007), which therefore narrow down the definition of a phrasal verb
to idiomatic constructions only. This seems to coincide with a common-sense notion of
what a phrasal verb is among ordinary users and learners of English.
On the other hand, some authors of grammars, notably Huddlestone and Pullum
(2000), reject the term phrasal verbs altogether, describing the constructions within
a wider context of phraseology. Such an approach is based primarily on syntactic criteria,
while the present study focuses on semantic and stylistic properties of verb-particle combinations in Welsh and the question of their standardness in relation to contact with English. As will be seen, the term phrasal verb is applied within the Welsh context to a number
of constructions with different syntactic and semantic properties. For that reason, the term
phrasal verb (henceforth PV) is used here in the most general sense as a combination of
verb and particle which is typically an adverb or preposition. These verb-particle constructions are studied here as single lexical units in the minds of speakers and focus is
placed on their lexical rather than syntactic properties. Similarly to the major works cited
112
above, this study will concentrate on idiomatic constructions, as it is their status that is
debatable in Welsh. However, some attention will be paid also to certain types of non- or
semi-idiomatic verb-particle combinations, since idiomaticity of PVs is a complex phenomenon which cannot be reduced to simple binary opposition (see 2.4.2).
2.1.2. Popular notions about phrasal verbs
Recent studies, such as Thim (2012), point to some notions which, despite being generalising and not necessarily true, emerge repeatedly both in academic and popular discourse on PVs. As these ideas appear also in the Welsh context, they require some attention. Three assumed aspects of PVs are of particular importance here: their informality,
replaceability by single words and idiosyncrasy, or special “Englishness”. These properties will be described in the subsections to follow.
2.1.2.1. Informality
The idea that phrasal verbs are stylistically marked and confined to colloquial registers
of English dates back to the age of prescriptivism in English linguistics. Some authors of
early studies on PVs claimed that they were a feature of the colloquial language “of the
uneducated” (Konishi 1958: 119) and as such should be avoided by speakers who want
to maintain high standards in their language. However, as Thim (2012: 215) convincingly
argues, there is no sound evidence of these constructions being associated with a particular register of the language before the 19th century, when prescriptive approaches gained
popularity.
Nevertheless, there is no doubt that a large number of commonly used PVs – for
instance cheer up, drop off, get along – refer to contexts which are inappropriate in a formal situation. For this reason one may encounter popular reference works, in particular
style guides and textbooks for academic writing, which forbid the use of PVs or recommend using single lexical items (e.g. Swales and Feak 2004: 18; Cory 2009: 14; Clark
and Pointon 2016: 317). University students around the world are often discouraged from
113
using PVs in formal language and advised to choose their Latin equivalents instead (Mendis 2010: 20–22).
On the other hand, numerous reference books of more general nature, such as dictionaries, offer a more liberal approach. For instance, the authors of the dictionary Macmillan Phrasal Verbs Plus draw attention to variety of registers where PVs are used:
Expert speakers use phrasal verbs in all kinds of contexts – not just in informal situations
such as conversations or e-mails, but quite often in formal and technical writing too. There
are many contexts where a PV is simply the best, most natural-sounding way of expressing
an idea, and so students should be encouraged to use them. (Rundell ed. 2007: vi)
Statements such as this are supported by evidence from corpus studies which have tried
to establish the distribution of PVs across registers and styles. The study quoted most
frequently in this context is Biber et al. (1999), who concluded that PVs occurred fairly
evenly in conversation, fiction and news. The only area where the constructions were
underrepresented was academic prose (1999: 408). There were some notable exceptions,
however, since expressions such as carry out (an analysis) or point out (“to conclude”)
were more common in academic writing than in other registers (Biber et al. 1999: 412).
Quirk et al. (1985: 1152) state that PVs are primarily informal. However, as noticed by
Thim (2012:42), some of the examples they provide, such go astray, reel back, touch
down (of a plane), or turn on (the light) cannot be seen as particularly colloquial. Dempsey et al. (2007) have shown that in computational studies PVs may significantly distinguish between both the spoken/written and formal/informal dimensions. On the other
hand, corpus studies (e.g. Campoy Cubillo 2002) have confirmed that PVs are widely
used in specialised texts.
All things considered, it can be claimed that although a great number of PVs in
English are colloquial in tone and occur more frequently in speech than in writing, they
are by no means restricted to colloquial language. As far as Welsh is concerned, there
have been no studies devoted to the register distribution of PVs and this issue will be
further discussed in the dissertation.
2.1.2.2. Replaceability by a single item
114
Another aspect that is frequently mentioned as a distinguishing feature of PVs is their
replaceability by a simple synonymous verb. As with the notion of informality discussed
above, this idea may derive from prescriptive approaches, which presented Latinbased expressions as superior to combinations of ‘native’ English words. Such beliefs
may evoke the condemnation of stranded prepositions in the 18th century (O’Dowd 1998:
7–8).
Yet even a superficial scrutiny of PVs shows that replaceability with a single item
is not achievable in many cases. Studies have demonstrated that this criterion is of little
usefulness for defining a PV (e.g. Bolinger 1971; Quirk et al. 1985: 1162; Thim 2012: 40).
Dictionaries for a general readership acknowledge the same idea:
It is often said that phrasal verbs tend to be rather ‘colloquial’ or ‘informal’ and more appropriate to spoken English than written, and even that it is better to avoid them and choose
single-word equivalents or synonyms instead. Yet in many cases phrasal verbs and their
synonyms have different ranges of use, meaning and collocation. Single-word synonyms
are often much more formal in style than phrasal verbs so that they seem out of place in
many contexts. (Sinclair ed. 1991: v)
The question of replaceability is, however, more complicated when it comes to
language contact, i.e. a situation when PVs are a borrowed novelty in the recipient language. In this case some speakers feel that the constructions should be avoided, having
their ‘proper’, native equivalents. This issue will be subsequently discussed in Chapters 4 and 5 and section 2.3 of this chapter.
2.1.2.3. Idiosyncrasy
There is no doubt that PVs occur in English in great numbers and, due to their idiomatic
properties, may be seen as a highly noticeable aspect of English grammar. In the early
20th century some researchers went as far as to claim that these constructions express the
“analytic genius" of the English nation, an approach which from today’s perspective contained an element of "cultural and linguistic chauvinism" (Bernstein 1974: 59, 63). Although such discourse has largely disappeared in the second half of the 20th century, the
idea of the idiosyncrasy of PVs lingers on.
115
The linguistic evidence, however, does not support the claim that PVs are unique
to English. It is a fact that particle verbs similar to English PVs are to be found in all
contemporary Germanic languages, including German, Dutch, Danish, Norwegian, Swedish, Faroese, Icelandic, West Frisian and Yiddish (Dehé 2002: 1; Thim 2012: 46). Although “syntactic properties of the English phrasal verbs are in essential aspects identical
to those of the particle verbs in the other Germanic languages” (Thim 2012:54), there are
major differences as regards word order and the position of particle. Yet what really distinguishes English from other Germanic languages is the immense number of idiomatic
constructions that are used. While both idiomatic and literal PVs occur in most Germanic
languages, in none of these are idiomatic constructions as widespread and productive as
in English. There is also evidence that PVs exists in some non-standard varieties of Romance languages but whether they are the outcome of contact with Germanic languages
or internal-language development is debatable (Treffers-Daller 2012: 61–62).
The idiomaticity of English verb-particle constructions makes them particularly
difficult to master by second language users. Studies confirm that PVs are one of the most
challenging aspects of learning English vocabulary and developing native-like proficiency in both spoken and written language (cf. e.g. Darwin and Grey 1999; Chen 2007;
Siyanova and Schmitt 2007). With the dominant position of English as a global language
and the ever growing number of learners, the conviction that PVs in English are something special and idiosyncratic has good a chance of spreading.
As discussed in the Introduction and as will be subsequently shown, in the Welsh
context PVs are associated with borrowing and calquing from English by commentators
who adopt more prescriptive approaches to the language. This conviction may be occasionally seen also in the academic literature of the subject, which will be reviewed in 2.3.
First, however, it is worth describing the phenomenon of calquing verb-particle constructions and the presence of idiomatic PVs in languages other than Welsh.
116
2.2. Phrasal verbs in language contact situations – evidence from other languages
2.2.1. Phrasal verbs in other languages in contact with English
Welsh is certainly not the only language capable of calquing verb-particle constructions
from English. Literature presents examples of borrowing PVs in a situation of intense
language contact with English in various languages, including other Germanic languages,
for instance Danish (Gottlieb 2004: 46, 2012: 192) and those in which PVs do not usually
occur. One example of the latter is Spanish, which normally does not use verb-particle
constructions; in translating from English they would be rendered by compound verbs
with prefixed roots and other translation techniques (Bernstein 1974). However, it has
been observed that Hispanic communities in the United States and other bilingual communities in intense contact with English, e.g. in South America and the Caribbean, use
the phrase para atrás ‘back’ to form constructions analogical to PVs in English (Lipski
1990: 90–91; Potowski 2010: 39). Similarly, varieties of spoken French spoken in North
America, such as Cajun and Acadian French, calque certain types of English PVs, particularly those which contain the particle back. Yet another phenomenon is applying French
morphology to the English verb while leaving the English adverbial, thus creating hybrid
constructions such as messer up ‘mess up’. (Rottet 2000: 112-113). Calquing of PVs has
been also noted in Ontario French (cf. Treffers-Daller 2012: 63) and Jèrriais, the Norman
French dialect spoken on Jersey (M. C. Jones 2006: 562).
Studies in second language learning have demonstrated that acquiring English idiomatic PVs is easier to speakers of those languages which already have verb-particle
constructions. For instance, Laufer and Eliasson’s (1993) study showed that learners
whose native language did not have idiomatic PVs (Hebrew) avoided these constructions,
while learners whose L1 had idiomatic PVs (Swedish) did not. Weibel (2007) compared
essays of Italian and German students and showed that while Italians used fewer PVs, the
Germans actually overused them. Analogically to the learning process, calquing is expected to be facilitated in languages which already contain PVs and may use existing
patterns and productive particles to create new constructions. Such is, for example, the
case of American Norwegian (Haugen 1969 [1953]; as cited by Rottet 2000: 111). The
phenomenon does not seem to occur in languages which use entirely different patterns to
117
express motion and location, such as Slavic languages, where prefixes are used as aspectualisers.
2.2.2. Phrasal verbs in other Celtic languages
For particular relevance to this thesis is the question of the existence of PVs in Celtic
languages as it may provide some valuable insights into the nature of these constructions
in Welsh. Rottet points to the fact that transparent PVs expressing spatial relations such
as go up, walk down, put out, etc. have been undoubtedly present in all Celtic languages
for centuries; they are well-attested in Middle Welsh, dating back to the Black Book of
Carmarthen, dated to c. 1200 (Rottet 2005: 44). Consequently, controversies over the status of PVs concern solely idiomatic constructions, which are evidently less frequent than
transparent combinations and therefore assumed to be borrowed. Two other major Celtic
languages to be considered in this context are Irish, which has been in intense contact
with English for centuries, and Breton, which has remained under the influence of French,
a language without PVs. While PVs in Irish have been studied in several papers (Stenson 1997; Doyle 2001; Veselinović 2006), no relevant works on Breton have been found.
The question of Breton is, however, raised by Rottet (2005).
Contemporary Irish contains a fairly large number of PVs and, analogically to
Welsh, many of them have their exact English counterparts. However, there is also a number of PVs without their English analogues, which is also true for Welsh (see 2.4.4).
From the syntactic point of view, Irish has three types of idiomatic phrasal verbs,
all of which appear also in Welsh and will be described in detail in 2.4.3:
a) idiomatic prepositional verbs – e.g. cuir le, lit. ‘put with’, ‘add to something’,
déan ar, lit. ‘do on’, ‘move toward’, déan as, lit. ‘do from’, ‘get away, escape’, tabhair as,
lit. ‘give from’, ‘spirit away’, tar as, lit. ‘come from’, ‘shrink, expand’ (Stenson
1997: 567);
b) phrasal verbs proper – e.g. cuir amach, lit. ‘put out’, ‘spit out, vomit; report’,
cuir síos, lit. ‘put down’, ‘describe’, including semi-idiomatic, sometimes pleonastic constructions such as fás aníos /fás suas ‘to grow up.’ (Stenson 1997: 562; Veselinović 2006:
183)
118
c) phrasal prepositional verbs, such as cuir isteach ar, lit. ‘put in on’, ‘interrupt,
disturb’ cuir suas le ‘put up with’ (Stenson 1997: 567).
As regards idiomaticity, Irish contains the whole spectrum of verb-particle combinations from transparent to completely idiomatic (Stenson 1997: 562). From the point
of view of language contact, it has been observed that Irish has a number of idiomatic
PVs, which appear to be native. Some of them are well-attested in older stages of the
language, for instance cuirid suas de ‘give up, renounce, repudiate’ (Veselinović 2006:
181). Certain Irish verbs, such as bain do not have a one-to-one English equivalent and
form PVs which do not have English counterparts (Veselinović 2006: 180). Other types
of expression, called here loan renditions (see 2.4.4), bear some resemblance to English
but are not exact equivalents, for example leag suas, lit. ‘lay up’, ‘knock up, make pregnant’ (Stenson 1997: 564). The third type are constructions such as obair amach ‘work
out’ which may be considered to be calqued from English, although, as Veselinović
(2006: 181) claims, it does not necessarily need to be the case.40 Irish PVs also include
loanblends, where a) the verb is a borrowing with a native -áil ending, parallel to Welsh
-(i)o, for example hintáil thart ‘hinting around’, b) where the particle is borrowed, for
example tabhair away ‘give away’ (Stenson 1997: 559) or c) both elements are borrowings, as in switcháil off ‘switch off’. This shows the productivity of PVs in Irish in the
situation of contact with English. Stenson sees it as a natural process of enhancing the
language’s vocabulary: “[m]ost borrowed particles fill lexical gaps or provide specialized
meanings distinct from similar native particles in ways typical of languages in contact
throughout the world” (Stenson 1997: 576).
Both Stenson and Veselinović claim that PVs are well-rooted in the grammatical
system of the Irish language and many of them might have grown independently from
English. From the point of view of syntax, they follow Irish rather than English patterns,
while the patterns for creating idioms of that type date to the earliest stages of the Irish
language, offering a framework for introducing English borrowings (Stenson 1997: 560,
575). Veselinović also suggests that contact with English does not have to be the only
source of loans: “the striking similarities between certain idioms probably have to be
traced back to contact, but others can be products of an independent but typologically
40
For a longer list of suspected calqued PVs in Irish see Veselinović (2006:181-182).
119
parallel development” (2006: 181). She shows that English and Irish have undergone similar development in that preverbal compounds were replaced with periphrastic constructions, including PVs (Veselinović 2006: 174). Moreover, Irish English has its own peculiar PVs borrowed from Irish, which suggests bidirectionality of contact.
Stenson conducted a study of Irish texts from the 16th–20th centuries and demonstrated that the use of both literal and idiomatic uses of directional adverbs increased
steadily throughout that period, with an explosion of new constructions in the 19th and
20th centuries (Stenson 1997: 568), which can be attributed to increased language contact.
The distribution of PVs in Irish across registers has not been well investigated.
Veselinović claims that no such relation is to be observed as “it seems as if the particle
verbs entered the formal language earlier” (2006: 188). Stenson, however, mentions the
prevalence of PVs in modern texts which are more colloquial in character (1997: 570),
which would be a usual tendency. Although no separate studies on attitudes of speakers
towards such construction have been found, Stenson uses the phrase “sociolinguistic stigmatization of the forms” (Stenson 1997: 576) suggesting that they might be despised by
some speakers as calques from English. Veselinović suggests that such constructions are
evidently becoming increasingly popular in the language of new speakers but she does
not offer any tentative observations of the attitude of native speakers towards them
(Veselinović 2006: 183).
As regards PVs in the Breton language, they have been briefly investigated by
Rottet (2005: 45-49). He observes that literal verb-adverb combinations occur naturally
in the language, similarly to other Celtic languages. However, his comparative analysis
of six chapters of Welsh and Breton Bibles showed that the number of literal PVs in
Breton was much lower. Literal PVs were also found in Cornish, which demonstrates
their common Brythonic origin.
When it comes to idiomatic constructions, it has been claimed that they do not
exist in Breton (Stenson 1997: 560). Rottet, however, quotes a number of examples found
in a dictionary of contemporary Breton. These include both PVs proper such as dont
‘barzh ‘to regain consciousness, recover after and illness’ (lit. ‘come in’) and idiomatic
prepositional verbs, such as ober war dro (unnan bennak) ‘to look after someone’ (lit. ‘to
do around sb.’). Although the number of idiomatic PVs in Breton appears to be considerably smaller, according to Rottet it proves the potential of Celtic languages to form noncompositional PVs regardless of English influence.
120
The available literature on calquing PVs from English shows that the extent of the
phenomenon is influenced by the intensity of contact and the existence of syntactic frameworks in recipient languages which facilitate influx of such combinations. Studies on PVs
in Celtic languages demonstrate that literal verb-particle constructions have been present
in them since the earliest written attestations. The popularity of idiomatic constructions,
however, appears to be associated with contact with English. On the other hand, Celtic
languages have been found to be capable of forming native idiomatic PVs which allow to
form new constructions on the English model. All in all, it can be concluded that the
presence of idiomatic PVs in Celtic languages is not always a simple case of calquing.
The data on the acceptability and integration of such constructions across registers are
very scarce.
Bearing all this in mind, the present study will proceed to investigate verb-particle
constructions in Welsh, beginning with a literature review on the subject, which will be
followed by my own observations derived from corpus analysis in Chapter 3.
2.3. Phrasal verbs in Welsh linguistics
This section presents an overview of research on PVs in Welsh linguistics from the beginning of the 20th century until today. It aims to identify major gaps in the description
of phrasal verbs and look at the metalinguistic discourse on these constructions from the
point of language ideologies manifested by scholars. Generally speaking, the body of literature is fairly small as it mostly consists of side notes in papers on other topics.
2.3.1. Fowkes 1945
One of the first studies which mentioned PVs in Welsh is Fowkes’s 1945 paper on the
calquing of English idioms. Among 46 examples analysed by the author, there are 10
PVs: darllen i mewn ‘read into’, gwneud i fyny ‘make up’, pigo allan ‘pick out’, torri
i lawr ‘break down’, dwyn i fyny ‘bring up’, edrych ar ôl ‘look after’, rhoi (clociau) yn ôl
‘set (clocks) back’, troi allan ‘turn out’, rhoi i fyny ‘give up’, cau allan ‘shut out’, and
three idioms containing phrasal verbs: mynd allan o ffasiwn ‘go out of fashion’, gwneud
121
dy feddwl i fyny ‘make up your mind’ and dod ato dy hunan ‘come to yourself’ (Fowkes
1945: 244–247).
In Fowkes’s view, the idiomaticity of all these constructions is something “entirely foreign” to the Welsh language. He claims that particles such as i mewn and i fyny
normally have only a literal spatial or temporal signification. Using the particles in a metaphorical way is “from a Celtic point of view, sheer nonsense. It is only as an idiomatic
borrowing that any meaning can be seen in it” (Fowkes 1945: 244).
Although Fowkes points out that borrowing from English has a long tradition in
Welsh, there are no PVs among the historical examples he cites. Yet, although the author
describes calquing of PVs as “sheer nonsense”, his final general conclusions are much
more positive as the author acknowledges that borrowings generally enrich the language’s
vocabulary:
It seems, however, that there can be no objection to this linguistic course of events beyond
that of exaggerated sentimentalism. It is, perhaps, unfortunate that Welshmen are found
deploring the "corruption" of their language, although this is by no means limited to the
Welsh, yet even non-scientific opinion will scarcely deplore the rich results seen as a consequence of the variety of sources which have fed the vocabulary of English. (Fowkes 1945:
248)
Although the author seems to distance himself from ideologies of purism, he presents the rather negative view on idiomatic borrowings in Welsh as “illogical” interference from English. It will be seen in the subsequent analysis how the examples quoted in
this early study re-emerge in contemporary metalinguistic discourse.
2.3.2. M. Jones (1979) and A. Thomas (1987)
Between Fowkes’s 1945 article and the 1990s considerably little was written on phrasal
verbs in Welsh. They are mentioned in two articles on the standardisation of contemporary Welsh as an example of English influence by M. Jones (1979) and A. Thomas (1987).
Both authors generally place phrasal verbs outside the linguistic norm.
In his article on the standardisation of spoken Welsh, Jones (1979:113) refers to
PVs as one of the areas in syntax where English influence is noticeable. The author mentions adverbial constructions where the particle is redundant, such as syrthio i lawr ‘fall
122
down’ and marw allan for ‘die out’. He provided examples of idiomatic calques as well
(gwneud i fyny ‘make up’, rhoi i fewn ‘give in’, torri i lawr ‘break down’) describing
them as “even more striking”. The author draws attention to similar, transparent and semiidiomatic constructions, which already exist in Welsh, for instance mynd yn ôl, ‘go back’,
mynd ymlaen ‘go on’, ‘continue’, concluding that the new development is an extension
of the existing pattern.
Similarly, in a discussion on the standard spoken language, A. Thomas (1987:107)
mentions PVs as a phenomenon which attracts the most criticism from language purists
due to the fact that many of such expressions have indigenous Welsh equivalents. At the
same time, the author considers PVs to be one of the most productive sources of extending
the lexicon of Welsh and concludes that: “[t]here is no question but that this is an area of
productivity which the purist has no choice but to accommodate”.
2.3.3. P.W. Thomas 1996
Phrasal verbs are not included in major grammars of Welsh available today (Williams
1980; King 1993; Thorne 2000) with the exception of Gramadeg y Gymraeg by P.W.
Thomas (1996), the most comprehensive grammatical study of Welsh published so far.
The author treats PVs as a marginal issue, placing them in a footnote section to the
chapter on prepositional verbs. Prepositional verbs, berfau arddodiadol, are understood
by the author as combinations of verb + preposition, such as gofalu am ‘take care of’,
angofio am ‘forget about’, where the relation between the two lexical elements is particularly close (Thomas 1996: 560). The note on PVs, berfau ymadroddol, opens with the
statement: “Welsh has a number of phrasal verbs, that is verbs often accompanied with
a particular adverbial41” (Thomas 1996: 560). The definition is followed by a categorisation of PVs in Welsh into three groups (Thomas 1996:561):
I.
berfau cyflawn, ‘intransitive verbs’. According to the author, this category includes intransitive sequences of verb + adverbial that do not have single-word
41
Mae gan y Gymraeg nifer o ferfau ymadroddol, sef berfau y bydd adferfolion penodol yn cyd-ddigwydd
â hwy’n aml.
123
equivalents and it is unusual to separate their elements. The examples provided are mynd i ffwrdd and mynd ymaith, both meaning ‘go away’.
II.
'berfau cyflawn ac anghyflawn, ‘intransitive and transitive verbs’. These include transitive and intransitive verbs that tend to be used only in the informal
style and may be replaced with a single verb. The verb and particle can be
separated with an object. Additionally, in a very formal, “if not outdated” style
the two elements can be separated by an adverbial (see 2.4.2). The following
examples are given:
mynd yn ôl ‘go back’, dod yn ôl ‘come back’, rhoi yn ôl ‘give back’,
all of which may be replaced by dychwelyd ‘return’;
cymryd yn ôl ‘take back’ to be replaced by atafaelu ‘distrain’ or
dychwelyd ‘return’;
dod ymlaen ‘come on’ to be replaced by datblygu ‘develop’ or cynyddu
‘increase’.
III.
The third category are sequences translated from English, which may include
some verbs from the second group. The examples provided are: llenwi allan
‘fill out’, codi i fyny ‘rise up’, torri i ffwrdd ‘break away’, eistedd i lawr ‘sit
down’, syrthio i lawr ‘fall down’, ysgrifennu i lawr ‘write down’. According
to Thomas, these verbs also belong to informal registers, particularly the spoken language, as in the formal language the particles would be omitted: “[o]ne
would not expect, for example, the Archdruid to ask the awarded bard to ‘sit
down in the peace of the Eisteddfod’”42 (Thomas 1996:561). These verbs can
also be complemented with an object inserted between the verb and particle.
Thomas’s categorisation described above raises a number of issues concerning
both the criteria for the division and the description of particular categories. First of all,
while PVs in Categories 1 and 2 are divided on the basis of a semantic criterion of replaceability by a single-word equivalent, verbs in Category 3 are distinguished by another
criterion i.e. their alleged English origin. The latter distinction seems to be troublesome
for the author himself, as acknowledged in his comment to Category 3: “it is possible that
“Ni ddisgwylid, er enghraifft, i’r Archdderwydd wahodd y bardd arobryn i ‘eistedd i lawr yn hedd yr
Eisteddfod’”
42
124
some of the above [i.e. Categories 1 and 2 – ML] can be attributed to the same process”43.
Thus the distinction between groups 2 and 3 is blurred and based solely on the linguist’s
guess whether an item was borrowed or not.
The criterion distinguishing between 1 and 2 i.e. the existence of one-word equivalents is questionable as well. To begin with Category 1, it can be argued that the examples provided, mynd ymaith and mynd i ffwrdd, could in fact be replaced with a single
verb, such as gadael ‘leave’. Indeed, GPC gives mynd ymaith as a definition/synonym of
gadael. The semantic difference between gadael and mynd ymaith does not seem to be
greater than between dychwelyd ‘return’ and dod yn ôl ‘come back’ attributed to Category 2.
Furthermore, it is arguable whether PVs in Category 2 should be labelled “informal”. For instance, a corpus search shows that mynd yn ôl and dod yn ôl are used extensively in formal contexts, e.g. in the Welsh National Assembly44 alongside dychwelyd,
and therefore might be, at best, described as neutral. At the same time, one of the proposed
single-word equivalents atafaelu is obsolete and belongs to a highly literary, fossilised
style45. It should also be noted that despite the statement that these PVs occur only informally, Thomas shows examples of the same verbs used in very formal, “if not outdated”
contexts. This obvious contradiction supports the claim that the constructions in question
are not confined to informal style. On the whole, since items in both Categories 1 and 2 may be replaced by other items and are not confined to a particular register,
the proposed categorisation based on replaceability and formality criteria is not valid.
As regards category 3, it is worth noticing that the majority of examples provided
belong to the type called here pleonastic PVs (see 2.4.4), that is constructions where the
second element may be seen as redundant due to the fact that its meaning is included in
the verb. Among the relevant examples given by Thomas are codi i fyny ‘rise up’, eistedd
i lawr ‘sit down’, syrthio i lawr ‘fall down’, ysgrifennu i lawr ‘write down’. The status of
these verbs is not given explicitly, but the author’s tongue-in-cheek comment that they
would not be used at the National Eisteddfod suggests their nonstandardness.
“Y mae’n bosib bod rhai o’r uchod i’w priodoli i’r un broses”
In the CCCC corpus the number of concordances for dychwelyd is 167, for dod yn ôl – 145, mynd yn ôl
– 131.
45
In the CEG and CCCC corpora atafaelu is used only in the sense ‘confiscate’ or ‘sequestrate (carbon)’,
never in the sense ‘get back’.
43
44
125
Quite surprisingly, except from dod ymlaen ‘develop’, ‘increase’, the footnote
does not mention any truly idiomatic PVs, which would be very unusual in any discussion
on the subject in English-language context. This suggests that in Thomas’s view idiomatic
PVs are outside the scope of description of standard Welsh. However, even if one accepts
this approach and limits the occurrence of PVs in Welsh to compositional constructions,
Thomas’s description is not devoid of faults, lacking clear and non-arbitrary criteria for
the proposed classification.
As described in 1.3.3., P.W. Thomas’s book was written within a prescriptive approach, following the tradition of John Morris-Jones. Since the author believes that Welsh
structures which emerged due to contact with English are erroneous, he treats phrasal
verbs as nonstandard and restricted to colloquial registers. In doing so, he appears to accept all three popular yet oversimplifying ideas concerning PVs mentioned in 2.1.2.: their
replaceability, informality and “Englishness”. As a result, the description of phrasal verbs
in a major textbook for anyone studying Welsh grammar is fairly inconsistent and nearly
ignores the existence of all idiomatic PVs. One of the primary aims of the present thesis
is to provide a more comprehensive description of Welsh phrasal verbs, problematizing
their status across different registers of the language.
2.3.4. Heinz 2003 and Asmus and Williams 2014
Phrasal verbs are also mentioned by Heinz (2003: 57-59) in her study of Welsh lexicography while describing the semantic aspects of the Welsh verb, defined as “verbs which
take adverbials as their complement”. The author refers to the section of P.W. Thomas
described above and concludes that “there is no reliable definition for phrasal verbs in the
Welsh language and their existence is arguable” since “by some academics they would
be recognised as being wrong in Welsh” (Heinz 2003:57). The author associates PVs in
Welsh with contact with English, mentioning “a tendency to introduce English structures
into the Welsh language”.
In her semantic classification of the Welsh verb, Heinz distinguishes four semantic categories. PVs constitute the fourth one, which however, as the author claims, is dubious for inclusion since, firstly their very existence is questionable, secondly, they can
126
mostly be replaced by other Welsh verbs, and thirdly they behave differently than the
other three categories as regards syntax.
Similarly to Thomas, Heinz suggests that PVs should be excluded from the description of standard Welsh as a marginal issue. The reason for exclusion is negative attitudes of Welsh scholars and an underlying claim that they are novel creations, as a rule
replaceable by already existing Welsh verbs. Interestingly, the definition of a PV provided
does not refer to the question of idiomaticity, implicating that even transparent constructions such as cymryd yn ôl ‘take back’ provided as an example are non-existent or nonstandard.
The reservations regarding the standardness of PVs are expressed more explicitly
in a recent paper co-authored by the same researcher (under the name of Asmus). It is
stated that the linguistic description of Welsh is faulty due to the fact that calques from
English are favoured over native idiomatic verbs. The only example of a PV given is
pleonastic codi lan: “[t]here is a focus on phrasal verbs (e.g. codi (lan) “get up (up)”,
which do not really form an essential part of Welsh. (…) Indeed, many of them are loan
translations from English” (Asmus and Williams 2014: 53).
The view represented in Heinz (2003) and Asmus and Williams (2014) appears to
be ideologically rather than linguistically based, as the authors imply that borrowed constructions are not a part of the Welsh language and therefore should not be included in
linguistic descriptions. Lack of a clear definition of a PV or distinction between transparent and idiomatic make this view appear somewhat simplistic. It can be seen as an example of rather purist attitude to language influenced by the prescriptive representation of
PVs in Thomas’s grammar.
2.3.5. Rottet 2000 and 2005
The only study devoted solely to phrasal verbs in Welsh is an article by Rottet (2005),
preceded by a more general paper by the same author on the calquing of PVs in language
contact (2000), which also mentions Welsh. The author uses McArthur’s (1992) definition of a phrasal verb which states that it is a verb which “operates more like a phrase
than a word” and comprises of a verb of movement and action and an adverbial particle
127
of direction and location. This definition narrows the scope of research to adverbial constructions only, excluding prepositional and phrasal-prepositional verbs.
Rottet draws attention to the fact that PVs are rarely acknowledged in descriptive
and pedagogical literature and that there are reasons to see them as a phenomenon of
language contact, as most Welsh PVs have exact English counterparts (2005: 40–41).
However, in the second section of his article, the author challenges this view by pointing
to the long history of PVs in Welsh and their presence in other Celtic languages (2005:
42-53), which has been discussed in 2.2.2. He also provides a description of Welsh adverbials, drawing attention to the fact that they are grammaticalised prepositional phrases.
In the third section, Rottet proceeds to establish a typology of Welsh PVs as contact phenomena, distinguishing between pleonastic calquing of particle, calquing of both
verb and particle, loanblends and wholesale borrowings (2005: 53-60). These categories
will be described in more detail in 2.4.4. as they prove to be very useful when discussing
the complexity of the phenomenon.
Importantly, the author also notices the existence of a few Welsh idiomatic PVs
which do not have their exact English counterparts, such as bwrw ymlaen ‘carry on’,
lit. ‘hit on’ and cael allan ‘find out’, lit. ‘get out’. This, combined with the evidence from
other Celtic languages, leads the author to conclude that Welsh has the capacity of creating idiomatic PVs on its own, which is immensely reinforced by the influence of English.
This is in accordance with Stenson’s (1997) and Veselinović’s (2006) observations regarding Irish (2.2.2). However, Rottet’s claims concerning the origins of ‘native’ constructions are largely intuitive and not supported by historical evidence (see 3.3.)
The fourth and final section of the article discusses what the author calls ‘attitudes’
to Welsh phrasal verbs. In this section the author presents a number of examples from
metalingustic discourse and quotes some remarks on the constructions made by language
specialists. Next, he analyses a number of dictionary entries from Geiriadur Prifysgol
Cymru and Geiriadur yr Academi to search for evidence for a pedagogical norm against
PVs.
Rottet is the first author who problematizes the use of phrasal verbs in different
registers. In the first of his two studies (2000), he refers to the replaceability criterion, but
notices that PVs are widely accepted by speakers of Welsh:
While most if not all of these phrasal verbs could also be expressed with single word Welsh
equivalents, there is little or no stigma attached to using the loan translation variants in
128
Welsh today. Many such calques are well enough established that they are no longer recognised as tokens of English influence. (Rottet 2000: 115)
In the same paragraph the author also draws attention to differences between the official
description of the language and actual usage:
However, not all attested combinations are officially acknowledged in dictionaries (…)
Some contemporary Welsh speakers freely calque even the most idiomatic English phrasal
verbs spontaneously during conversation, and since all speakers today are also fluent in
English, there is little or no danger of not being understood. (Rottet 2000: 115)
In contrast to previous studies, Rottet does not make evaluative judgments, following, as
he does, a descriptive approach. Having provided numerous examples from the press and
literature, he concludes that idiomatic PVs are a natural feature of contemporary Welsh,
largely integrated in the language of bilingual speakers, although there seems to be a prescriptive norm against them, reflected in dictionaries and pedagogical materials.
As a pioneering study, Rottet’s article provides a number of important insights
into the nature of phrasal verbs in Welsh. Importantly, it is the first study which considers
the idiomaticity of phrasal verbs while discussing their status as borrowings from English
and offers some information about the evolution of these constructions throughout history. Naturally due to its small scope, Rottet’s study has also a number of limitations. The
evidence provided by the author is based on examples taken from a small corpus of texts
which is not described in detail by the author. Upon closer scrutiny one can see the quoted
examples come from two weeklies Y Cymro (19 issues from the years 2000-2003) and
Golwg (two issues 2000 and 2001) and around fifteen literary works, mostly novels, written between 1885 and 2000.The presentation of literary examples is slightly misleading
with regard to their chronology as the author provides dates of reprints of some novels.
Therefore, for example, by citing Daniel Owen as “Owen 1993” he leads the reader to
think that the quotation of this 19th-century author comes from the 1990s. While this does
not impede the validity of Rottet’s general claims, it leaves room for a more systematic,
synchronic study of phrasal verbs in contemporary Welsh. The author also does not give
much consideration to differences in acceptability across various registers of Welsh but
merely acknowledges their presence in the informal as well as the written language. Furthermore, in describing the “pedagogical” norm he analyses mainly dictionaries and only
one strictly pedagogical contemporary book (i.e. Gruffudd 2000). Lastly, without giving
129
wider consideration to the complexities of defining a phrasal verb, Rottet narrows the
scope of research only to constructions with adverbial particles.
2.3.6. Hirata 2012
A contribution to the research on syntactic behaviour of transitive PVs in Welsh has been
made by Hirata (2012) in his study on preposition stranding in Welsh. He demonstrates
that preposition stranding is possible in colloquial Welsh and suggests that it is a recent
innovation caused by contact with English. The thesis touches on other issues related to
language contact. Following Borsley et al. (2007: 116), the author believes that the appearance of preposition stranding is a twentieth-century innovation modelled on English
syntax. Although PVs are not the main subject of Hirata’s study, he pays some attention
to particle placement in verb-particle constructions. The study involved twelve Welsh
native speakers who were given sets of alternative sentences and marked their acceptability on a five-point scale. Six of the 72 sets contained transitive PVs: troi i lawr, sortio
allan (used in two sets), troi ymlaen, troi i ffwrdd, torri i lawr.
The results have demonstrated the acceptability of optional particle placement. It
was shown that PVs in Welsh prefer the Verb – Object – Particle order. This order is most
acceptable when the PV contains a complex particle, while a Verb – Particle – Object
order is more likely to be accepted with single particles. This phenomenon is explained
by “the assumption that a single particle is easier to merge in morphology to make a
complex verb” (Hirata 2012: 166). Moreover, and importantly for the present study,
Hirata’s results suggest that PVs were less accepted than their one-word equivalents available to participants in the sentence sets – gwrthod for ‘troi i lawr ‘turn down, reject’,
diffodd for troi i ffwrdd ‘turn off’, torri for torri i lawr ‘break down’ and trefnu for sortio
allan ‘sort out’ (Hirata 2012: 162–163). Importantly, the fifth studied phrasal verb
rhoi/troi ymlaen ‘turn on’, does not have a non-phrasal equivalent.
Without referring to PVs in particular, the author suggests that acceptability of
contact-induced constructions may be associated with the participants’ exposure to the
literary language and also their age:
In my judgement tests, there are some informants consistently disallow [sic] P-stranding
sentences and some prefers [sic] them. The amount of exposure to the literary language
130
seems to be related to the acceptability. However, some other factors such as age might be
more relevant (Hirata 2012: 193).
Although the study does not focus on phrasal verbs, it offers some evidence supporting the above claims of nonstandardness of certain PVs among native speakers of
Welsh.
2.3.7. Other studies
Occasional mention of PVs is to be found in a small number of papers, most of which
point to the phenomenon of calquing verb-particle constructions from English. Jones
(1998) provides examples of PVs in her study of the Rhosllannerchrugog and
Gwenhwyseg dialects. She places them among other constructions calqued from English
and proposes their one-word equivalents:
in Gwenhwyseg dialect: diffodd ‘extinguish’ for troi i ffwrdd ‘turn off’, dihuno ‘wake’
for dihuno lan ‘wake up’, ysgrifennu ‘write’ for ysgrifennu i ffwrdd ‘write away’
(1998: 84);
in Rhosllannerchrugog dialect: trwsio ‘repair’ for neud fyny ‘do up’, profi ‘test’ for
troi allan ‘try out’, ffrwydro ‘explode’ for chwythu i fyny ‘blow up’ digwydd ‘happen’
for mynd ymlaen ‘go on’, dweud ‘say’ for gwneud allan ‘make out’, dosbarthu ‘distribute’ for rhoi allan ‘give out’ (1998: 184–185).
The author believes that calquing is a sign of language obsolescence as the phenomenon
is largely unidirectional and PVs are chosen by speakers despite the existence of native
terms. In Gwenhwyseg dialect calques were more frequent in the speech of informants
aged less than 60, which suggests their prevalence in younger speakers’ speech
(1998: 86).
In a study by Woolridge (2011, see 1.3.5.), some PVs have been listed among
common examples of word-for-word unconscious translation of English idiomatic
phrases and marked by professional translators as incorrect (Woolridge 2011:64). The
author noted that this type of interference could be eliminated by making the translation
process more conscious. PVs are also mentioned by Deuchar as examples of loan translation (2005: 616), and Ball and Thomas (2012: 120), who say that verb-particle construc-
131
tions are “relatively rare” in Welsh and become “widespread due to calques from English”. Finally, Robert (2013: 101) presents various types of Welsh loans from English
within Winford’s (2003) framework of borrowing. In her classification, constructions
such as setio i fyny ‘set up’ and sefyll allan ‘to stand out’ are loan translations (calques)
belonging to the category of loanshifts (loan meanings).
2.3.8. Summary of literature on PVs in Welsh
The above overview shows major gaps in the description of phrasal verbs in contemporary Welsh linguistics and gives grounds to assume that the small amount of research in
that area might have been influenced by ideologies of prescriptivism and purism.
Phrasal verbs are absent from all Welsh grammars available to Welsh speakers
with the exception of P.W. Thomas’s Gramadeg y Gymraeg, which devotes a long footnote to the constructions. It appears that the author has not given the subject much consideration since the description appears somewhat intuitive and lacks coherence. Moreover, the author ignores non-transparent constructions, presumably due to the fact that he
considers them calques.
Idiomatic PVs have been occasionally mentioned in academic studies, mostly in
the context of borrowing from English. The opinions of researchers on these constructions
differ. Some view PVs in a rather negative way, claiming that they are something unnatural to Welsh (Fowkes 1945) or even do not belong to the ‘real’ language (Heinz 2003;
Asmus and Williams 2014). Extensive calquing in colloquial speech is seen as a sign of
language deterioration (Jones 1998) and pleonastic constructions are considered redundant (Jones 1979, Thomas 1996). Other researchers take a more neutral view, simply acknowledging the existence of PVs (Ball and Thomas 2012; Hirata 2012; Robert 2013).
Negative opinions on using verb-particle constructions are based on the idea that Welsh
PVs have their indigenous equivalents which may become endangered if replaced with
loans from English. However, it has been noticed that not all PVs – such as troi ymlaen
for ‘turn on’ – have single-word counterparts. As a consequence, they fill lexical gaps
and enhance the lexicon of the language (Thomas 1987). Yet, the prevailing notion appears to stress the fact that Welsh should not become too similar to English and lose its
native idioms.
132
The most comprehensive and ground-breaking descriptive study of PVs in Welsh
is Rottet (2000, 2005), who points to the complexity of the issue and pays attention to the
great number of idiomatic constructions. He also acknowledges different types of PVs,
an issue which had been previously neglected. Moreover, he draws attention to some prescriptive stances expressed in Welsh lexicography and concludes that different types of
PVs have been fully integrated into Welsh by now, which reflects the bilingual reality in
Wales.
Little attention has been paid to the diachronic aspect of PVs in Welsh. Jones
(1979:113) and Rottet (2005) put forward the claim that contact with English reinforces
the elements inherent to Welsh, extending the patterns already existing in the language to
form new idiomatic combinations. The latter author supports it with examples dating back
to medieval times.
Altogether, the body of literature on PVs in Welsh is small and, interestingly,
a substantial part of research on the subject has been conducted by academics from outside Wales. This reluctance on the part of Welsh linguistics to engage with the subject is
perhaps related to the general lack of studies on borrowings from English (see 1.3.1.1).
2.4. Phrasal verbs in the present study
The present study intends to continue the investigation of the issues outlined in the presented literature. As shown in the above review, one can observe that PVs are generally
excluded from the linguistic description of the language, particularly in grammar books
and no sufficient in-depth description has been provided so far. This gap is rather curious
in view of the prevalence of these constructions and the fact that they have been documented in the Welsh language for centuries. The main aim of the dissertation is to create
such description, which will be presented in the Conclusions. With this goal in mind, this
section will specify the definitions used in my study, classify the investigated constructions and describe their features relevant to the thesis. I will then review and expand on
Rottet’s findings, firstly by conducting a larger and more rigorous corpus study offering
some quantitative as well as qualitative data. For the reasons discussed in 1.5, the present
research will not explore the area of the spoken language to a great extent. It will, however, contribute to the debate on the integration of the constructions in standard Welsh.
133
Apart from providing a corpus-based analysis of phrasal verbs in the press and fiction
(Chapter 3) and metalinguistic discourse in lexicographic and pedagogical materials
(Chapter 4), it will measure the stated acceptability of PVs among proficient Welsh speakers, a matter which had not been researched so far, except for Hirata’s (2012) suggestion
that some professional speakers might prefer one-word equivalents of PVs and that this
might be related to their exposure to literary language and age.
Based on this evidence, the final description of PVs will also aim to establish the
position of PVs within the Welsh linguistic norm in view of language ideologies. Rottet
(2000:115) has claimed that there is generally no stigma attached to PVs, although there
might be a pedagogical and linguistic norm against them. In the present dissertation, I
would like to verify this claim on the basis of a more structured corpus and field study,
hypothesising that, in the last 20 years, certain PVs have become even more established
in the Welsh language. Following studies on Irish by Stenson (1997) and Veselinović
(2006), this study will not be limited to verb-adverb constructions but include also idiomatic prepositional verbs and phrasal prepositional verbs, which were not previously investigated.
2.4.1. Definition and characteristics of phrasal verbs in Welsh
Phrasal verbs are defined in this study as combinations of verb and particle, which form
a single unit in the minds of the speakers. Similarly to English, Welsh phrasal verbs typically include common lexical verbs associated with physical movement. In English
nouns or adjectives without a verbalizing suffix may be used as verbs, e.g. egg on, pile
up, dumb down46, while in Welsh they are verbalized by adding a morphological verbal
endings, the most frequent of which is -(i)o, e.g. bordio i fyny ‘board up’, arafu i lawr
‘slow down’. Another difference between English and Welsh as regards word formation
is that nominalisation of PVs in Welsh is hardly possible, except for wholesale borrowings such as getawe ‘getaway’, gif-yp ‘giveup’, which would be accompanied by a Welsh
verb (cael getawe, rhoi gif-yp, etc.).
46
Thim (2012: 30) distinguishes here between transparent PVs such as sex up and lexicalised, semantically
opaque constructions termed cranberry verbs, such as eke out.
134
The term particle is used here to designate prepositions, prepositional adverbs or
spatial adverbs which follow the lexical verb and semantically denote location and/or
direction in space or time. A prepositional adverb, following Quirk et al. (1985), is an
adverb which is formally identical to or related to a preposition but has a different syntactic status and properties:
Thus in the sentence A car drove past the door, past is a preposition, while in the sentence
A car drove past, past is a prepositional adverb. It is capable of standing alone and in contrast to simple prepositions it normally receives stress (Quirk et al. 1985: 713–714).
A list of particles in Welsh verb-particle constructions was proposed by Rottet
(2005: 42). The list will be reviewed in 3.4 in the corpus study. The number of particles
which may be used to form idiomatic PVs is more limited than in the case of transparent
ones. In contrast to English, Welsh particles are morphologically variable and in that respect can be divided into the following groups:
a) simple spatial adverbs, such as allan ‘outside’47, draw ‘over’ (there), ymaith
‘away’. Some of those, are derived from prepositions, e.g. drosodd ‘over’ from
dros ‘over’ (cf. Thomas 1996: 430);
b) complex spatial adverbs, which are prepositional phrases consisting of the preposition i and a noun referring to a topographical feature “which iconically suggests
a direction or a location” (Rottet 2005: 42). Examples include: i ffwrdd ‘away’
from ffordd ‘road’, i fyny ‘up’ from mynydd ‘mountain’, i lan ‘up’ from glan
‘shore’, i lawr ‘down’ from llawr ‘floor, ground’, mas (ma’s) ‘out’ from maes
‘field’ and bant ‘off, away’ from pant ‘valley’. In the latter two cases the preposition ‘i’ has disappeared. These particles may be followed by a noun, but cannot
have a pronoun inserted between the components;
c) simple non-inflected prepositions, such as â ‘with’, gyda ‘with’; some simple
prepositions such as heibio ‘past’ may also function as prepositional adverbs;
d) simple inflected prepositions, such as am ’about, around, for’, ar ‘on’, at ‘to, towards’, i ‘to, for’, dros ‘over, across’, trwy ‘through’ which take a morphological
ending according to the person and number. The ending is added to the preposition
47
The English translations of particles in this section are approximate, since most of the Welsh prepositions
may have other of English equivalents. Correspondence of meanings is analysed in detail in 3.3.
135
stem. As shown in Table 2, there is a considerable variation of inflected forms
across registers and dialects (cf. Thomas 1996: 343);
Table 2. Morphological variation of the preposition trwy ‘through’, after Jones (2013: 210).
Literary
Official
and Colloquial
Northern
Southern
1st. pers. sg.
trwof fi/i
drwyddo i
drwydda i
drwyddo
2nd. pers. sg.
trwot ti
drwyddot ti
drwyddat ti/chdi
drwyddot ti
3rd. pers. sg.
trwyddo ef
drwyddo fe/fo
drwyddo fo
drwyddo fe
trwyddi hi
drwyddi hi
drwyddi hi
drwyddi hi
1st. pers. pl.
trwom ni
drwyddon ni
drwyddan ni
drwyddon ni
2nd. pers. pl.
trwoch chwi
drwyddoch chi
drwyddach chi
drwyddoch chi
3rd. pers. pl.
trwyddynt hwy
drwyddyn nhw
drwyddan nhw
drwyddon nhw
e) complex inflected preposition, such as i mewn i ‘into’, oddi wrth ‘from’, which
consist of two elements, only the second of which is inflected;
f) complex prepositions, such as ar ôl ‘after’, ar draws ‘across’, o gwmpas ‘around’.
yn ôl ‘back’, ymlaen ‘ahead, forward, on’. These are prepositional phrases grammaticalised to form a preposition or an adverbial. If an object of such a phrase is
a pronoun, it is inserted between the elements of the phrase. Thus:
(20) Cer yn dy flaen chydig bach…. (Prysor 2010:52)
‘Go ahead a little bit.’
In the example (20) the particle ymlaen is split, and its original component blaen ‘front’
undergoes soft mutation as it is preceded by the prefixed pronoun for 2nd pers. sg. dy.
Some Welsh particles have dialectal variants. The four most common examples
are presented in Table 3. As can be seen, in this case the Northern Welsh variety is considered the standard, except for the preposition efo, which is avoided in formal registers,
replaced by Southern gyda or standard â.
Table 3. Dialectal variants of PV particles in Welsh
Standard Welsh
Southern Welsh
Northern Welsh
English equivalent
i ffwrdd/ymaith
bant
i ffwrdd
away/off
136
i fyny
(i) lan
(i) fyny
up
allan
mas
allan
out
â/gyda
gyda
efo, hefo
with
2.4.2. Semantic classification – Idiomaticity
Idiomaticity, or compositionality of meaning, is a crucial issue in discussing the semantic
properties of verb-particle combinations called here phrasal verbs. In the English language, the most basic classification of these verbs as regards idiomaticity is based on the
division between compositional (literal, transparent, non-idiomatic) and non-compositional (idiomatic, opaque) items. Obviously, no sharp dividing line between the two
groups can be drawn; a large proportion of PVs are polysemous and the same combination
may appear in different degrees of idiomaticity e.g. he came in the sense of “entered”
(transparent), the tide came in, ”moved higher up the beach” (semi-idiomatic), short coats
are coming in “are becoming fashionable” (idiomatic). In some cases idiomaticity may
be thus regarded not only as continuum but also as a cline related to historical developments of the meaning of a given PV from transparent to idiomatic. A major work presenting the evolution of particle meanings from the cognitive perspective is Rudzka-Ostyn
(2003). Her model will be used in the semantic analysis of Welsh particles in 3.3.
Idiomatic combinations of verb and particle occurred already in Old English and
had fully developed by the Middle English period (Thim 2012: 182–185). Thus far, researchers have found it difficult to convincingly describe long-term changes in the frequency of occurrence of the English particle verbs, although there is some evidence for
the increase of these constructions in the 19th century (for discussion of the issue cf. Thim
2012). In present-day English, PVs appear to be extremely productive, which can be seen
for instance in the phenomenon of verbalising nouns and adjectives which serve as the
base verb. Researchers have noted that phrases of this type are often strongly culturally
bound and related to contemporary issues; for instance, space out was used for “taking
drugs” in the time of space exploration (Bernstein 1974: 65). In Welsh, however, such
productivity is limited (see 2.4.1 above).
There is a lack of agreement among researchers regarding the role of idiomaticity
in determining the syntactical and lexical status of PVs (for an overview cf. Cappelle et al.
137
2010: 190). Many authors regard idiomaticity as a basic criterion for distinguishing PVs
from other multi-word items. Quirk et al. (1985:1162) see idiomatic vs. non-idiomatic
status as a major distinction between multi-word items and free combinations. Similarly,
Biber et al. (1999) define multi-word lexical verbs as “combinations which comprise relatively idiomatic units”.
In contrast, some authors, such as Fraser (1976), believe that transparent and idiomatic combinations are not the same phenomenon in view of structural differences between them, for instance a possibility of coordinating adverbs in transparent constructions
(see 2.4.2.1). However, there is also a number of arguments in favour of the opposite
view. One of them is that some transparent constructions are among the most widely used
multi-word combinations in English, which has been demonstrated by corpus studies e.g.
Gardner and Davies (2007). Consequently, these items appear to be retrieved as “readymade” items in the minds of speakers rather than being constructed anew each time they
are used (Cappelle et al. 2010: 190). Moreover, historically speaking, most non-compositional PVs have developed by way of metaphorical extension of meaning of the transparent constructions and it is often impossible to draw a clear-cut line between transparent
and idiomatic senses of polysemous combinations. All in all, there is much convincing
evidence that transparent and idiomatic PVs are, as concluded by Lindner, manifestations
of the same phenomenon differing only in terms of semantic features (Lindner 1981: 31).
In view of the above, in the present work idiomaticity is viewed as a continuum
with three main degrees based on Thim’s (2012) classification. These are: 1) compositional phrasal verbs, divided into a) transparent and b) aspectual, and 2) non-compositional idiomatic phrasal verbs.
1) Compositional combinations can be divided into two types:
a) transparent - combinations which are semantically compositional, where the verb retains its literal meaning and expresses verbal action, while the particle expresses directionality of the action. Hence, transparent PVs are free in that both the verb and particle
are exchangeable, with the verb expressing the kind of action and the particle expressing
the direction, as shown in Figure 1.
138
The above examples have their nearly exact counterparts in Welsh. In this context
Welsh expresses the motion-in-space relation in the same way as English, with obvious
differences in the SVO – VSO word order between the languages (Figure 2).
George
tossed
took
put
carried
threw
the food
up.
in.
away.
back.
out.
Figure 1. Exchangeability of verb and particle in English PVs (after Thim 2012: 14)
Taflodd1
Cymerodd
Rhodd
Cariodd
George y bwyd
i fyny.
i mewn.
i ffwrdd.
yn ôl.
allan.
taflu translates as both ‘toss’ and ‘throw’.
1
Figure 2. Exchangeability of verb and particle in Welsh PVs.
There is no doubt that transparent verb-particle combinations of this type occur
naturally in the Welsh language. Some of them, however, appear to function as set phrases
in the minds of the speakers and in some cases are replaceable by single items. For that
reason, some Welsh authors, e.g. P.W. Thomas (1996), have claimed that single verbs
such as dychwelyd ‘return’ are preferred over transparent phrases such as dod yn ôl ‘come
back’ (see 2.3.3).
b) aspectual – combination of verbs and particle whose meaning is usually fully transparent and readily understandable since the verb retains its literal meaning, while the particle
adds an aspectual interpretation of the verb. The most frequent aspectual particle in English is up; others include down, over, through, on, along, around and away. Characteristically, groups of such PVs show resemblances in particle meaning; for example, up may
imply completion, as in drink up, break up, while away suggests persistence, as in work
away, fire away (Quirk et al. 1985: 1163)48. In quite a few cases, however, the exact
meaning of a particle is more difficult to isolate, as in find out and in many instances the
48
For a comprehensive analysis of particle meanings cf. Rudzka-Ostyn (2003) and section 3.3.
139
aspectual meaning may overlap with directional one (Thim 2012: 19). Aspectual phrasal
verbs are free in the sense that ad hoc formations are possible, but since substitution is constrained by limited productivity they can be seen as semi-idiomatic
(cf. Thim 2012: 57)
There is a considerable number of aspectual PVs in Welsh, bearing close similarities with English ones, although not all English ones will have their Welsh counterparts.
An interesting feature of aspectual PVs is that in some cases the aspectual element is
logically redundant, for example up in finish up or grow up. Hence, such constructions
might be called pleonastic. The “redundancy” is less common with directional particles,
although not impossible, for example return back (Thim 2012: 17). The Welsh corpus
constructed for this study includes similar examples with the particle yn ôl ‘back’:
dychwelyd yn ôl ‘return back’, bacio yn ôl ‘back back’, cilio yn ôl ‘retreat back’.
It is these two properties of aspectual PVs – their semi-idiomatic status and possible redundancy – that attract the attention of Welsh linguists and teachers. It appears
that, being less productive in Welsh, such constructions may sound unnatural to some
Welsh speakers and consequently the idiomatic meaning of a particle is seen as being
calqued from English. For example, the particle i fyny ‘up’ is defined and translated in
GPC as
le neu safle uwch, i’r lan, tuag at le neu safle uwch (hefyd gyda bf. fel gosod, llyncu, rhoddi,
torri, etc., i drosi priod-ddull Saesneg):
up, upwards (also used with certain verbs to translate English idioms).
As shown above, aspectual PVs have also been mentioned by Jones (1979),
P.W. Thomas (1996) and Asmus and Williams (2014) as examples of incorrect or unnecessary borrowings from English.
2) The second group of PVs are non-compositional, idiomatic verb-particle combinations,
where the meaning of the phrase is not fully predictable from the meaning of its constituents and there is no possibility of contrastive substitution of the verb or particle, as in:
work out “solve a problem”, put something off “postpone”, bring someone round “persuade”, come across something “find by chance” or run out of something “have nothing
left”. However, the metaphorical meaning is not necessarily completely opaque, as in run
into someone “meet by chance” or see through “not be deceived”.
140
Idiomatic PVs are present in contemporary Welsh and the vast majority of them
have exact English counterparts e.g. gweithio allan ‘work out’, gadael i lawr ‘let down’.
For that reason, as in the case of aspectual PVs, it is possible to view them as calques
from English.
2.4.2.1. Criteria for distinguishing between compositional and idiomatic phrasal
verbs
There have been numerous attempts within English linguistics to distinguish between
compositional and non-compositional verb-particle constructions. In his classic study,
Bolinger (1971: 6-22), lists nine tests for exclusion/inclusion of multi-word items as PVs.
Some of these criteria were used by Quirk et al. (1985) in their classification, which distinguishes (idiomatic/aspectual) PVs from free combinations (called transparent PVs in
this study). Although the semantic criterion of idiomaticity is the main one within Quirk
et al.’s approach, they also mention a number of syntactic features distinguishing between
the two categories. Claridge (2000: 32) notices that Quirk et al.’s approach has some
drawbacks as they apply the semantic-based classification only to some types of multi
verb items, which leads to inconsistencies. In the present study, the primary criteria for
deciding whether a PV is idiomatic or not is its semantic transparency or opaqueness.
There are cases however, where this distinction might be blurred (cf. Quirk et al.
1985: 1153 and the "metaphorical appropriateness" of verbs such as bring up).
Additionally, several syntactic criteria mentioned by Quirk et al. (1985) are, to a
certain degree, applicable to the Welsh language as well as English, despite syntactic
differences between the two languages.
Firstly, transparent constructions allow inversion of the particle (21), in contrast
to idiomatic and aspectual combinations (22), which confirms the latter’s semi-idiomatic
status (Quirk et al. 1985: 523).
(21) a.
b.
(22) a.
b.
out came the spider
up went the flag
*out he carried the test.
*up he drank all the water
141
Parallel inversion of the particle is possible in Welsh transparent PVs, but impossible with
idiomatic ones, as shown in examples (23) and (24).
(23) a.
bant aeth e yno ei hunan
‘off he went there on his own’
b.
ymlaen aeth y car
‘on went the car’
(24) a.
*bant trodd hi’r goleuadau.
‘off she turned the lights’
b.
*ymlaen cariodd e’r prawf.
‘on he carried the test’
Secondly, in many transparent combinations a modifying adverb right or straight
may be placed between the verb and the particle (Quirk et al. 1985: 1153). This is not
possible with idiomatic PVs and the vast majority of aspectual ones. In Welsh, the adverb
yn syth ‘straight’ can also be used in the same way as in (25) and (26).
(25) Gan ei bod yn hwyrach na hynny, cynigiaf ein bod yn mynd yn syth ymlaen i bleidleisio. (CCCC)
‘As it is later than that, I propose that we move straight to voting.’
(26) Yna dyma Simon Pedr yn cyrraedd ar ei ôl ac yn mynd yn syth i mewn i'r bedd.
(John 20.6, beibl.net)
‘Then Simon Peter came along behind him and went straight into the tomb.’
One can also find examples of a similar use of the borrowing reit ‘right’ (27).
(27) Rhoddodd ei wynab reit i fyny at y ffenast nes bod ei wynt yn stemio'r gwydr.
(Prysor 2006: 70)
‘He put his face right up against the window until his breath steamed the glass.’
142
According to P. W. Thomas in a very formal, “if not outdated” style, Welsh allows the
placement an adverbial between the verb and particle as in the quoted folk song:
(28) Cyn delwyf i Gymru’n ôl (Thomas 1996:561)
‘Before I come back to Wales’
The third distinguishing feature of transparent combinations is coordination of particles,
as in:
(29) Maen nhw'n gyrru fel ffyliaid i fyny ac i lawr y lôn yna bob awr o'r dydd a'r nos.
(CEG)
‘They are driving like fools up and down that lane every hour of day and night.’
Such coordination is not possible with fully idiomatic PVS. However, there are some
borderline cases of semi-idiomatic PVs as in (30), where ‘move up’ and move down’ are
used metaphorically in the sense ‘climb the social ladder’.
(30) Am symud lan ydw i, nid lawr. (Roberts 2013: 169)
‘I’m intending to move up, not down.’
Another commonly used criterion for idiomatic status is replaceability by a single
word. As demonstrated in section 2.1.2.2, however, it is not a reliable criterion for several
reasons. Firstly, a considerable number of idiomatic PVs cannot be rephrased by a single
item, for example settle down, let someone off, run out of. Secondly, it is possible to find
one-word equivalents for a number of transparent PVs as well e.g. come in – enter, come
back – return (cf. Quirk et al. 1985: 1162). Another crucial aspect are semantic differences between PVs and their one-word equivalents as regards style and register e.g. throw
up and vomit. This makes the criterion of replaceability even more dubious.
Despite a number of useful criteria for establishing the idiomatic or non-idiomatic
status of PVs, it can be seen that idiomaticity of PVs in both English and Welsh forms
a continuum with plenty of difficult cases. Ultimately, as Claridge (2000: 41) says “It is
actually a (vicious?) circle we are moving in: we all know in some way what, e.g. a phrasal
143
verb is, but a full and theoretically adequate proof of this intuitive knowledge seems impossible”. It seems quite inevitable that in some cases including a particular item may be
an arbitrary decision of the researcher.
2.4.3. Syntactic classification of phrasal verbs
The terms regarding syntactic classification of PVs used in this dissertation are based on
Quirk et al. (1985). It should be remembered that these authors include only non-compositional constructions in their definition of multi-verb items. Transparent combinations,
where the verb and the adverb have distinct meanings, would be termed “free combinations” by the authors, while in PVs “the meaning cannot be predicted from the meaning
of verb and particle in isolation.” (Quirk et al. 1985: 1152). Quirk et al. divide multi-word
verbs into three categories: Phrasal Verbs, Prepositional Verbs and Phrasal-Prepositional Verbs. This classification will be used in this thesis, since all three categories occur
in Welsh and examples of each category are cited in literature, although it is mainly
Phrasal Verbs and Phrasal-Prepositional Verbs which have attracted attention. The category Phrasal Verbs will be renamed here Adverbial Phrasal Verbs to avoid confusion
with the more general term and names of the categories will not be capitalised.
2.4.3.1. Adverbial phrasal verbs
Adverbial phrasal verbs (APVs) consist of a verb and adverbial particle. In Welsh, as in
English, they might be intransitive e.g. mynd ymlaen ‘go on’ or transitive chwythu
rhywbeth i fyny ‘blow something up’. With regard to the syntactic behaviour of transitive
phrasal verbs, there are differences between Welsh and English. In English the particle
may be placed before or after the direct object e.g. He switched the lights on/ He switched
on the lights; however, if the object is a pronoun, the particle normally comes after it: He
switched them on. In Welsh both orders are possible, although placing the object between
the verb and particle seems to be more common. Jones (1979), however, claimed that the
Verb – Particle – Object order is impossible. Thus:
144
(31) Mae John wedi sortio’r papurau allan
*Mae John wedi sortio allan y papurau (Jones 1979: 113)
‘John has sorted the papers out’
Rottet (2005) and Hirata (2012) do not support this claim, providing numerous examples
of both word orders. The latter’s study showed both orders were judged to be grammatical
by native Welsh speakers, although the Verb – Object – Particle order was generally preferred. Moreover, Hirata notices that this order was preferred if the particle was complex,
while a single particle was more acceptable in the Verb – Particle – Object order. The
issue will be further investigated in the subsequent chapter (3.2.3.3).
When the object in transitive constructions is a pronoun, Welsh uses different syntactic patterns depending on whether the verb is conjugated or a verbal noun is used. The
pronoun can be prefixed (32), affixed (33) or suffixed (34) to the verb. Suffixed pronouns
may be optionally added when prefixed or affixed pronouns are used as in (35).
(32) pan fo'r ergyd nesa a'r un nesa a'r un nesa yn fy nharo i lawr
‘when another and another and another blow strikes me down’
(33) phob math o bethau i'w hannog ymlaen
‘all sort of things urging them on’
(34) cyfra fi mas
‘count me out’
(35) dwi'n ei rentu e mas
‘I rent it out’49
2.4.3.2. Phrasal prepositional verbs
49
Examples 33-36 taken from my corpus (see Chapter 3).
145
Phrasal prepositional verbs (PPVs) contain two particles, the second of which is
a preposition. In English they are largely restricted to informal language: look forward to,
put up with, get away with, do away with. They can be transitive or intransitive (Quirk
et al. 1985: 1160) and the majority of them show a degree of idiomaticity (Thim 2012:
28). Common Welsh examples include: edrych ymlaen at ‘look forward to’, bwrw
ymlaen â ‘continue with’, dal i fyny â ‘catch up with’, rhedeg allan o ‘run out of’. A very
limited number of PPVs in Welsh may take a direct object, such as golchi
rhywbeth i lawr â rhywbeth ‘wash something down with something’.
2.4.3.3. Prepositional verbs
According to Quirk et al., prepositional verbs (PPs) consist of a verb and preposition which is semantically and/or syntactically associated; they may be distinguished
from APVs by the inability to move the particle after the following noun phrase (Quirk
et al. 1985: 1155–1156). They may or may not have a direct object – cf. He invested in
property and He invested his money in property. Both the verb and particle in PPs can be
used quite literally or metaphorically, along the idiomaticity continuum e.g. look at, look
into, look after. Of interest to this thesis are only PPs which are idiomatic in meaning and,
following Keiser’s framework model, form a semantic whole although syntactically they
are marginally separable (Keizer 2009: 1201). While from a purely syntactic point of
view constructions such as come across something, look after something are a different
phenomenon to APVs, they are commonly included in dictionaries of PVs due to their
idiomatic properties.
Welsh has a number of idiomatic PPs, including phrases which translate directly
into English, such as dod ar draws rhywbeth ‘come across something’, dod rownd rhywun
‘come around someone’, edrych ar ôl rhywun’ ‘look after someone’ and native constructions such as taro ar rywun ‘bump into someone’, lit. ‘hit on someone’, torri ar draws
rywun ‘interrupt someone’, lit. ‘cut across someone’. There is also a type of idiomatic
expressions characteristic to Welsh, where the preposition is inflected in the 3rd person
singular feminine, but there is no entity to which the prepositional phrase refers, for instance mynd ati ‘go into action, set to work’, lit. ‘go at it’, dal ati ‘continue, keep up’ lit.
‘keep to it’, cymryd arni ‘pretend’, lit. ‘take on it’.
146
2.4.4. Language contact classification
A further categorisation of PVs, being of crucial importance to this thesis, is related to
language contact phenomena. The classification proposed here is based on Rottet
(2005:50-60) and Stenson for Irish (1997). Welsh PVs can be divided into:
1.
Native Welsh constructions (N). These include all or the majority of transparent
PVs since, as explained above (2.4.2), they are a natural way to express motion
through location. There is also a relatively small group of idiomatic verbs which
can be assumed to be native since they do not have English equivalents. Rottet
(2005:50-52) provides a list of such idioms, e.g. codi allan ‘get around’, lit. ‘rise
out’, hwylio i lawr ‘blow over’, lit. ‘sail down’.
2.
Loan renditions (LR). The term originally coined by Weinreich (1953) describes
items where the translation diverges from the donor language rather than reproduces it element by element. In other words, a PV might be considered an LR if
one of the elements corresponds closely to English. For example, the phrase dod
dros ‘get over’, ‘overcome’, literally meaning ‘come over’ may be a LR, especially in view of the fact that English get is an all-encompassing verb which does
not have an obvious Welsh equivalent.
3.
Pleonastic PVs (PL). These are semi-idiomatic, aspectual constructions, in which
one of the elements is, from a logical standpoint, redundant. Common Welsh examples are eistedd i lawr ‘sit down’ and tyfu i fyny ‘grow up’. Rottet demonstrates
that the use of a particle in many such cases is optional (2005:53) and the version
without a particle may be considered more formal.
4.
Directly translatable PVs (DT). This category includes idiomatic verb-particle
constructions where both the verb and particle are directly translatable into English and may therefore be assumed to have been calqued, for example: troi lan
‘turn up’, chwythu rhywbeth i fyny ‘blow something up’, dod ar draws rhywbeth
‘come across something’, rhedeg allan o rywbeth ‘run out of something’. A single
Welsh PV may sometimes serve as an equivalent of several English PVs, due to
polysemy of verbs such as torri ‘break’, ‘cut’; there are also cases when a single
English PV may be translated in several different ways e.g. miss may be rendered
as methu or colli (Rottet 2005: 56).
147
5.
Loanblends (LB). These are constructions which combine borrowed and native
morphemes. Most frequently the borrowed item is a verb, usually with the ending
-(i)o, for example cyrlio i fyny ‘curl up’, ffeindio allan ‘find out’. The particle may
also be borrowed. In fact, only two particles are used in that respect: off and rownd
e.g. dod rownd ‘come round’, mynd off ‘go off’. In some cases both elements may
be loanwords, as in hitio off ‘hit off’.
6.
Wholesale borrowings (WH). These are the items where the whole linguistic material is borrowed from English. Such examples may be difficult to distinguish
from code-switching. In the present study, which focuses on the written language,
an item is considered a WH if the spelling reflects Welsh pronunciation. As observed by Rottet (2005), WH items are very rare in Welsh; examples include only
nominalised items such as cael getawe. In the present corpus, set phrases, such as
cym on ‘come on’ and howld on ‘hold on’ were found.
2.4.5. Scope and focus of the present study
On the basis of the relatively small body of available literature, this study focuses on the
verb-particle constructions in Welsh which arise controversies as regards language contact and standardisation. As shown above, transparent PVs are a natural way of expressing
motion through location in the Welsh language: there is hardly a way to replace expressions such as mynd i fyny ‘go up’, dod i mewn ‘come in, enter’, neidio allan ‘jump out’.
Nevertheless, replaceability is achievable in some cases such as dod yn ôl ‘come back’ –
dychwelyd ‘return’. This is pointed out by P. W. Thomas, who appears to suggest that
replaceability with single items makes certain PVs redundant. For that reason, the case of
the transparent PV dod yn ôl has been investigated in the field study presented in Chapter 5. Other than this example, the present research will not be concerned with transparent
PVs. It does, however, include the semi-idiomatic aspectual phrases, with a focus on constructions called here pleonastic PVs, such as tyfu i fyny, eistedd i lawr, since these are
also mentioned in literature as nonstandard borrowings from English. The main focus of
the study are idiomatic PVs, classified syntactically into three categories: APVs, PPs and
PPVs (2.4.3). In distinguishing between transparent and idiomatic PVs, the criteria described in section 2.4.2.1 are used. The following chapter presents an analysis of a corpus
148
of written texts designed for this study. Chapter 4 focuses on the presentation of PVs in
Welsh dictionaries and pedagogical materials. Finally, the results of a field study on the
acceptability of PVs among professional speakers of Welsh are presented in Chapter 5.
149
Chapter 3: Phrasal verbs in Welsh – a corpus study
3.1. Introduction
In this chapter, the description of phrasal verbs in Welsh emerging from the literature on
the subject reviewed in Chapter 2 is contrasted with findings based on a small sample
corpus of written texts and the available lexicographic data. As could be seen, observations regarding PVs made by linguists thus far have been largely intuitive rather than
based on a systematic analysis. This study attempts to fill this gap and provide the basis
for a detailed linguistic description of the phenomenon of PVs in Welsh, underlining its
major complexities. It also aims to validate Rottet’s (2005) claim about the widespread
usage of PVs in press and fiction.
Despite its small size, the sample of texts analysed can be seen as representative
of contemporary Welsh in that it consists of texts with different degrees of formality, from
written spoken through semi-formal and formal, thus giving an idea of the language accessed on an everyday basis by literate speakers of Welsh.
The corpus is presented in the first part of the chapter, followed by a quantitative
and qualitative analysis based on the categorisations presented in the previous chapter,
examining distributions of PVs according to their syntactic properties and status in relation to language contact. Next, the sample is analysed qualitatively with focus on stylistic
markedness. The final section investigates correspondence of meanings of Welsh and
English particles, attempting to find regularities in transfer phenomena, using a cognitive
model of particle semantics based on Rudzka-Ostyn (2003).
150
3.1.1. Aims of the study
The primary aims of building the corpus were to:
1. create a database of examples of idiomatic and semi-idiomatic verb-particle constructions in written Welsh;
2. determine the most prevalent idiomatic PVs and examine their various senses and
uses;
3. review the list of particles entering verbal constructions provided by Rottet (2005);
4. examine the productivity of specific verbs and particles with reference to language
contact;
5. analyse the syntactic behaviour of idiomatic PVs in Welsh;
6. investigate the status of idiomatic PVs as borrowed or native items;
7. analyse the distribution of various categories of PVs across registers of Welsh regarding their native or non-native status;
8. investigate the markedness of borrowed and non-borrowed PVs.
3.1.2. Scope of the research
In view of the aims presented above, two types of constructions were included in the
corpus:
a) idiomatic phrasal verbs, including three syntactic types: adverbial phrasal verbs
(APVs), phrasal prepositional verbs (PPVs) and prepositional verbs (PPs) (see
2.4.3);
b) pleonastic, semi-idiomatic phrasal verbs of all three syntactic types.
Consequently, the following constructions were excluded from the corpus:
a) any non-compositional PVs other than pleonastic ones;
b) native idiomatic verb–preposition-verb patterns. There are few such constructions
in Welsh, the most common example being bod am wneud rhywbeth, lit. ‘be about
doing something’, ‘want, intend to do something’;
c) native idiomatic constructions with a preposition inflected in the 3rd person singular feminine and no entity to which the prepositional phrase refers. These include for example: bwrw ati (i) (lit. ‘hit towards it”, ‘continue something, begin’),
151
cymryd arni (lit. ‘take on it’, ‘pretend’), mynd ati (i) (lit. ‘go towards it (to)’, ‘go
into action, set to work’);
d) reflexive constructions, such as dod at ei hunan ‘come to oneself’, cadw rhywbeth
at ei hunan ‘keep something to oneself’;
e) full idioms, such as proverbs, sayings, fixed phrases – constructions in which not
only the verb and particle, but also other elements, such as grammatical form or
direct object are fixed. Idioms with PVs found in the course of the research were
mostly translated directly from English e.g. gwneud dy feddwl i fyny ‘make up
your mind’, gallu gwneud gyda rhywbeth ‘can do with something’, bod wedi torri
allan i wneud rhywbeth ‘be cut out to do something’, brathu i ffwrdd mwy nag y
gallwch gnoi ‘bite off more than you can chew’. Some of these calques were mentioned already by Fowkes (1945). The few native idioms found usually contained
PPs rather than PVs e.g. mae hi wedi canu ar rywun, lit. ‘it has sung on someone’,
‘someone is in trouble’, mynd o flaen gofid, lit. ‘go ahead of distress’, ‘to go to
meet trouble’;
f) example of code-switching into English, indicated by italics.
3.2. Analysis of the corpus
3.2.1. The corpus
Due to lack of digitalised material and easily searchable electronic corpora of Welsh,
a corpus was created especially for the purpose of this study. The corpus consists of
printed texts: works of fiction and periodicals, which were annotated manually. Automatic extraction of PVs was not feasible or reliable for a number of reasons. First, the
vast majority of texts were not available in digital form. Secondly, conducting an automatic corpus search for PVs would require at least a closed list of particles. The only
existing list provided by Rottet (2005: 42) cannot be treated as exhaustive due to different
categorisation used in the present study, which includes prepositional verbs. Moreover,
the great multitude of morphological and dialectal variants of Welsh verbs and particles
would make computer extraction extremely challenging. For instance, the most frequent
152
particle, ymlaen, may have a number of reduced and quite unpredictable spellings e.g.
‘mlaen, ‘mlân, ‘m’lan in written colloquial language. The same concerns verbs: for example, the verb edrych ‘to look’, has at least 20 inflected forms which might be reduced
to spoken forms such as ‘drych. What is more, in South Welsh dialects the verb can be
substituted by another verb, disgwyl, which can further be reduced to forms such as
‘shgwl. Finally, bearing in mind the scope of the study has been confined to idiomatic
constructions, manual annotation of the whole corpus would be required in any case. For
these reasons, this method was decided to be the least time-consuming and most reliable,
allowing for an in-depth study of the relatively small sample.
The corpus was aimed at representing the varieties of written language available
to ordinary speakers of Welsh. The sample consisted of two subcorpora: works of fiction
and press, which represented primarily semi-formal and informal registers of Welsh. Consequently, some formal registers, such as academic and official administrative texts, are
underrepresented in the corpus. The motivation behind not including them was the prevalence of English to Welsh translations in the production of such documents (Screen
2016; 1-3) and their little use among speakers of Welsh (see 1.2.4.1, Evas and Cunliffe
2016), therefore their low impact on the perceived linguistic norm. The corpus is a smallsized one, estimated at no more than 1 million words. As the texts were not digitalised
due to technological and time limitations, exact word count of the sample has not been
performed.
The selected fictional works were five novels awarded Gwobr Goffa Daniel Owen
(Daniel Owen Memorial Prize) in the years 2009-2013: Y Llyfrgell (2009) by Fflur Dafydd, Adenydd glöyn byw (2010) by Grace Roberts, Tair rheol anhrefn (2011) by Daniel
Davies, Afallon (2012) by Robat Gruffudd and Craciau (2013) by Bet Jones. The Daniel
Owen Memorial Prize is awarded annually at the National Eisteddfod of Wales for a novel
of at least 50,000 words. The novels are adjudicated based on the author’s language proficiency and skilful use of idiomatic Welsh, which is why they were chosen for the study
as representative high-quality texts.
The press corpus consisted of newspapers and magazines published in Wales
in a single period, July-August 2014. The sample included the weeklies Golwg and
Y Cymro, 4 issues each, the monthly quality magazine Barn, cultural quarterlies Y Faner
Newydd, Y Wawr and Y Casglwr, Christian magazines Cristion Gorff and Gwyliedydd,
a satirical magazine LOL published annually during the National Eisteddfod and 7 issues
153
of local newspapers, the so-called papurau bro from Mid Wales and Carmarthenshire.
The full list is available in Appendix A.
The two subcorpora were of a differing nature in that literary representations are
attempts on the part of authors to reproduce what they see as ‘authentic’ language,
whereas the press corpus data are more primary in nature. The corpus includes a broad
variety of registers, from formal and literary in quality press and narrative parts of novels
to written colloquial in dialogues and popular magazines. Texts in the press corpus represent a wide selection of fields and topics, from politics, business and current matters,
through art, culture, religion and history, to pastimes, hobbies and humour. Regarding
dialectal differences, the corpus contains samples of all major geographical varieties of
Welsh, with the limitation that some Northern and Southern Welsh varieties are underrepresented in the sample of local papers due to lack of access to them at the time the
material was collected.
3.2.2. Methods
All texts in the sample were read twice. Concordances with PVs were extracted and annotated manually in an Excel spreadsheet. If a PV was repeated within one clause in
a narrative or two consecutive sentences in a dialogue, it was counted only once.
The spreadsheet created consisted of the following data: a) source code b) page
number c) phrasal verb c) dialectal or colloquial variant (if applicable) d) English translation e) concordance f) narration type g) verb h) particle i) second particle (for PPVs)
j) syntactic category k) language contact category. A sample of the created spreadsheet is
included in Appendix A.
Dialectal or colloquial variants.
When particles and verbs had their dialectal or colloquial variants, the form considered
standard were listed as the cardinal one, while the form occurring in the concordance was
placed in the “Variant” column. The variants of particles and their frequencies in the corpus are presented in Table 4.
The variants of verbs constituting PVs listed in the corpus were: chwara
and whare for chwarae; dala for dal; dwad and dŵad for dod; disgwyl, disgwl, dishgwl
154
and edrach for edrych; ista for eistedd; ffindo, ffindio for ffeindio, ffito for ffitio, gadel for
gadael, llanw for llenwi, pigo for picio, raso for rasio, slammo for slamio, sorto for sortio,
towlu for taflu, sgwennu for ysgrifennu.
Table 4. Dialectal or colloquial variants of particles and their frequencies in the corpus.
Standard
form
Total no.
of tokens
var. 1
no.
ymlaen
369
mlân
8
allan
i fyny
i lawr
â
i mewn
ar ôl
yn ôl
i ffwrdd
o gwmpas
drosodd
trwy
167
96
75
55*
49
48
39
18
14
11
11
mas
lan
lawr
gyda
mewn
'rôl
'nôl
bant
ymbytu
drosto
drwy
61
40
10
6
9
1
5
10
1
1
1
Variants and number of tokens
var. 2
no var. 3 no.
var. 4
.
mlaen
5
mla'n 4
'mla'n
fyny
2
‘da
miwn
5
1
hefo
fewn
5
1
ymaith
1
i ffwr'
1
efo
no.
var. 5
no.
2
‘mla
en
1
4
da
1
*the number includes occurrences of â as both the first and second particle.
English translation
English equivalents were provided for each Welsh verb-particle construction. Dictionaries of Welsh: GPC, GA, Gwe were consulted when needed. English dictionaries used to
distinguish the senses of borrowed PVs were Oxford English Dictionary (3rd. edition,
2017) and Macmillan Dictionary Online (Macmillan Dictionary Online 2017).
Mode of narration
Each PV concordance was assigned to one of two modes: narrative (N) or conversation
(C). N examples were those found in the narrative parts of novels and in press. C examples
were those occurring in dialogues or internal monologues in the novels and direct quotations or interviews in the press. This distinction allowed to distinguish representations of
more informal spoken registers.
Syntactic category
PVs were tagged according to 4 syntactic categories: prepositional verbs (PP), intransitive
adverbial phrasal verbs (APVi), transitive adverbial phrasal verbs (APVt) and phrasal
155
prepositional verbs (PPV) (see 2.4.3). Similarly to APVs, PPVs could follow two syntactic patterns: Verb + Particle + Preposition + Direct Object, e.g. rhedeg allan o rywbeth
‘run out of something’, and Verb + Direct Object + Particle+ Preposition + Direct Object,
e.g. cadw rhywun allan o rywbeth ‘keep someone out of something’. However, as examples of the second pattern were extremely rare in the corpus, this distinction was not taken
into consideration.
Language contact category
PVs were tagged according to 5 language contact categories described in 2.4.4: native
constructions or loan renditions (N/LR), pleonastic semi-idiomatic PVs (PL), verb-particle constructions which are directly translatable into English and may often be calques
(DT), loanblends (LB) and wholesale borrowings (WH).
3.2.3. Quantitative analysis – results
3.2.3.1. Verbs and particles
The PVs in the corpus were made up of 127 different verbs, excluding dialectal and colloquial variants. Among those, 45 verbs (35%) were borrowings from English; 37 of them
were formed with the -(i)o ending, 3 with -ian ending (hongian ‘hang’, prowlian ‘prowl’,
socian ‘soak’), 2 with -u ending helpu ‘help’ and rhentu ‘rent’ and 3 were phonetic adaptations of English verbs occurring in wholesale borrowings, cym ‘come’, ffyc ‘fuck’,
howld ‘hold’. It should be noted that according to GPC the majority of these borrowings
were first recorded in Welsh before the 20th century, with some verbs going back to the
medieval period (Table 5). Seven items were not found in GPC: these were modern colloquial borrowings, including taboo words. The list of verbs demonstrates the productivity of PVs, which are created with both well-established and recently borrowed items.
Table 5. First recorded uses of borrowed verbs constituting PVs in the corpus according to GPC.
Century of 1st
recorded use
14th
No. of borrowed verb in
the corpus (types)
1
Verbs
cnocio
156
15th
16th
17th
18th
19th
20th
21st
not in GPC
TOTAL
6
6
6
8
6
4
1
7
45
cyrlio, cario, helpu, pacio, pasio, rhentu
gwatsio, picio, pwyntio, setio, socian, sortio
clirio, ffeindio, ffitio, hitio, hongian, pigo
actio, bacio, bordio, howld, popio, setlo, stretsho, sticio
contractio, patsio, prowlian, pwmpio, rasio, snapio
ffonio, optio, slamio, teipio
logio
cashio, copio, cwlio, cym, ffwcio, ffyc, pitsho
The most productive verbs in the sample were mynd ‘to go’ and dod ‘to come’
entering combinations with 22 and 19 different particles, respectively, including various
combinations of two particles in PPVs. The verbs combining with 5 or more different
particles are presented in Table 6.
Table 6. The most productive verbs in the corpus.
Verb
English
translation
Particles
No. of types
mynd
go
22
dod
come, get
19
40
troi
turn
13
16
tynnu
12
18
cadw
pull, draw,
drag, take.
keep
12
16
edrych
look
11
13
rhoi
give/put
10
19
torri
break/cut
8
13
galw
call
â, allan, allan â, allan at, allan o, am, ar ôl, draw,
dros, trwodd/drwodd, heibio, heibio i, i ffwrdd,
i fyny, i lawr, i mewn, o gwmpas, rhagddo,
trwy, ymlaen, ymlaen â, yn erbyn
allan, allan â, allan o, ar draws, ar ôl, draw,
dros, drosodd, drwodd, i fyny, i fyny â, i lawr,
i mewn, i mewn i, oddi ar, rownd, trwy, ymlaen, ymlaen â
allan, allan am, allan yn, ar, at, drosodd, i
ffwrdd, i fyny, i lawr, i mewn, i mewn i, ymlaen, yn ôl ar
allan, allan o, am, ar ôl, at, i fyny, i lawr, i
mewn, i mewn i, oddi wrth, yn ôl, yn ôl o
allan, allan o, ar, draw, draw o, draw oddi wrth,
i ffwrdd, i fyny, i fyny â, i lawr, o gwmpas, ymlaen
allan, am, ar ôl, trwy, ymlaen, ymlaen am, ymlaen ar, ymlaen at, ymlaen i, yn ôl, yn ôl dros
allan, ar draws, drosodd, heibio, i ffwrdd, i
fyny, i fyny ar, i lawr, i mewn, ymlaen
allan, allan i, allan o, ar draws, i fyny, i lawr, i
mewn, yn ôl
â, allan, draw, heibio, i fyny, i mewn
No. of types
with
distinguished senses
32
6
8
symud
move
6
10
bod
be
6
8
cymryd
gwneud
take
make/do
i ffwrdd, i fyny, i lawr, i mewn, ymlaen, ymlaen
at
allan o, ar ôl, drosodd, i fyny i, o gwmpas, ymlaen
allan, drosodd, drosodd o, i ffwrdd, ymlaen
â, allan o, heibio, i fyny, i fyny am
5
5
8
6
157
The PVs in the corpus were made up of 28 different particles, excluding dialectal and
reduced colloquial variants listed above. The particles and their frequencies are presented
in Table 7. Out of these, the following prepositional particles formed PPs: â (gyda/efo),
am, ar, ar draws, ar ôl, at, dros, heb, heibio, i mewn, o gwmpas, oddi ar, oddi wrth,
rownd, trwy, yn erbyn. The following adverbial particles combined with prepositions to
form PPVs: allan, draw, drosodd, drwodd, heibio, i fyny, i lawr, i mewn, o gwmpas,
ymlaen and yn ôl. Among those allan (mas) was the most productive particle, forming
30 out of 73 PPV types. The prepositions used as the second particle in PPVs were:
â (gyda/efo), am, ar, at, dros, i, o, oddi wrth and yn.
Table 7. Particles constituting PVs in the corpus and their frequencies.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Particle*
No. of types
No. of tokens
allan (mas)
i fyny (fyny/lan)
ymlaen
i lawr (lawr)
i mewn
yn ôl
i ffwrdd (ymaith, bant)
heibio
o gwmpas (ymbytu)
ar ôl
drosodd
draw
trwy
ar draws
ar
dros
drwodd
am
â (gyda, efo)
at
off
on
rhagddo
yn erbyn
heb
oddi ar
oddi wrth
rownd
TOTAL
100
54
49
34
28
16
13
12
11
9
9
7
7
6
6
6
6
5
4
3
3
2
2
2
1
1
1
1
398
167
96
369
75
49
39
19
50
14
48
11
32
11
68
16
15
7
9
7
7
7
10
2
2
1
1
1
1
1134
*excluding the second particle in PPVs.
158
3.2.3.2. Frequency count
A total of 1134 concordances with verb-particle constructions of interest were found in
the corpus. Among those, there were 271 PV types, taken as a combination of a given
verb and particle, and 332, if counting syntactic variants of the same PV separately (transitive/intransitive, with or without a second particle). Within these, 398 different senses
of PVs were distinguished. This is the number which will be taken into consideration
henceforth in discussing PV types, unless marked otherwise.
A list of 23 most frequent PVs in the corpus, occurring more than 10 times, is
presented in Table 8. The left-hand side of the table presents PV cardinal forms
(verb + particle), with the number of tokens including all its syntactic variants, with transitive/intransitive distinction and variants with different prepositions. For instance, the
most frequent PV, edrych ymlaen usually appears as a PPV with the preposition at
(125 tokens). However, three other prepositions which combine with this verb were
found: i (14), am (7) and ar (1). The verb occurred 14 times as an APVi without a second
particle. This produces 161 concordances altogether. On the right-hand side of the table,
different forms of the same PV are counted separately – thus edrych ymlaen at, edrych
ymlaen and edrych ymlaen i occupy three positions on the list.
Table 8. The 23 most frequent PVs in the corpus.
with syntactic variants included
syntactic variants separately
no.
phrasal verb
no. of tokens
no.
phrasal verb
no. of tokens
1
edrych ymlaen
161
1
edrych ymlaen at
125
2
mynd ymlaen
94
2
mynd ymlaen
86
3
edrych ar ôl
35
3
edrych ar ôl
35
4
torri ar draws
35
4
torri ar draws
35
5
dod ar draws
30
5
dod ar draws
30
mynd heibio
29
6
mynd heibio
30
6
7
cario ymlaen
20
7
dod draw
18
8
dod draw
18
8
cario ymlaen
17
9
edrych yn ôl
17
9
edrych yn ôl
16
10
symud ymlaen
16
10
eistedd i lawr
15
11
mynd allan
16
11
edrych ymlaen
14
12
eistedd i lawr
15
12
edrych ymlaen i
14
13
dod ymlaen
15
13
symud ymlaen
14
troi i fyny
14
dod ymlaen
13
14
troi i fyny
14
14
15
torri i lawr
13
15
159
16
12
16
torri i mewn
12
bod ymlaen
12
17
bod ymlaen
12
18
bwrw ymlaen
12
18
rhoi ymlaen
11
19
rhoi ymlaen
11
19
sortio allan
11
20
sortio allan
11
20
torri i lawr
11
21
tyfu i fyny
11
21
tyfu i fyny
11
22
ffeindio allan
11
22
ffeindio allan
11
23
taro ar
10
23
taro ar
10
17
torri i mewn
The most frequent PVs will be analysed in more detail in 4.5, with regards to their semantic properties and their presence in dictionaries of Welsh.
3.2.3.3. Syntactic categories
The distribution of syntactic categories of PVs in the corpus is presented in Table 9.
Table 9. The syntactic distribution of verb-particle constructions.
Tokens
Adverbial
phrasal verbs
676
Intransitive
511
Transitive
165
Phrasal
prep. verbs
269
Prep.
verbs
189
Total
1134
%
60%
45%
15%
24%
17%
100%
Types
271
171
100
73
54
398
68%
43%
25%
18%
14%
100%
%
The most frequent type of constructions were APVs, constituting about two thirds of both
tokens and types. The most frequent of those were: mynd ymlaen ‘go on’, ‘continue’,
‘happen’, ‘play’ (86 tokens), mynd heibio ‘go by’, pass’ (30), dod draw ‘come over’,
‘visit’ (29), cario ymlaen ‘carry on’, ‘continue’(17), edrych yn ôl ‘look back’, ‘review’
(16), eistedd i lawr ‘sit down’ (15), edrych ymlaen ‘look forward’ (14), symud ymlaen
‘move on’, ‘make progress’, ‘keep living’, ‘start something new’ (14), troi i fyny ’turn
up’, ‘come’, ‘appear’ (14), dod ymlaen ‘come on’, ‘hurry’, ‘progress’, ‘improve’, ‘get
on’, ‘have a good relationship’ (13), torri i mewn ‘break in’, ‘commit robbery’, ‘appear’
(12), bod ymlaen ‘be on’, ‘happen’, ‘be switched on’, ‘be scheduled’(12), rhoi ymlaen
‘put on’, ‘switch on’ (11), sortio allan ‘sort out’, ‘solve a problem’ (11), torri i lawr
‘break down’, ‘stop working’, ‘have a nervous breakdown’(11), tyfu i fyny ‘grow up’(11).
160
Transitive APVs – the position of the object
Verb + adverbial particle constructions were additionally divided into intransitive and
transitive. The purpose of this distinction was to examine the position of noun object in
transitive constructions due to inconsistencies in the literature described in 2.4.3.1.
The instances with pronominal direct object were excluded, since only the VerbObject-Particle is possible in such cases. Out of 114 concordances which contained transitive APVs with a direct object which was not a pronoun, 101 (89%) – followed the
Verb-Object-Particle order. Only in 14 sentences (11%) did the object follow the particle.
Three of these exceptional cases, shown in (36) - (38), represented archaic, literary
style: a quotation from the Bible, a quotation from a 19th century author and an archaically stylised satirical poem by a contemporary author. Noticeably, all three examples are
native PVs with the highly productive particle allan.
(36) a bwriodd allan bawb oedd yn prynu a gwerthu yn y deml (CRAC, 13)
‘and he drove out all those who were buying and selling in the temple’
(37) sgrech annaearol pan yn gosod allan arswyd y disgyblion (FN H/2014, 22)
‘unearthly scream while showing the fear of the disciples’
(38) estynnaf allan iddynt ffrwyth fy nghywilydd (LOL 46/2014, 6)
‘I extend to them the fruit of my shame’
Another three examples of Verb-Particle-Object order shown in (39)-(41) contained the
PV gosod allan ‘set forth, ‘display’, which appears to be centuries old; GPC gives 1585
as the date of its first recorded use and the phrase is used in Morgan’s Bible. This might
be the reason for the fossilised Verb-Particle-Object order used with this particular construction.
(39) tipyn o gamp wrth osod allan yr holl egwyddorion (Cas113/2015, 35)
‘a bit of a feat to set forth all the principles’
(40) i geisio gosod allan holl ffenomenau (Cas113/2015, 26)
‘trying to show all the phenomena’
(41) crynodeb o'n hadroddiad ymchwiliad (sydd yn gosod allan sylwedd y gŵyn)
(G 44/2014, 4)
‘a summary of our research report (which sets forth the basis of the complaint)’
161
Changes in the word order can indeed be demonstrated by comparing three versions of the Welsh Bible. The classic William Morgan’s version and 1988 Beibl Cymraeg
Newydd appear to prefer the object after the particle while the most recent beibl.net uses
the other order. The examples in (42) are translations of the sentence “Behold the Lamb
of God who takes away the sins of the world” (J 1, 29):
(42) a. Wele Oen Duw, yr hwn sydd yn tynnu ymaith bechodau'r byd. (Beibl William
Morgan)
b. Dyma Oen Duw, sy'n cymryd ymaith bechod y byd! (Beibl Cymraeg Newydd)
c. Dacw Oen Duw, yr un sy'n cymryd pechod y byd i ffwrdd. (beibl.net)
In two cases, (43) and (44), the moving of the object after the particle could be motivated
by the fact that the object is long.
(43) os ydi hynny'n amod i gadw allan y “Trydydd Rhyfel Byd” a oedd yn cael ei fygwth
yn y 1950au (G 45/2014, 21)
‘if this is the condition to keep away the “Third World War” that we were threatened with in the 1950s’
(44) rhoes heibio bob gobaith o gyfansoddi ei darn cerddoriaeth (AGB, 109)
‘she gave up all hope of composing her piece of music’
Four other examples of Verb-Particle-Object order came from dialogues in the novel Afallon. They might be characteristic of the dialect used in the novel:
(45) allan ni adel mas y politics? (AFALL, 306)
‘can we leave the politics out?’
(46) rwy'n pigo lan camera rhif tri (AFALL, 30)
‘I’m picking up camera number three’
(47) ro'n i wedi rhoi heibio bethau bachgennaidd (AFALL, 309)
‘I have put aside childish things’
(48) rwy i wedi rhoi lan syniade eraill (AFALL, 194)
162
‘I have given up other ideas’
The remaining two examples from press were also colloquial in tone:
(49) gwyliwch allan siopa (YD396, 9)
‘watch out for shops’
(50) wrth imi ddisgwyl ymhlith y cefnogwyr ffyddlon (hen ac ifanc) a socian i fyny'r
awyrgylch hafaidd (C 11.07.14, 17)
‘while I was waiting among the faithful supporters (old and young) and sucking
up the summer atmosphere’
In one odd case the Welsh PV, although directly translatable into English, followed different word order than the English phrase (51).
(51) gwylio'r teledu, talu'r morgais, newid y car lan, mynd am wyliau…. (AFALL,
288)
‘watch TV, pay the mortgage, change up the car, go on holidays…’
Altogether, the Verb-Particle-Object appeared in two types of registers: archaic and contemporary colloquial, mostly with multi-word object.
PPVs
There were 73 types of phrasal prepositional verbs in the corpus, constituting 18% of all
types (269 tokens – 24%). Out of these, 31 types were variants of APVs, e.g. rhedeg allan
o ‘run out of’ as a variant of rhedeg allan, cario ymlaen â ‘carry on with’ as a variant of
cario ymlaen. The other types had independent meanings.
The most frequent PPVs, with more than 5 tokens, were edrych ymlaen at ‘look
forward to’ (125 tokens50), edrych ymlaen i ‘look forward to’ (14), bwrw ymlaen â ‘continue with’ (8), dal i fyny â ‘catch up with’ (7), mynd ymlaen â ‘go on with’ (7), edrych
ymlaen am ‘look forward to’ (6).
50
The extremely high frequency of this PV should be attributed to the presence of many local newspapers
in the sample where the PV is used as a stock phrase (see 3.2.4.2).
163
The vast majority of PPVs in the sample (92%) were intransitive, with only 6 transitive constructions: cadw rhywbeth/rhywun allan o rywbeth ‘keep something/someone
out of something’, cael rhywbeth allan o rywbeth ‘get something out of something’, golchi rhywbeth i lawr â rhywbeth ‘wash something down with something’, gwneud
rhywbeth allan o rhywbeth ‘make something out of something’, setio rhywun i fyny
â rhywun ‘set someone up with someone’, tynnu rhywun i mewn i rywbeth ‘pull someone
into something’.
PPs
Idiomatic prepositional verbs constituted 17% of tokens (189) and 14% of types (54).
This category was characterised by higher than average proportion of native constructions/loan renditions (see next section) in comparison with other syntactic types. The most
frequent PPs were edrych ar ôl ‘look after’ (35 tokens), torri ar draws51 ‘interrupt’ (35),
dod ar draws ‘come across’ (30), taro ar ‘hit upon, bump into’ (10), dod dros ‘come over’
(of an emotion) (9).
3.2.3.4. Language contact categories and mode of narration
Table 10 presents the distribution of PVs in the corpus (distinguishing between different
senses of the same item) according to their status as borrowings. It can be seen that the
percentages of respective categories are comparable for tokens and types.
Table 10. The distribution of PVs in the corpus according to language contact categorisation.
Status
Native/loan
renditions
129
Loanblends
Pleonastic
Tokens
Directly
translatable
830
Total
59
Wholesale
borrowings
11
105
% of tokens
73%
11%
9%
5%
1%
100%
Types
263
61
45
26
3
398
% of types
66%
15%
11%
7%
1%
100%
1134
51
Similarly, it should be noted that high frequency of torri ar draws is due to the fact the expression was
often used in dialogues in novels after the words quoted.
164
The majority of PVs in the sample, 263 types (66%), were attributed to the directly translatable (DT) category as they have their exact English equivalents. The second most frequent category were loan renditions or items which are not directly translatable (N/LR),
with 61 types (15%). Loanblends (LB) constituted about one tenth of PVs, with 45 different types. Pleonastic items (PL) were less frequent, with 26 types (7%). Wholesale
borrowings (WH) were a marginal phenomenon, with only 3 types found (1%).
The distribution of items depending on the mode of narration was fairly even, with
666 (59%) tokens used in narrative mode and 468 (41%) in conversation 52. The differences were more marked within language contact categories. Native PVs and loan renditions were more prevalent in N53, whilst wholesale borrowings occurred almost exclusively in C (see below). In the other three categories there were no marked differences in
the distribution (Table 11).
Table 11. The distribution of PVs in narrative and conversation, according to language
contact categories.
Mode
% of
tokens
59%
N/LR
%
PL
%
CL
%
LB
%
WH
%
Narrative
No. of
tokens
666
94
74%
34
58%
491
59%
46
44%
1
9%
Conversation
468
41%
35
26%
25
42%
339
41%
59
56%
10
91%
Total
1134
129
59
830
105
11
Native PVs and loan renditions
Verb-particle constructions which did not have their one-to-one English equivalents were
attributed to the category native items/loan renditions (N/LR). Historical dictionaries,
GPC and OED were consulted in each case to check for obsolete English donor constructions. However, one cannot exclude the possibility that a Welsh PV might have been
calqued from an English item which is not recorded in dictionaries, for example due to
geographically restricted usage. Therefore, the observations presented below should be
treated with some caution.
In total, the sample contained 129 N/LR tokens and 61 types. There were several
cases of polysemous PVs with one calqued and one non-calqued meaning. For example
52
If one considered the narrative parts in the novel Afallon, written in a rather informal register, as C the
distribution would be exactly 50% (with 567 tokens in each category).
53
This figure should be treated as biased, however, due to the prevalence of native PP torri ar draws in
narrative parts of novels (see footnote 51).
165
dod ar draws was considered DT when meaning ‘come across’, ‘to find’, however in one
case it meant ‘get in someone’s way’ and thus was counted as N/LR.
There were relative few examples of APVs and PPVs likely to be native rather
than loan renditions. The most frequent of them (12 tokens) was bwrw ymlaen (â), already
mentioned by Rottet (2005:51). Five PVs contained the particle draw which is variously
translated into English and emerged to function as a native Welsh particle in non-calqued
PVs. The particle and its English equivalents will be described in detail in 3.3.2.10. Other
examples of native items are presented in Table 12.
Table 12. Examples of native APVs and PPVs in the corpus.
PV
brasgamu ymlaen
bwrw ymlaen (â)
codi allan
diferu heibio
estyn allan (at)
lledaenu i lawr
mynd heibio i
picio draw
rhygnu ymlaen
taro i mewn
Literal English translation
stride/lope forwards
throw/ hit/spend (time) on
rise out
drop by
stretch/extend out (towards)
spread down
go past
pick over (there)
rub/grate on
hit in
Meaning
develop rapidly
continue, carry on (with)
stir out of doors, be out and about
flow by (about time)
extend/reach out for
extend
exceed, go beyond
pop over, visit briefly
harp on
call in, pop in
The degree of idiomaticity of these items varies. Among more idiomatic expressions were
bwrw ymlaen, codi allan, rhygnu ymlaen. Semi-idiomatic items include: brasgamu
ymlaen, llaedanu i lawr, diferu heibio, taro i mewn. Some of these constructions appear
to be original metaphors created by individual authors, rather than set phrases, which
confirms the naturalness of such expressions in Welsh (52).
(52) bod agweddau tuag at bobl hoyw wedi brasgamu ymlaen dros y blynyddoedd diwethaf (Gwyl189/2014, 4)
‘that the attitudes towards gay people have loped/stridden forward in recent years’
(53) yn gwylio amser yn diferu heibio (LLYFR, 46)
‘watching the time dropping by’
166
Distinguishing between native PVs and loan renditions was a challenging task.
Some constructions might be classified as LR due to the fact they contain the same particle as the English equivalent. The similarity might be coincidental, however, when the
metaphorical meaning of the particle is salient (see 3.3). More obvious candidates for
loan renditions are constructions with dod ‘come’, which regularly replaces the English
‘get’ in cases where the verb indicates movement, while the usual Welsh equivalent of
‘get’, cael, does not have this connotation54. Thus, PVs such as dod dros ‘get over’, dod
drwodd ‘get through’, dod rownd rhywun ‘get around someone’, dod trwy ‘get through’,
dod ymlaen â ‘get on with, get along’ may be considered loan renditions. This phenomenon has been observed also by Rottet (2005: 52). The same concerns cases where dod
corresponds to English go: dod allan â ‘go out with someone’, ‘have a relationship’,
dod drwodd ‘go through something’, ‘examine something carefully’.
An evident loan rendition in the sample was llwytho i fyny, lit. ‘load something
up’, an equivalent of English ‘upload’. This recent coinage competes with calqued compounds such as lanlwytho, uwchlwytho55.
One outlying case was gwylio allan ‘watch out’, meaning “watch out for”. Here
the Welsh PV was classified as LR since it was identical with PPV English equivalent,
save from the fact that the Welsh construction did not use the second particle.
In some cases N/LR items appear to compete with calqued PVS. For example, the
sample contained an even number of tokens of bwrw ymlaen (32) and cario ymlaen (32)
(cf. Rottet 2005: 52), both meaning ‘carry on’. Similar examples were native PVs with
draw: cadw draw ‘keep away’ (3) vs. cadw i ffwrdd (1); galw draw ‘call in’, ‘visit’ (3)
vs. galw heibio (5) or galw i mewn (5), where in the second variant the particle is likely
to be calqued from English.
Considering the distribution of syntactic types in the N/LR category, out of the
61 types found, there were 32 PVs (52%), 19 PPs (31%) and 10 PPVs (16%). Noticeably,
the proportion of PPs within this category was almost two times higher than in general
distribution (see 3.2.3.3). The N/LR PPs are listed in Table 13.
54
The verb cael was found to be unproductive in forming PVs, since only three types with this verbs were
found: two PPVs with the particle allan: cael allan o ‘get out of’ (trouble), cael rhywbeth allan o rywbeth
‘get something out of something’, ‘benefit’ and cael rhywun yn ôl, get someone back’, ‘revenge’.
55
An entry for upload is not to be found in GPC, GA or GWE. Uwchlwytho is proposed by TSD, Termiadur
Addysg suggests llwytho i fyny, while Geiriadur Termau'r Coleg Cymraeg Cenedlaethol gives llwytho i
fyny, and also optional uwchlwytho; the entry also contains fynylwytho and lanwytho crossed out as unacceptable.
167
Table 13. Prepositional PVs (PPs) in the corpus classified as native items or loan renditions.
PP
Literal English translation
Meaning
cadw ar
colli ar
crafu â
keep on (someone)
lose on (someone)
scrape with (someone)
defend
lose your senses
butter someone up
dod ar draws
dod dros
dod trwy
galw â
gwneud â
llefaru ar draws
mynd ar ôl
taro ar
taro ar
taro ar draws
torri ar draws
tynnu am
tynnu at
come across (someone)
come over (something)
come through (something)
call with (someone)
do with (someone)
utter across (someone)
go after (someone)
hit on (something)
hit on (someone)
hit across (something)
break/cut across (someone)
pull on (something)
pull towards (something)
cross, get in someone’s way
get over, overcome
get through, analyse
call in, come round, visit
get on with, have a good relationship
interrupt
talk about
hit upon, stumble upon, find
bump into
hit upon, stumble upon, find
interrupt
get on for (about time or age)
turn (a certain age)
tynnu ar ôl
pull after
take after, resemble a family member
Possible loan
rendition (*)
*
*
*
*
*
*
*
Directly translatable items – calques and loanblends
The largest group of PVs – 830 tokens, 263 types – were assigned to the category directly
translatable (DT), while 105 items, with 45 types, belonged to the category loanblends
(LB). Items of both categories occurred with equal frequency in narrative and conversation mode. They constitute the majority of all PVs found (82%).
It should be emphasised that, despite the fact that items in the DT category can be
translated word for word into English retaining the same meaning, they are not necessarily
calques, especially in the case of semi-idiomatic PVs. Obviously, deciding whether an
item is a borrowing or whether the similarity is coincidental is a difficult, if not an impossible task, which would require extensive diachronic research going beyond the scope
of this thesis. However, some issues connected with transfer will be analysed in detail in
3.3, regarding the semantics of PV particles in order to shed light on the calquing phenomenon.
Wholesale borrowings
As mentioned above, wholesale borrowings were a marginal phenomenon in the sample.
Only 11 tokens of WH items were found, with 3 types: cym on ‘come on’, howld on ‘hold
on’ and ffyc off ‘fuck off’. All three are well-established set phrases, which is indicated
168
by their Welsh spelling. All WH tokens were used in the conversation mode, safe from
one example from the satirical paper LOL, which used slang in narrative.
Pleonastic PVs
Finally, there were 59 PVs identified as pleonastic constructions, with 26 types. There
were no major differences in distribution of these items between N and C modes. The list
of PL items is presented in Table 14. It contains two types of constructions: spatial, such
as cilio yn ôl ‘retreat back’ and aspectual, such as chwarae o gwmpas ‘play around’.
Transfer of meanings of aspectual particles will be described in more detail in section 3.3.
Table 14. Pleonastic PVs in the corpus.
PV
English translation
bacio yn ôl
back 'back'
chwarae o gwmpas
play around (play, have fun)
cilio yn ôl
retreat back
clirio allan
clear away
cloi i lawr
lock down
cychwyn allan
start out
dadlau ymlaen
argue on
danfon allan
send out
darllen allan
read out
disgyn i lawr
fall down
dod i fyny
come up (come)
dod i lawr
come down
dringo i fyny
climb up
dychwelyd yn ôl
return back
eistedd i lawr
sit down
esgyn i fyny
rise up
glanhau allan
clean out
helpu allan
help out
pasio heibio
pass by
picio i fyny
pick up (lift)
pori i lawr
graze down
torri i fyny
break up
tyfu i fyny
grow up
ymchwilio i mewn
research into
ysgrifennu i lawr
write down
ysgubo ymaith
sweep away
169
3.2.4. Stylistic markedness
The above results have presented largely quantitative data on the form and syntactic properties of PVs. This section will focus on the stylistics of Welsh PVs, offering a qualitative
analysis of the examples of their stylistic markedness in the corpus. The aim is to identify
constructions which are considered nonstandard and search for evidence of a norm against
PVs in fiction and the print media.
Following Zwicky’s (1978) definition, the term stylistic markedness is used here
to denote a way in which an element of a language can stand out or be remarkable by
being associated with a particular register, style, social dialect or regional variety of the
language; in other words a stylistically marked item is not usable by all speakers in all
contexts (Zwicky 1978: 29). Before examples of such items are presented, some remarks
will be made on the differences in frequencies of the PVs in the texts and their stylistic
feature in order to provide an appropriate context for the investigation.
3.2.4.1. Fiction
In the course of analysing the fiction sample, differences between the authors’ idiolects
have been observed as regards the number of verb-particle constructions used. The intuitive observations have been confirmed by a frequency count presented in Table 15. The
mean number of pages in each of the novels was estimated based on the number of words
in four randomly selected pages. The number was than multiplied by the number of pages
in the novel to estimate the total number of words. Next, the mean number of PV tokens
per 1000 words was calculated.
Table 15. Number of PVs per 1000 words in novels.
Title
Pages
Afallon
Adenydd glöyn byw
Craciau
Tair rheol anhrefn
Y Llyfrgell
327
343
182
258
246
Mean no. of
words/page
318
335
330
277
319
Words
103,986
114,905
60,060
71,466
78,474
PV tokens
210
207
82
71
75
PVs/1000
words
2.02
1.80
1.37
0.99
0.96
170
The results show significant differences in the number of PVs used by different authors:
Davies (2011) and Dafydd (2009) use approximately 1 PV per 1000 words, while in
Gruffydd (2012) and Roberts (2010) the number is twice as high. Differences were also
observed in the proportion of PVs used in dialogues (Table 16).
Table 16. Proportion of PVs used in dialogues in the analysed novels.
Title
Afallon
No. of PV
tokens
210
No.
of
PV types
77
No. of PVs
in dialogues
111
% of PVs in
dialogues
53%
Adenydd glöyn byw
207
73
144
70%
Craciau
82
29
30
37%
Y Llyfrgell
75
34
26
35%
Tair rheol anhrefn
71
12
40
56%
A closer scrutiny of the language and lexicon used by the authors may provide some
interesting observations regarding PVs.
The highest number of PVs occurred in Afallon (Gruffudd 2012), the only novel
in the corpus written in the first person. The novel’s narrator, Rhys, is a middle aged
Welshman who settles in Swansea after years of living abroad. The novel’s plot revolves
around his unsuccessful attempts to run a restaurant and unravel a murder mystery connected with the presence of American soldiers in Wales. The language of the narrative as
well as the dialogues is highly colloquial with frequent instances of code-switching and
numerous lexical borrowings. This may be seen as an intentional strategy of the author as
the novel is a political satire criticising the inertia and inability of Welsh people to deal
with various social and political problems. Rhys is an anti-hero criticising the decline of
traditional Welsh values, yet himself contributing to the situation. The character’s language “reflects the language of everyone around him” (personal correspondence with the
author), however the author intended to make the narrative parts “a bit more neural in
comparison to dialogue” (personal correspondence).
With regard to PVs, the novel contains a relatively low number of N/LR items and
a high number of recent calques and loanblends. Characteristically, the latter are used
both in narrative (e.g. cario ymlaen, setlo lawr, pigo lan, pasio heibio, snapio mas) and
dialogue (cario 'mlaen, cwlio lawr, ffeindio mas, ffwcio lan, gwatsio ar ôl, pitsho i mewn,
pwmpio lan ar, sortio allan/mas). It should be noted that some of these items appear in
171
conversations which “really” are held in English, since many characters in the novel do
not speak Welsh. The author seems to indicate this fact by extended use of borrowings,
calques and code-switching, but also by reducing dialectal elements in the speech of nonWelsh-speaking characters. All in all, the author’s lexical choices and the high number of
PVs might be influenced by factors such as: a) informality of the language b) use of
Swansea dialect in the narrative c) intentional or unintentional calquing from English in
dialogues featuring English-speaking characters c) intentional nonstandardness of the language related to the novel’s main theme of decline.
The other novel which made abundant use of PVs was Adenydd glöyn byw (Roberts 2010). The novel is set in rural North Wales and tells a story of a family of three
women: a grandmother, a mother, and a teenage granddaughter, Eira. The plot revolves
around Eira’s love affair with her young teacher, and the novel’s main theme are changing
attitudes toward gender roles, female sexuality and moral values. From the linguistic point
of view, the novel is characteristic for its rich dialogue and use of highly individualised
language, reflecting the dialect (with some protagonists coming from South Wales) and
the age of characters (a portrayal of three different generations).
As can be seen in Table 16 above, a significant proportion of PVs used in the
novel (70%) occurred in conversation mode. This is not only due to abundance of dialogue but also stylistic features of the author’s language, in particular the use of teenage
slang. Teenagers play an important role in the novel and the author strives to realistically
portray their speech. The author indicates that teenagers’ Welsh is heavily influenced by
English although they can “speak it properly” if they make an effort. In the following
excerpt Eira is first scolded by her mother for calquing the expression get some gen and
then self-corrects a grammatical mistake of omitting a prefixed pronoun and mutation to
indicate possession.
‘Adag honno dda’th o at Mrs Peters i ga’l dipyn o gen arna fi!’
‘Eira!’ ochneidiodd Rhiannon. ‘Dw i’n gwybod bod y stoncar ‘ma wedi dy gynhyrfu di’n
lân, ond tria gofio sut ma siarad Cymraeg, nei di?
‘Sorri, Mam. Dw i’n sgwennu fo’n iawn ‘sti. Tafod fi… fy nhafod i sy’n rhedag o flaen…
fy mhen i, ‘gorffennodd Eira’n ofalus. (AGB, 13)
[‘At that time he came to Mrs Peters to get some gen about me!’
‘Eira!’ sighed Rhiannon. “I know you are totally excited over this stonker, but try to remember how to speak good Welsh, will you?
172
‘Sorry, mam. I write it well, you know. It’s my tongue…my tongue running ahead…of my
head,’ finished Eira carefully.]
As a rule, the author portrays the language of teenagers by using stylistically
marked lexicon, of which PVs are an important part. The markedness is clearly visible in
the following fragment, in which the grandmother does not understand a calque used by
Eira:
‘Naethoch chi gopio off noson honno, Nain?
‘’Nes i be?
‘Ca’l cariad.’ (AGB, 22)
[‘Did you cop off that night, grandma?
‘Did I do what?
‘Make love.’]
Although PVs are used extensively by characters of all age groups in the novel, members
of the older generation are shown to use more N/LR items, such as cadw draw ‘stay away
from’, dod allan â ‘go out’, ‘have a date’, mynd drwodd ‘get through’, ‘finish’, while
young people use noticeably more loanblends (hitio off, patsio i fyny, setio rhywun i fyny).
There are also instances of wholesale borrowings and code-switching in their speech:
Cym on, Mam, spill the beans! (AGB, 167)
[Come on, mam, spill the beans!]
A rŵan, Mam, chill out am sbel, a wedyn newidia i fynd allan. (AGB,56)
[And now, mam, chill out for a while and then change to go out.]
Nominalisation of PV in the form of wholesale borrowings such as getawê (see 2.4.1) or
code-switching occurs as well, e.g.:
‘Be? Ti’n ffansïo fo rwan?
‘’Mond i lais o, siŵr dduw! Mae o’n ditsiyr, ‘tydi? Complete blydi turn-off!’ (AGB, 305)
[‘What? You fancy him now?’
‘Only his voice, surely! He is a teacher, isn’t he? Complete bloody turn-off!’]
‘Dan ni wedi ca’l gormod o getawê’n barod. Neith ‘yn lwc ni byth ddal.’ (AGB, 300)
173
[‘We’ve got away too many times [lit. We’ve had too much getaway] already. Our luck
can’t hold forever.’]
In contrast to the dialogues, the narrative contains only a small number of common
PVs, with no loanblends (except for setlo i lawr ‘settle down’, ‘sit comfortably’) or
wholesale borrowings. The author has a tendency to use pleonastic constructions with
back in the narrative, such as bacio yn ôl ‘back’ and cilio yn ôl ‘retreat back’.
Markedness of teenage language is also visible in the novel Craciau (B. Jones
2013), which follows about a dozen characters during a couple of dramatic days when an
earthquake takes place on the Isle of Anglesey. The novel has a dynamic plot and contains
much dialogue. The characters represent different ages and social classes and nearly all
of them speak the Anglesey dialect. Similarly to Adenydd glöyn byw, the differences in
the use of PVs by different characters are related to age: loanblends and wholesale borrowings are used by teenage characters (e.g. cnocio allan, sortio allan, ffyc off) while
N/LR items (e.g. dod draw) are characteristic of the speech of elderly people. The author
uses numerous PVs in the narrative as well, including loanblends such as cyrlio i fyny,
bordio i fyny, but mostly well-established constructions such as torri ar draws, mynd
ymlaen.
The two novels where noticeably fewer PVs were found are Y Llyfrgell (Dafydd
2009) and Tair rheol anhrefn (Davies 2011). Both novels are thrillers with elements of
a political satire, filled with action and humour. While Tair rheol anhrefn features plenty
of conversation, Y Llyfrgell uses relatively little dialogue in comparison to the other novels. In both books the majority of characters speak the Ceredigion dialect. Both authors
use a relatively small set of PVs and avoid loanblends. The only two LB items in Tair
Rheol occur in the speech of a bus driver, e.g.:
Sortiwch e i gyd mas ymysg eich gilydd. (TAIR, 38)
[Sort it all out among yourselves.]
In fact, the very short scene with the bus driver contains 4 out of 12 PV types used in the
whole novel. This may suggest that for the author calqued PVs are stylistically marked as
belonging to the language of uneducated people (in contracts to most characters in the
book who hold a university degree). Similarly, in Y Llyfrgell the majority of LBs and
evident calques occur in dialogues between a porter and canteen staff in the library, once
174
again associating this type of language with the working class. Tair rheol anhrefn features
a number of characters who are not Welsh-speaking, but in contrast to Afallon, their language is not marked with contact-induced elements.
In the narrative, both authors appear to avoid loanblends and obvious calques from
English. The narration in Y Llyfrgell is characteristic for using original metaphorical PVs,
such as diferu heibio ‘drip by’ or llosgi i mewn ‘burn in’.
3.2.4.2. Press
Similarly to the novels, differences in distribution of PVs and their categories were observed across magazines and newspapers in the sample. Table 17 presents the number of
tokens in each issue and the number of PVs per page. These numbers should be treated
with caution since the average amount of text per page varied in each source. For the
same reason, it is extremely difficult to calculate the number of words in the sample without digitalising the whole corpus. Notwithstanding these difficulties, the available data
demonstrate some general tendencies, confirmed by qualitative analysis presented below.
Table 17. Frequency and distribution of PVs in the press sample.
Title code
Tokens
C tokens
Pages
PVs/page
DT
N/LR
LB
WH
PL
Weekly magazines/newspapers
C 18/7
29
6
20
1.5
18
4
2
0
5
C 8/8
25
8
20
1.3
20
5
0
0
0
G 45
33
21
27
1.2
23
4
0
0
6
C 25/7
24
1
20
1.2
19
3
2
0
0
G 46
39
23
35
1.1
32
1
4
0
2
G 44
30
18
27
1.1
21
2
6
0
1
G 47
34
21
35
1.0
28
1
1
0
4
C 11/ 7
14
3
20
0.7
10
2
1
0
1
136
83
124
1.1
104
8
11
0
13
92
18
80
1.2
67
14
5
0
6
total Golwg
total Y Cymro
Local bulletins (papurau bro)
YD
35
1
28
1.3
26
4
3
0
2
YA
16
1
16
1.0
11
2
3
0
0
CL
25
1
28
0.9
22
1
0
0
2
CW
15
0
20
0.8
13
2
0
0
0
BG
20
1
28
0.7
19
0
1
0
0
175
LLA
10
0
16
0.6
10
0
0
0
0
T
12
0
24
0.5
9
1
1
0
1
Monthlies, quarterlies, yearlies
LOL
37
4
24
1.5
23
3
7
2
2
Gwyl
20
0
16
1.3
16
3
0
0
1
W
26
7
42
0.6
18
4
2
0
2
Cas
13
0
28
0.5
10
3
0
0
0
FN
15
0
60
0.3
12
3
0
0
0
6
0
24
0.3
4
2
0
0
0
11
1
50
0.2
11
0
0
0
0
CR
BARN
The table presents the total number of tokens found in each issue as well as the number
of tokens used in conversation, i.e. direct quotations or interviews. The right-hand side of
the table demonstrates the distribution of PVs across language contact categories.
The first part of the table compares the use of PVs in two Welsh weeklies – the
magazine Golwg and the newspaper Y Cymro. Both media use the semi-formal register
of Welsh; the level of formality depends also on the type of texts, with sport news and
individual columns tending to use more colloquial language. PVs also appeared more
frequently in this type of texts. The language of Golwg appears to be more informal than
that of Y Cymro as regards grammatical forms, however when it comes to lexicon, major
differences are not visible. The mean number of PVs per page was almost even in the two
publications. As can be seen from the second column of the table, the proportion of PVs
used in conversation was much higher in Golwg. This is due to the fact that the popular
magazine contains plenty of interviews in contrast to Y Cymro, which as a quality newspaper focuses on facts and opinions. Many instances of calqued PVs found in Golwg seem
to stem from word-for-word translations of interviews conducted in English, as in:
(…) does dim angen i fusnesau lleol gredu bod y "dref yn colli mas" G 46, 5
[there is no need for local businesses to believe that the “town is missing out”]
The second group of publications were local community bulletins, known as papurau bro. The style used in this type of publication varies from semi-formal to informal,
with a tendency to use traditional literary verb forms, such as impersonal and conjugated
past habitual. This may be attributed to the fact that authors of articles often belong to the
older generation. A slightly more conservative approach can be seen in the vocabulary
176
although it should be remembered that since the bulletins generally contain few interviews, most of the items occurred in the N mode. Most PVs found in papurau bro were
directly translatable into English and practically no LB or WH items were used. Out of
133 tokens in all papurau bro, there were 48 PV types. Curiously, 55 tokens (over 40%)
were instances of edrych ymlaen due to the fact that “looking forward” to future events is
a conventional phrase used in most reports about local community life.
Much variation can be seen in the Welsh magazines analysed. Three of them:
quality opinion magazines Barn and Y Faner Newydd, and the religious bulletin Cristion
contained the lowest numbers of PVs, 0.2-0.3 per page, with no LB or WH items. The
register used of these publications was found to be quite formal both in terms of grammatical forms and lexicon, which may explain the avoidance of calques from English.
The PVs used were mostly well-established inconspicuous items, many of them semiidiomatic. Evidence for a prescriptive norm against the English element was found particularly in Y Faner Newydd. In contrast to other Welsh periodicals, the magazine appears
to have a policy of avoiding citing in English as in the following example:
Y geiriau 'Croeso i bawb o bob oed' (yn Saesneg wrth gwrs) wnaeth ein denu yno mae’n
debyg. (FN, 19)
[What probably attracted us to come there were the words “Welcome to every one of all
ages” (in English of course).]
One of the articles contained a strong-worded passage against using Welsh “corrupted” by English words and phrases:
It is irresponsible and inexcusable to allow using corrupted language – language of errors,
full of English words and phrases (…) Children and young people will accept bratiaith as
standard and consider it a pattern (…) Like other languages, Welsh has different registers
but bratiaith is not one of them, (…) by allowing to use bratiaith, the service contributes
to the process of undermining Welsh words and phrases with English words and phrases. 56
(FN, 3)
Only 15 PV tokens were found in the 60 pages of dense text in this magazine,
two of them being quotations from a 19th-century book. The avoidance of constructions
Gweithred anghyfrifol ac anesgusodol yw caniatáu defnyddio iaith lygredig - iaith wallus ac iaith sy'n
frith o eiriau ac ymadroddion Saesneg (...) Byddai plant a phobl ifanc yn derbyn y fratiaith fel safon, ac yn
ei hystyried yn batrwm. (...) Fel ieithoedd eraill, mae gan y Gymraeg wahanol gyweiriau, ond nid yw bratiaith yn un ohonynt. (...) Drwy ganiatáu defnyddio bratiaith mae'r gwasanaeth yn cyfrannu i'r broses o
ddisodli geiriau ac ymadroddion Cymraeg gyda geiriau ac ymadroddion Saesneg.
56
177
coming from English is associated with the character of the magazine who is highly supportive of Welsh-language activism and thus in general manifests a purist language ideology.
A moderate number of PVs was used in organisation bulletins Y Wawr and
Casglwr. Similarly to papurau bro, the language used in these magazines was quite formal in terms of grammar. As a literary magazine, Casglwr used also more formal vocabulary and appeared to avoid PVs in general and loanblends in particular (0 tokens). Avoidance of direct calquing can be seen in the following example, in which the phrase “tocking
away” used by Dylan Thomas is translated with an idiomatic Welsh expression rather
than a calqued PV:
Yn y cyswllt hwn mae'r ymadrodd 'tocio'r byd i'r byw' ('tocking the earth away’) yn arwyddocaol iawn. (Cas, 4)
[In this context the phrase ‘tocking the earth away’ is very significant.]
A slightly higher number of PVs was found in the magazine Y Wawr, the bulletin
of Merched y Wawr ('Women of the Dawn'), a women’s movement which promotes the
Welsh language as well as doing charity work and organising social events in local clubs
across Wales. Due to its educational role, the organisation’s magazine uses semi-formal
language with numerous traditional literary forms. As far as the PVs are concerned, most
items which could be seen as calques from English occurred in interviews. In one of them,
a member of Merched y Wawr made a comment about the need to accept changes which
take place in the language:
How can the Welsh language be made ‘cool’?
You have to understand that language is not a static thing. Its’s alive and it changes all the
time. So we have to accept the changes and the way young people use it. In the end it’s the
use of language that keeps it alive. (W, 31)57
57
Sut mae gwneud yr iaith Gymraeg yn ‘cool’?
Ma raid deall nad yw iaith yn beth statig. Mae'n fyw ac yn newid a thyfu o hyd. Felly ma' rhaid i
ni dderbyn y newidiau yn y defnydd ohoni gan bobl ifanc. Y defnydd o i'r iaith gydd yn ei chadw'n
fyw yn y pen draw.
178
The highest number of PVs occurred in the satirical magazine LOL. This was the
only magazine with a significant number of loanblends and the only one in which wholesale borrowings occurred. This is hardly surprising, bearing in mind the character of the
publication and the fact that is makes extensive use of slang. The magazine is an illustrative example of stylistic markedness of certain PVs. The most striking text in that respect
is a fake diary of Prime Minister Carwyn Jones, who is criticised by some for his reluctance for speaking Welsh and lack of support for the language. As the text mocks the
politician, the diary is intended to be written in very “bad” Welsh. Alongside numerous
borrowings and calqued idioms, the one-page-long text contains as many as 10 PVs,
which are evident calques from English or loanblends, e.g.
er i fi sefyll lan dros y diawl (LOL, 8)
[although I stood up for the bastard]
nyttyr llwyr sy'n ffwcio addysg lan ar draws y wlad (LOL, 8)
[total nutter who fucks education up across the country]
o'n i'n ishws yn dishgwl mlân at rownd neu ddou o golf (LOL, 9)
[I used to look forward for a round or two of golf]
mae'n rhaid i fi sorto hwnna mas (LOL, 8)
[I’ve got to sort that one out]
A similar text in LOL mocks the Welsh language commissioner Meri Huws,
whose “letters” also contain plenty of calqued constructions such as sorto pethe fel hyn
mas ‘sort things like that out’, chi'n gwybod sut i sretsho pethe mas ‘you know how to
stretch things out’. (LOL, 7)
One more example of humorous criticism towards calquing was found in one of
the issues of Golwg, in a column which presents examples of bad translation of public
signs into Welsh (known as “Sgymraeg”). The column shows a picture of a poster with
the sentence “Mae Mehefin yn archebu allan bob dros!”, a word-for-word translation of
“June is booking out all over”.
In two cases, PVs were marked with inverted commas. In the first example, the
phrase come out in the context of sexual orientation is relatively new, which can explain
its markedness in Welsh.
179
a'i longyfarch am ei ddwerder yn 'dod mas' a datgelu ei wir rywioldeb (G 46, 10)
[and congratulate him on his courage to 'come out’and reveal his true sexuality]
In the second example, the author of an article about a horse exhibition seems to
ironically underline the phrase dod ymlaen, which is one of the most polysemous PVs in
the corpus. In this instance it is probably a loan rendition of get on or come along in the
sense ‘succeed’. For the author, the item appears to be a token of ‘bad’ Welsh.
(…) gerddodd gŵr ifanc heibio ac mi ofynnwyd iddo sut yr oedd wedi 'dod ymlaen' gyda'i
geffyl. Atebodd yntau yn ei Gymraeg gorau “Seventeenth”, a oedd yn ddarlun yn ei hunan
o safon a dyfnder y sioe. (C 8/8, 18)
[a young man walked by and he was asked how he was ‘getting on’ with his horse. He
answered in his best Welsh “Seventeenth”, which was in itself a picture of the standard and
deepness of the show]
The above data demonstrate that the vast majority of PVs are well-established in semiformal varieties of Welsh. The cases of stylistic markedness were rare and concerned
those PVs which were associated with slang, in particular teenage speech. There was little
evidence of prescriptive norm against these constructions in press, but they were sometimes seen as tokens of ‘bad’ Welsh if seen as careless and purposeless calquing from
English.
3.3. Correspondence of meanings – a comparative analysis of particle semantics
The corpus analysis has provided evidence for different levels of correspondence in
meaning between Welsh and English phrasal verbs. The vast majority of items were directly translatable into English. However, the role of contact-induced mechanisms in the
emergence of these constructions cannot be always taken for granted. Only in the case of
loanblends and wholesale borrowings does the form of the PV evidently point to transfer
from English. For other categories, correspondence with English does not necessarily
prove direct influence of language contact. On the one hand, it is possible that an idiomatic extension of particle meanings might have occurred in Welsh independently of
English. As claimed by Rottet, idiomatic uses of phrasal verbs can be found as early on
as in medieval texts (2005:44) and he has assumed that the presence of untranslatable
items such as bwrw ymlaen indicated that they developed independently from English
180
(2005:50-51). Although the claim seems logical, it could just as well be assumed that
such constructions were modelled on the basis of those already borrowed from English.
This question is very difficult to resolve for the lack of diachronic data and theoretical
frameworks for examining correspondences of meaning in cases of loan translations. Diachronic cognitive linguistics are a relatively new field of research focusing around prototype theory and directionality of semantic extension and change (see Winters 2010) and
methods for comparative studies dealing with the emergence of loan translations have not
been developed so far. An in-depth diachronic analysis is certainly beyond the scope of
this study.
However, what is of some interest here are the mechanisms of transfer involved
in the emergence of Welsh idiomatic PVs and the stylistic markedness of these items. The
level of one-to-one correspondence between Welsh and English PVs can hardly be coincidental. Moreover, the above analysis has shown that some directly translatable items
with native Welsh verbs, (e.g. sefyll i fyny dros rywun ‘stand up for someone’, archebu
allan ‘book out’) can be stylistically marked in the Welsh written language as calques
from English, while others with the same particles (e.g. darllen allan ‘read out’, dal i fyny
‘catch up’) are not. Therefore, one can assume that the meaning of the particle is a potential factor at play regarding the integration of Welsh PVs. It can be hypothesised that
some particle senses are considerably better integrated in Welsh than others as this has a
bearing on the markedness of a given PV. Therefore an analysis of semantic correspondence of particles between Welsh an English might provide valuable information on the
nature of PVs in Welsh relevant for the planned general description of this linguistic phenomenon.
Hence, in this section I would like to examine the meaning of the 28 PV particles
identified in the corpus study and compare them with their English equivalents. The particles will be examined one by one using a cognitive model of particle semantics based
on Rudzka Ostyn (2003). The analysis will pay attention to the frequency of different
particle senses in the corpus and their earliest attestation in the historical dictionary of
Welsh Geiriadur Prifysgol Cymru. Additional examples of PVs attested in the dictionary
will be also sought as they could provide more evidence for the evolution of particle
meanings. Of course such an approach has its limitation. For one thing, as often noted
181
(e.g. by Poplack 2017: 43) borrowed items may exist for a long time without being attested in a dictionary. This is especially true of phrasal verbs, which might be omitted on
purpose because of the authors’ prescriptive approach.
Notwithstanding these limitations, there are a few issues the analysis may help to
investigate. Firstly, there is the question of the capability of Welsh to create idiomatic
PVs on its own. It would be useful to look beyond single items and search for particle
senses which do not have their English counterparts. The presence of such senses would
provide stronger evidence that Welsh can develop idiomatic PVs independently from
English. Secondly, comparative data can be used to examine the integration of some PVs
in the linguistic norm. One can assume that if a given sense appears only in constructions
which are directly translatable into English and additionally are not attested in the historical dictionary, there are grounds to assume that this sense is new and has been transferred
from English. This will help identify the most marked particle senses and those which are
unique for Welsh. It will be interesting to compare the results of the analysis with the
representation of phrasal verbs in general dictionaries and teaching materials analysed in
the next chapter.
3.3.1. A cognitive model for analysing the semantics of PV particles
The cognitive semantic model used for the analysis takes a ‘meaningful particle’ approach
which stands in opposition to the common perception of PVs as highly idiomatic items in
which the choice of particle is arbitrary and the meaning unpredictable. Particles, semantically derived from prepositions, have been studied within Cognitive Linguistics for several decades (Lindner 1981; Lakoff 1987; Zelinsky-Wibbelt ed. 1993; Boers 1996; Navarro i Ferrando 1998; Tyler and Evans 2003; Rudzka-Ostyn 2003). Within the cognitive
model, prepositions are understood as semantic networks of related senses with a core, or
basic, meaning which is related to the cognitive domain of physical space. This lexical
meaning is then extended towards more and more abstract, metaphorical senses (Porto
Requejo and Pena Díaz 2008: 113). Therefore, at the conceptual level the form of a particle does not represent a single meaning, but a set of different but related meanings,
which together form a semantic network (Tyler and Evans 2003: 7).
182
A major work on the semantics of particles in English PVs is Rudzka-Ostyn
(2003). Untypically, the book is practical rather than theory-oriented, as it is intended to
serve as a textbook for students of English. Despite being written in a simple manner, the
work contains a comprehensive list of English PV particles, with a network of meanings
assigned to each of them and each sense being illustrated by a large number of examples
of commonly used PVs. This work has been chosen as the basis for the analysis of possible transfer or extension of senses in Welsh PVs.
The basic cognitive grammar terminology used in the following section is based
on Langacker (1991), Navarro i Ferrando (1998) and Rudzka-Ostyn (2003). A particle is
a linguistic unit, that is “a thoroughly mastered structure, i.e. one that a speaker can activate as a preassembled whole without attending to the specifics of its internal composition" (Langacker 1991: 15). Every unit consists of a phonological and semantic pole. The
semantic pole of a unit is the semantic structure or predication (Langacker 1991: 553)
and it is represented in capital letters. Thus, for example, the lexical unit allan corresponds
to the semantic structure ALLAN. In the case of PVs, the predications are relational,
which means that they require two entities. An entity is “a maximally general term for
anything one might conceive of or refer to for analytical purposes” (Langacker 1991:548).
The two relational entities are called the trajector and the landmark. The trajector is the
entity which is foregrounded or focused on, while the landmark is the container or surface
providing the frame of reference or background to the trajector (Navarro i Ferrando
1998: 19; Rudzka-Ostyn 2003: 9).
The subsequent sections of this chapter will examine semantic structures of the
28 Welsh particles found in the corpus study, comparing them with the corresponding
English structures described on the basis of the respective sections in Rudzka-Ostyn
(2003). Examples of PVs relevant for each sense will be drawn from the corpus and also
from the historical dictionary of Welsh, Geiriadur Prifysgol Cymru, online version
(GPC). Each item from the corpus will be also cross-checked with GPC, in an attempt to
trace the evolution of various meanings. The date of the earliest recorded use of an item
will be given in brackets with the abbreviation “GPC”. Occasionally, the dictionary includes an item without a date, but a label Ar lafar, ‘spoken’. This will be indicated by:
(GPC, “spoken”). Lack of the label next to a PV indicates that the construction is not
included in GPC. The particles will be analysed in the order of their frequency in the
corpus (cf. Table 7, 3.2.3.1).
183
3.3.2. The particles
3.3.2.1. ALLAN and OUT
The Welsh adverbial particle allan (mas in southern dialects of Welsh) corresponds to
English out. Allan was the most frequent particle in terms of the number of PV types. Out
of 100 types containing this particle, 80 were identified as DT or LB, 6 as PL and 14 as
N/LR items. Concerning the number of tokens, allan was the second most frequent particle after ymlaen, with 167 tokens.
Rudzka-Ostyn distinguishes 6 meanings within the predication OUT. They are
summarised in Figure 3.
Figure 3. Network of meanings of OUT according to Rudzka-Ostyn (2003:41).
The basic meaning of OUT is spatial and includes the concept of an entity moving
out of a container (Rudzka-Ostyn 2003: 14). This corresponds to sense 1b. of allan in
GPC: “indicating movement, direction or extension to the point beyond an external border
of a place, building etc. (…) also figuratively” [translations of all definitions from GPC
mine – ML]. Idiomatic Welsh PVs conveying this meaning found in the corpus were
numerous, for example: gadael allan ‘leave out’, ‘omit’ (GPC 1595), troi allan ‘turn out’,
‘expel’ (GPC 1604-07), tynnu allan ‘take out’, ‘withdraw money’, agor allan ‘open out’,
184
bwrw allan ‘cast out’, ‘expel’ (GPC 1551). They could also be formed with borrowed
verbs, e.g. sticio allan ‘stick out’ (GPC 20th c.). All these items translate directly into
English, but some PVs without English word-for-word equivalents were found as well:
codi allan lit. ‘rise out’, ‘stir out of doors’, ‘be out and about’ (GPC “spoken Northern”)
and gosod allan ‘set out or forth’ (GPC 16th c.). GPC contains more relevant examples,
with the earliest, sefyll allan ‘stand (up) an come out’, dated to the 13th century. Others
include: mynnu allan, lit. ‘insist out’, ‘draw out’ – c. 1400, taflu allan ‘throw out’ –
c. 1400, cau allan ‘close out’ – 1651, camosod allan, ‘misrepresent’ – 1688. The most
recent items are logio allan ‘log out’ (GPC 1999) and clocio allan ‘clock out’
(GPC 2003).
Within the second meaning of OUT the landmark is home or social space
(Rudzka-Ostyn 2003: 18): it occurs in the context of leaving home to see other people
(e.g. go out) or social exclusion (e.g. throw someone out). PVs with allan conveying the
socialising aspect were for instance mynd allan and dod allan, both meaning ‘go out’ (to
see someone). The sense of social exclusion can be attributed to phrases such as troi allan
‘turn (someone) out’, ‘expel’ (GPC 17th c.), cwympo allan ‘fall out (with someone)’,
‘argue’ (GPC 17th c.).
The third meaning involves choosing something out of a group or set of things
(Rudzka-Ostyn 2003: 19). Items belonging to this category found in the corpus were:
cyfrif allan ‘count out’, ffeindio allan ‘find out’ (GPC 18th c.), pwyntio allan ‘point out’
(GPC 18th c.), sefyll allan ‘stand out’ (GPC 16th c.), sortio allan (GPC “spoken”). Other
GPC examples not present in the corpus are: pennu allan, lit. ‘appoint, determine out’,
‘single out’ (GPC 1733), penodi allan ‘point out’(GPC 1759), pigo allan ‘pick out’ (GPC
1604), sawru allan ‘sniff out’ (GPC 1815).
In the fourth sense, bodies, minds and mouths are viewed as containers (RudzkaOstyn 2003: 19) out of which something, e.g. sound, is released. Welsh examples from
the corpus include: darllen allan ‘read out’, cadw allan o ‘keep out of’, estyn allan ‘reach
out’, galw allan ‘call out’, gweddi allan ‘shout out’, llenwi (llanw) allan ‘fill out’, ‘put
on weight’. Although none of these PVs has been found in GPC, the dictionary contains
at least two examples of similar constructions: saethu allan ‘to ejaculate (of prayer)’
(GPC 1599) and dysgu allan ‘learn off, learn by heart’ (GPC 1913, “spoken Northern”).
In the fifth meaning the landmark is a state or situation (Rudzka-Ostyn 2003: 22)
and the particle may convey change of state, e.g. from possession to lack of possession
185
(be out of something), from consciousness to non-consciousness (knock out). Welsh examples found in the corpus include: dod allan ‘come out (of trouble)’, bod allan o ‘be out
of (something)’, cnocio allan ‘knock out’, ‘make unconscious’, rhedeg allan ‘run out (of
something)’ (GPC “spoken”), tynnu allan ‘withdraw’, optio allan ‘opt out’. Apart from
rhedeg allan, none of these items is found in GPC. However, the dictionary contains at
least two PVs expressing this sense: chwythu allan ‘blow out’, ‘extinguish’ (GPC 1606)
and taeru allan lit. ‘claim out’, ‘refute, argue against’ (GPC 1716). “Being out” in the
sense of “being extinguished” (of fire, light) is distinguished as sense 5. of out in GPC
(1790, “spoken”).
The sixth meaning is related to states of non-existence, ignorance or invisibility,
which metaphorically function as containers i.e. the particle conveys the sense of something starting to exist, appearing, becoming known (Rudzka-Ostyn 2003: 25). The examples in the corpus include: dod allan ‘come out’, ‘emerge’, gweithio allan ‘work out’
(GPC 17th c.), gwneud allan ‘make out’, helpu allan ‘help out’, rhoi allan ‘publish’
(GPC 16th c.), torri allan ‘break out’, ‘start’ (GPC “spoken”), troi allan ‘turn out’ (GPC
19th c.). These examples correspond to sense 2. of out according to GPC: “out (of story,
etc.), abroad, in (to) the public arena, in public, publicly, openly.” Quotations of this sense
of out cited span from the 14th century to modern spoken “coming out” referring to homosexuals.
The final, seventh meaning is related to an increase to maximal boundaries i.e. it
conveys a sense of completeness of an action (Rudzka-Ostyn 2003: 32). Only four examples of this sense were found in the corpus: clirio allan ‘clear out’, glanhau allan ‘clean
out’, archebu allan ‘book out’, stretcho allan ‘stretch out’. Characteristically, three of
them were stylistically marked, belonging to teenage language or used for the purpose of
satire (see 3.2.4 above). None of these examples can be found in GPC. The only item with
a similar sense of the particle found in the dictionary was golchi allan ‘purge soluble
matter away, wash out’ (GPC 20th c.)
The above data show that allan is a highly productive particle in Welsh as it occurs
in evidently native constructions recorded many centuries ago as well as very recent expressions, in many cases calqued from English. The extension of meaning of ALLAN
from spatial containers to social, physical and abstract dimension appears to historically
follow the same path in Welsh as it does in English. However, the most metaphorical
186
meaning of completeness appears to be recently borrowed and consequently is still quite
strongly marked.
3.3.2.2. I FYNY and UP
The adverbial particle i fyny corresponds to English up, with the basic meaning “indicating movement to a higher place or position” (GPC). I fyny was the second most frequent
particle in the corpus as regards the number of types (54), with 6 PL constructions,
11 LBs, 35 DTs and 2 N/LR items. With regard to token number, i fyny was the third most
frequent particle (96 tokens).
Rudzka-Ostyn distinguishes 5 meanings of the predication UP, summarised in
Figure 4.
Figure 4. Network of meanings of UP according to Rudzka-Ostyn (2003: 103).
The basic, spatial meaning of UP/I FYNY, related to movement to a higher position or
being in a higher position occurs usually in non-compositional constructions, however
some metaphorical uses of the particle in this sense were found in the corpus e.g. golchi
i fyny ‘wash up (about sea)’. Other examples of figurative PVs within this sense found in
GPC are bwrw i fyny (1588) ‘throw up’, ‘vomit’, eistedd i fyny ‘sit up’ (1588), taflu i fyny
‘throw up’, ‘vomit’ (1849).
187
The sample also contained pleonastic constructions such as: dringo i fyny ‘climb
up’, esgyn i fyny ‘rise up’, picio i fyny ‘pick up’ (GPC 18th c.), tyfu i fyny ‘grow up’ (GPC
1691). Out of these only tyfu i fyny appeared more than once (11 tokens). Like English
UP, I FYNY may also indicate moving in the northern direction (Rudzka-Ostyn 2003:
80), as in dod i fyny, ‘come up’, ‘come to the North’.
The first non-spatial meaning of UP described by Rudzka-Ostyn (2003:77) is connected with the trajector aiming at or reaching a goal, an end or a limit. Examples from
the corpus were mostly directly translatable items: galw i fyny ‘call (someone) up’, dal
i fyny ‘catch up’, rhoi i fyny ‘give something up’, ‘resign’ (GPC 16th c.), cadw i fyny
‘keep up’, cwrdd i fyny ‘meet up’, including loanblends setio i fyny â ‘set (someone) up
with (someone else)’, picio i fyny ‘pick (someone) up’. There was also one loan rendition:
llwytho i fyny ‘upload’ (see 3.2.3.4). Another recently coined PV of this type found in
GPC is bacio i fyny ‘back up (a computer file)’ (20th c.). Altogether, this sense of i fyny
appears to be recent.
The second meaning involves an entity moving to a higher degree, value or measure (Rudzka-Ostyn 2003: 80). Examples of Welsh PVs conveying this sense from the
corpus are: chwythu i fyny ‘blow something up’, ‘exaggerate’, dal i fyny â ‘catch up with’,
‘reach the same standard’, newid i fyny ‘change up’, gwisgo i fyny ‘dress up’ mynd i fyny
‘go up’, ‘progress’, symud i fyny ‘move up’, ‘make progress’, camu i fyny ‘step up’, ‘progress’. None of these are to be found in GPC, however the dictionary contains several
other examples: cymryd i fyny ‘take up’, ‘assume’ (GPC 1658), dwyn i fyny ‘bring up’
(GPC 1588), traenio i fyny ‘train up’ (GPC 1696), and also troi i fyny ‘turn up (sound)
(GPC “spoken”).
Next, akin to English UP, I FYNY may indicate that an entity becomes more visible, accessible or known (Rudzka-Ostyn 2003: 85). This sense was found in the following Welsh PVs: sefyll i fyny dros ‘stand up for’ (GPC 1783), siarad i fyny ‘speak up’,
‘make your point’, dod i fyny ‘come up’, ‘appear’, rhoi i fyny ‘put up’, ‘raise to make
more visible’, picio i fyny (pigo lan) ‘pick up’, ‘learn’, popio i fyny ‘pop up’, ‘appear’,
troi i fyny ‘turn up’, ‘come, appear’ (GPC 1858, “spoken”). The PV dod i fyny was stylistically marked in satire (see 3.2.4.2).
The final meaning distinguished conveys the sense of completeness, reaching the
highest limit of something or covering an area completely (Rudzka-Ostyn 2003: 86).
Welsh PVs expressing this sense are gwneud i fyny am ‘make up for’, ‘compensate’,
188
torri i fyny ‘break up’, ‘break into pieces’, chwythu i fyny ‘blow up’ (GPC 1675),
rhoi i fyny â ‘put up with’, ‘tolerate’ (GPC 1855), and a number of loanblends: bordio i fyny ‘board up’, cyrlio i fyny ‘curl up’, ffwcio i fyny ‘fuck up’, pacio i fyny ‘pack up’,
patsio i fyny ‘patch up’, pwmpio i fyny ar ‘pump up on’, socian i fyny ‘soak up’. Many of
these examples belonged to the informal registers, they were found in teenage speech or
dialect. Other examples from GPC are: cau i fyny ‘close up’ (GPC 1588), cloi i fyny ‘lock
up’ GPC 1658, llosgi i fyny ‘burn up’ (GPC 1658), llyncu i fyny ‘swallow up’ (GPC 1744),
nyddu i fyny lit. ‘spin up’, ‘draw to an end’ (GPC 18th c.), pacio i fyny ‘pack up’, ‘finish’
(GPC 1853), torri i fyny ‘go bankrupt’ (GPC 1743) and also topio i fyny ‘top up’ (GPC
“spoken”).
In contrast to out, there are no examples of idiomatic usages of i fyny recorded
before 1588. The role of English interference in the development of these items seems
probable although most metaphorical senses of the particle are not recent, with single
examples dating back to 17th and 18th centuries. GPC mentions this aspect in the definition of i fyny: ‘also used with certain verbs to translate English idioms’.
3.3.2.3. YMLAEN, RHAGDDO, AR and ON
In the majority of cases, the English adverbial particle on corresponded to Welsh adverb
ymlaen, denoting “forward, on(ward), ahead, in (the) front; on(ward) (of time), in the
future; on (denoting continuation of action, etc.); also fig.” (sense a. adv. of ymlaen,
GPC). Ymlaen was the most frequently used particle in the corpus with regard to number
of tokens (369) and third with regard to number of types. There were 49 PV types with
ymlaen in the corpus, with 24 different verbs. 7 types were identified as N/LR items, one
as PL, 37 as DTs and 4 as LBs. Only one PV type with a particle corresponding to the
adverbial on contained a different particle, rhagddo. Rhagddo is the 3rd pers. sing. masc.
inflected form of the preposition rhag. It may be synonymous to ymlaen, indicating progress in time. The item found in the corpus was mynd rhagddo ‘go on’ (GPC 177258),
a variant of mynd ymlaen.
58
Not in a separate entry, but under rhag.
189
PVs with ymlaen in the corpus represented three meanings of ON distinguished
by Rudzka-Ostyn. The first one refers to contact or getting closer to make contact (2003:
149). This basic sense of on usually occurs in non-idiomatic constructions, however it
can also refer to starting a device or switching something on. As could be expected, most
Welsh PVs expressing this meaning appeared to be recent: dod ymlaen ‘put on’, troi
ymlaen ‘turn on’ (GPC 20 c.), rhoi ymlaen ‘put on’(GPC “spoken”). Another example
found in GPC, but not in the corpus is logio ymlaen ‘log on’ (GPC 2000). The PV rhoi
ymlaen was also used once in the context of putting on weight, a probable calque from
English59.
The second meaning of YMLAEN/ON is temporal and implies progress in time
(Rudzka-Ostyn 2003: 152). This sense occurred only in the exceptionally frequent PV
edrych ymlaen ‘look forward (to)’, which curiously, is not to be found in GPC. However,
other PVs of this type are in the dictionary, all marked as spoken and not directly translatable into English: gweithio ymlaen, lit. ‘work on’, ‘work overtime’ (GPC “spoken”),
tynnu ymlaen lit. ‘pull on’, ‘be getting on (of time, age)’ (GPC 1893, “spoken”), wado
ymlaen, lit. ‘beat on’, ‘grow old, persevere’ (20th c., “spoken”). As for edrych ymlaen
‘look forward’, it should be noted that it usually occurs as a PPV with a second particle
equivalent to English ‘to’. The unmarked, most frequent second particle in the corpus was
at ‘to, towards’, competing with i ‘to’ (see 3.2.3.2, 3.2.3.3), which appears to be calqued
from English. The slight difference in the semantics of the second particles and the presence of a calqued variant might suggest that the PV is a loan rendition or a native idiom
rather than a direct loan translation.
The most frequent idiomatic meaning of ymlaen refers to continuation of an action
or situation. Among numerous examples from the corpus directly translatable into English are: annog ymlaen ‘urge on’, bod ymlaen ‘be on’, cadw ymlaen ‘keep on’, dod ymlaen
‘come on’ (GPC 1735), gyrru ymlaen lit. ‘drive on’, ‘carry on’ (GPC 1632), llusgo ymlaen
‘drag on’, mynd ymlaen ‘continue, proceed’, palu ymlaen ‘dig on’, symud ymlaen ‘move
on’ and loanblends: cario ymlaen ‘carry on’ (GPC 1793), pasio ymlaen ‘pass on’. Other
examples found in GPC are not directly translatable into English: dwyn ymlaen, lit. ‘bring
on’, ‘further promote’ (GPC 1658), hwylio ymlaen, lit. ‘sail on’, ‘promote, further, advance’ (GPC 1677).
According to Rudzka-Ostyn (2003: 150) put on weight also belongs to the “contact” category as ON
implies “getting closer” to an entity, which in this case is particular weight.
59
190
English on may also be a preposition occurring in prepositional verbs. The usual
Welsh equivalent, as far as literal, spatial meaning is concerned, is the preposition ar.
With regard to idiomatic PVs there were only 6 types with the preposition ar. Three of
them had direct English equivalents: taro ar ‘hit on’, ‘discover’(GPC 1603), troi ar ‘turn
on someone’ ‘attack’ (GPC c. 1400), tyfu ar ‘grow on someone’. Taro ar could also mean
‘bump into someone’, a possible loan rendition or native idiom, especially that it had
a variant taro ar draws, lit. ‘hit across’.
The remaining two idioms were not directly translatable into English: cadw ar, lit.
‘keep on’, ‘defend’ (GPC c.1330), colli ar, lit. ‘lose on’, ‘lose your senses’. In all these
items, AR, like ON, seems to involve two entities in contact (Rudzka-Ostyn 2003: 149154). GPC contains a large number of other idiomatic PPs with ar, first recorded in the
medieval period, e.g. dal ar, lit. ‘hold on’, ‘to affect; be affected with’ (GPC 13th c.),
syrthio ar, lit. ‘fall on’, ‘to affect gradually, overtake, possess’ (GPC 1346) dodi ar, lit.
‘put on’, ‘to name’ (GPC 13th c.), medru ar, lit. ‘can on’, ‘to come upon, come across,
light upon; find (out)’ (GPC 1346), gyrru ar, lit. ‘drive on’, ‘to accuse, charge, prosecute,
convict’ (GPC 13th c.). The dictionary provides plenty of similar examples, spanning
through many centuries60.
GPC also contains one example in which ar is use instead of ymlaen to calque the
English on: troi ar ‘to switch or turn on’ (light, radio, etc.); ‘arouse’ (sexually), ‘turn on’
(GPC 20th c., “spoken”). Such an example was not found in the corpus, as it belongs to
highly informal language.
All in all, idiomatic PPs with ar are well-established in the Welsh language and
few phrases are borrowed in that respect; many idiomatic English PPs with on such as act
on, chew on, draw on, wait on do not have their direct equivalents in Welsh.
Finally, the particle on was borrowed directly as on in 2 wholesale borrowings:
cym on and howld on, which are colloquial set phrases strongly marked in terms of style.
The particle should be therefore considered unproductive.
To summarise, it can be seen that the predications YMLAEN and AR have several
salient meanings in Welsh. Instances of transfer occur within these senses and new ones
have been added as well in order to convey modern concepts, such as switching on devices.
60
Cario ar, cyfyngu ar: cynyddu ar, darfod ar, dyrnu ar, golchi ar, gosod ar, lladd ar, rhedeg ar, etc.
191
3.3.2.4. I LAWR and DOWN
The Welsh adverbial particle i lawr corresponds to English down, implying negative verticality (Rudzka-Ostyn 2003: 104). I lawr occurred in 34 PV types in the corpus, with
1 probable N item, 7 PLs, 22 DTs and 5 LBs. It was the fourth most frequent particle
concerning the number of types and the fourth concerning the number of tokens (75).
The basic meaning of I LAWR, as in English DOWN, involves the trajector moving from a higher to a lower place (Rudzka-Ostyn 2003: 104). Semi-idiomatic pleonastic
constructions found in the corpus were disgyn i lawr ‘fall down’ and eistedd i lawr ‘sit
down’ (GPC 1567). Similar examples found in GPC are syrthio i lawr ‘fall down’
(GPC 12th c.) and gorwedd i lawr ‘lie down’ (GPC 1759). GPC contains several other
examples of transparent PVs with i lawr: dodi i lawr ‘sit down, relax for a while’
(GPC 1632), dod i lawr ‘come down’, ‘descend’ (GPC 1797), cloddio i lawr ‘dig down’
(GPC 1567), gosod i lawr ‘lay, put or set down’ (GPC 1632).
The second meaning according to Rudzka-Ostyn is related to time and geographically orientated motion (2003: 106), e.g. go down (in history), go down (south). No similar examples were found in the corpus or GPC, although the spatial meaning of going
south has been undoubtedly present in Welsh for decades, as in the following example:
(54) ac yn mynd i lawr i’r Sowth am bythefnos o holidays (Griffith 1975:4).
‘and going down to the South for two weeks of holidays.’
Yet another sense of I LAWR/DOWN is associated with a decrease in intensity,
quality, quantity, size, degree, value, activity, status, strength etc. (Rudzka-Ostyn 2003:
107). Examples found in the corpus were: symud i lawr ‘move down’, ‘lose status’, mynd
i lawr ‘go down’, ‘go to prison’, taro i lawr ‘strike/knock down’, ‘make weak’,
torri i lawr ‘break down’, ‘have a nervous breakdown’ or ‘destroy’ (GPC “spoken”), troi
i lawr ‘turn down (of sound)’ (GPC 1935 “spoken”), tynnu i lawr ‘pull down’, ‘corrupt’,
gollwng i lawr ‘let down’, ‘betray’, and a loanblend cwlio i lawr ‘cool down’. Other examples from GPC are torri i lawr ar ‘cut down on’ (GPC 20th c., “spoken”), hwylio i lawr
‘to fall (of wind)’ (GPC 1913 “spoken”), edrych i lawr ar ‘look down on’, ‘despise’(1632), pwyso i lawr ‘weigh down, oppress’ (GPC 1588), rhedeg i lawr ‘run down’,
192
‘criticise’(GPC 1740), sefyll i lawr ‘stand down (e.g. in an election)’ (GPC “spoken”),
troi i lawr ‘turn down’, ‘reject’, (GPC “spoken”).
The fourth meaning is connected with the entity reaching a goal, completion or
extreme limit down the scale (Rudzka-Ostyn 2003: 111). Examples in the corpus were:
pori i lawr ‘graze down’, torri i lawr ‘break down’, taro i lawr’ ‘knock down’ and
cloi i lawr ‘lock down’. Other examples from GPC are: berwi i lawr ‘boil down’ (GPC
1938), llosgi i lawr ‘burn down’ (GPC 1764), talu i lawr ‘pay down (cash)’ (GPC 1778),
setlo i lawr ‘settle down’ (1921, “spoken”).
Finally, Rudzka-Ostyn (2003: 112) distinguishes the sense of DOWN related to
downward movement associated with eating or writing. For I LAWR this was expressed
in PVs such as golchi i lawr ‘wash something down’ (e.g. with wine), ysgrifennu i lawr
‘write down’. These items were not found in GPC, but the dictionary contains another PV
connected with writing, taro i lawr lit. ‘strike down’, ‘jot down’ (1742).
The above data indicate that figurative meanings of I LAWR are not well-established and belong to more informal registers of the language, however examples of figurative senses in the particle are found already in the 16th-18th centuries. On the basis of
the available data it is problematic to decide whether contact with English triggered or
merely reinforced the process of meaning extension.
3.3.2.5. I MEWN, I MEWN I and IN, INTO
The English in is rendered in Welsh as mewn or yn if it is a preposition, and as i mewn
when it is an adverb. The particle i mewn and, more frequently, i mewn i corresponds to
English into. Like in and into, i mewn and i mewn i are closely related and sometimes
interchangeable, with IN/I MEWN related to entering or being inside a container
(Rudzka-Ostyn 2003: 48) and INTO/I MEWN I conveying the idea of motion into a container (Rudzka-Ostyn 2003: 69). No idiomatic PPs with yn or mewn were found in the
corpus, while there were 19 types of PVs with i mewn, 8 types with i mewn i and one PPV
with i mewn and ar as the second particle (cashio i mewn ar ‘cash in on’). In total, there
were 28 PV types with i mewn (i) in the corpus, with 4 N/LR items, one PL, 18 DTs and
5 LBs. Taken together, i mewn and i mewn i ranked fifth in frequency count concerning
the number of types and seventh concerning the number of tokens (49).
193
Parallel to English IN, the basic and most frequent meaning of I MEWN relates
to entities being inside or entering a container (Rudzka-Ostyn 2003: 48). Examples of
idiomatic PVs in the corpus conveying this sense were: camu i mewn ‘step in’, ffitio
i mewn ‘fit in’, ffonio i mewn ‘phone in’, galw i mewn ‘call in’, llosgi i mewn ‘burn in’,
pitsho i mewn ‘pitch in’ (GPC “spoken”), symud i mewn ‘move in’, taro i mewn ‘pop in’
(GPC 1886, “spoken”), teipio i mewn ‘type in’, torri i mewn ‘break in’ (GPC 1567), troi
i mewn ‘drive in (of cattle)’, tynnu i mewn ‘pull in (of vehicle)’. Other examples from
GPC include: cau i mewn (GPC 1588), cymryd i mewn ‘take in’ (GPC 1632),
edrych i mewn ‘look into’, ‘inspect’ (GPC 1632), gollwng i mewn ‘let in’, ’inject’ (GPC
1740), gorwedd i mewn ‘lie in’, ‘lie in childbed’ (GPC 1771). More recent examples from
the dictionary are: popio i mewn ‘pop in’ (GPC 1936), clocio i mewn ‘clock in’, (GPC
1995), gwrando i mewn ‘listen in (to the radio)’ (GPC 20th c.)’and logio i mewn ‘log in’
(GPC 1999).
The remaining PVs with I MEWN (I) in the corpus referred to situations, circumstances viewed as containers (Rudzka-Ostyn 2003: 55): cashio i mewn ar ‘cash in on’,
llusgo i mewn ‘drag into’, dod i mewn i ‘come into’, rhoi i mewn ‘put in’, ‘invest’ or
‘submit’, tynnu i mewn i ‘pull (someone) into something’. Most of these items appeared
in dialogues in the novel Afallon or in the satirical magazine LOL and therefore can be
considered non-standard. Another example, ymchwilio i mewn ‘research into’, was included as a pleonastic construction. Two other examples were found in GPC: cymryd
i mewn ‘take in, ‘involve’ (1775), rhedeg i mewn ‘run in’, ‘shrink’ (1730).
The sense of IN related to psychological or physical states viewed as containers,
distinguished by Rudzka-Ostyn (2003: 57), expressed in PVs such as cram in, take something in, was not found in the Welsh corpus.
The metaphorical sense of INTO, implying change, motion from one state into
another (Rudzka-Ostyn 2003: 70) was not found in the corpus safe from one PPV, in
which yn was the second particle: troi mas yn ‘turn into’. The example came from the
novel Afallon and was used in conversation by a character who “really” spoke English,
therefore intentional calquing might be possible. Generally speaking I MEWN I was not
used to refer to states such as habits or emotions – Welsh equivalents of constructions
such as break into (tears), fall into (disgrace) were not found. The phrase bump into has
its Welsh equivalent in a phrase with a different particle, taro ar.
194
All in all, particles i mewn and i mewn i were less productive in Welsh than in
English. The spatial meanings of the particle are relatively rarely extended to the abstract
and appear to be non-standard. However, the earliest examples of metaphorical extension
found in GPC date to the 16th century.
3.3.2.6. YN ÔL and BACK
The Welsh particle yn ôl corresponds to English back. YN ÔL/BACK mostly occurs in
transparent constructions, conveying the sense of “returning to or staying at an earlier
location”. Such items were excluded from the corpus, apart from 3 pleonastic constructions: bacio yn ôl, cilio yn ôl and dychwelyd yn ôl. The sample contained 13 other types
of PVs (39 tokens) where the particle conveys the metaphorical sense of returning to an
earlier state, time or situation (Rudzka-Ostyn 2003: 174). Among these were:
tynnu yn ôl (o) ‘pull back’, ‘not engage’, eistedd yn ôl ‘sit back’, torri yn ôl ‘cut back’,
‘make savings’, meddwl yn ôl (dros) ‘think back (over)’, edrych yn ôl ‘look back’, dal yn
ôl ‘hold back’ ‘stop progress’ or ‘hide’ (e.g. feelings), cael yn ôl ‘get (someone) back’,
‘revenge’.
Only three PVs with yn ôl are recorded in GPC, presumably due to the transparency of such constructions. These are: dwyn yn ôl ‘bring back’ (1620), galw yn ôl ‘recall,
call back’ (1632), rhoi yn ôl ‘return, replace’ (1632)61.
3.3.2.7. I FFWRDD, YMAITH, OFF, ODDI AR and AWAY, OFF
The Welsh particles i ffwrdd and ymaith, used interchangeably, and bant, a variant of
i ffwrdd used mostly in southern dialects of Welsh, are equivalent to two English particles: away and off. There were 13 types of PVs containing i ffwrdd and its variants, 19
tokens in total, including 10 tokens with bant, 7 with i ffwrdd and 2 with ymaith. All of
these were directly translatable into English. There was also one pleonastic construction
All three recorded in John Davies’s Dictionarium Duplex, 1632; dwyn yn ôl recorded earlier in a translation of the Old Testament.
61
195
ysgubo ymaith ‘sweep away’. In 10 PV types the particle was translated as away and in
4 as off.
What is more, in informal speech the particle off is occasionally borrowed directly
from English. Three PV types with this particle were found in the corpus: LBs copio off
‘cop off’ and hitio off ‘hit off’, and one wholesale borrowing ffyc off ‘fuck off’. All of
these were stylistically marked as teenagers’ slang.
Rudzka-Ostyn distinguishes three main meanings of AWAY. The first and basic
one refers to the entity leaving a place and not being in it (Rudzka-Ostyn 2003: 139).
A relatively high number of such non-idiomatic or semi-idiomatic constructions with
i ffwrdd and ymaith can be found in GPC, with the earliest dating back to the medieval
period: gyrru ymaith ‘send away’, ‘drive out’ (GPC c.1300), bwrw ymaith ‘cast away’
(GPC 14th c.), dyrru ymaith (GPC 14th c.) ‘drive or send away’, torri ymaith ‘cut away’
(GPC 15-16 c. ), troi i ffwrdd ‘turn away’ (GPC 1567), dyrru i ffwrdd ‘drive or send
away’(GPC 1568), dwyn ymaith ‘take away’ (GPC 1588), gollwng ymaith ‘dismiss, send
away’ (GPC 1588), gyrru i ffwrdd ‘send away’ (GPC 1604-7), rhoddi ymaith ‘put away’,
‘cast off, store’ (GPC 1604-7), galw ymaith ‘call away’ (GPC 1632), pacio i ffwrdd ‘pack
away’, ‘send packing’ (GPC 17th c.), pechu ymaith ‘drive away by sin’ (GPC 1730), taflu
ymaith ‘throw away’ (GPC 1595). Examples found in the corpus, but not in GPC were:
symud i ffwrdd ‘move away’, ‘withdraw’, cymryd i ffwrdd ‘take away’, ‘deprive’, cadw
i ffwrdd ‘keep away’, ‘avoid’.
The second meaning of English AWAY, associated with gradual and continuously
growing distance – in constructions such as drink away, fade away, rot away, work away,
wash away (Rudzka-Ostyn 2003: 141) – could not be attributed to any of the Welsh PVs.
The only relevant example found in GPC was the colloquial wado bant, lit. ‘thrash away’,
‘work hard, work like fury’ (GPC 1929, “spoken”).
The third meaning of AWAY implies complete disappearance of an entity
(Rudzka-Ostyn 2003: 143). PVs conveying this sense found in the corpus were: ysgubo
ymaith ‘sweep away’, chwysu ymaith ‘sweat away’, rhoi bant ‘give away’. Items found
in GPC convening the latter sense of the particle are: chwarae ymaith, lit. ‘play away’
‘squander, fritter away’ (1713), cwympo ymaith ‘fall away’, ‘fail’ (GPC 1567), golchi
ymaith ‘wash away’ (1588), gwisgo ymaith ‘wear away’ (GPC 1595), ynfydu ymaith,
lit. ‘fool away’, ‘squander’ (GPC 1727). It appears that in the past the particle ymaith used
to be the productive particle to convey this sense, while currently i ffwrdd is preferred for
196
calquing English idioms. This is confirmed by the presence of one more recent entry conveying the sense of disappearance: gwneud i ffwrdd â ‘do away with’ (GPC 20th c.).
The English particle off does not seem to have an obvious Welsh equivalent, which
explains the above mentioned direct borrowing in PVs such as hitio off. The basic spatial
meaning of OFF, “loss of contact and separation” (Rudzka-Ostyn 2003: 121) can be rendered by DRAW or I FFWRDD/YMAITH. GPC contains an example of recent PV conveying this sense logio i ffwrdd ‘log off’ (GPC 1999).
Moving on to figurative senses, the English OFF expressing separation as interruption of flow or supply (Rudzka-Ostyn 2003:125) was conveyed in Welsh with
I FFWRDD (BANT): troi i ffwrdd ‘turn off’, ‘switch off’ (GPC 20th c., “spoken”), mynd
i ffwrdd ‘go off’, ‘stop working’. Another sense – separation due to motion away from its
former state, condition or point of reference (Rudzka-Ostyn 2003: 126) was identified in
only one PV in the corpus: mynd i ffwrdd ‘go off’, ‘explode’. Calquing of this sense appears to be very rare, with no direct Welsh equivalents such as spark off, ease off, bit off,
write off, buy off, pay off, show off.
Off as a preposition can be equivalent to Welsh oddi ar (for basic spatial meanings
cf. e.g. GA entry for off). Only one such example of an idiomatic PV was found in the
corpus: dod oddi ar ‘come off’ (e.g. drugs)’ in a dialogue from the novel Afallon.
The above evidence suggests that the particles ymaith and i ffwrdd have been used
in Welsh since medieval times in semi-idiomatic context. Few figurative meanings corresponding to English AWAY emerged over the centuries probably due to partial extension through contact with English. Similarly, there are relatively few cases of calquing
English off. The only well-established calqued sense of I FFWRDD is that of “switching
off”, parallel to the extension of meaning of YMLAEN for ON. Cases of direct borrowing
of off were rare and stylistically marked as slang. Rottet (2005: 59) suggests that the borrowing of off could be seen as freeing up i ffwrdd and bant to function as equivalents of
away. However, in my corpus i ffwrdd and bant prevailed in translating English off.
3.3.2.8. HEIBIO and BY, ASIDE
Heibio is an adverbial particle meaning “along some point or specific place (in time or
space) and beyond it, out and forwards after passing closely” (GPC). The particle heibio
197
occurred in 12 PV types in the corpus (51 tokens). In the vast majority of the cases it
corresponded to English by. The PV galw heibio ‘call by’ (GPC; “common spoken”) was
the only idiomatic PV in the corpus expressing the core sense of BY/HEIBIO “location
or motion near or at the side of an entity” (Rudzka-Ostyn 2003: 200). With 10 tokens in
the sample, galw heibio competed with galw draw, which occurred 3 times. A similar PV
is recorded in GPC: taro heibio ‘drop in’ (GPC 20th c.).
More often, the meaning of HEIBIO in Welsh PVs corresponded to the second
meaning of BY – closeness or location in time and measurement units (Rudzka-Ostyn
2003: 201). The PVs expressing the temporal sense of passing were: mynd heibio ‘go
by’62, and related hedfan heibio ‘fly by’, diferu heibio ‘drop by’, pasio heibio ‘pass by’
(GPC 1615), rasio heibio ‘race by’. Similar PVs recorded in GPC are: bwrw’r amser
heibio ‘while away the time’ (1562), gollwng heibio ‘let pass, let slip, pass by’ (16th c.).
In one case heibio combined with preposition i to express the sense of “beyond”
– mynd heibio i’r targed ariannol ‘go beyond the financial target’. The particle also occurred in the PV rhoi heibio which had three senses: ‘put aside, away’ (GPC 16th c.),
‘give up’ and ‘put on, entrust’. Similar meaning can be found in the following PVs recorded in GPC: gadael heibio ‘leave off, lay aside’ (15th c.), bwrw heibio ‘cast away, lay
aside’ (1655), adaw heibio ‘leave off, cease, omit’ (1659), troi heibio ‘give up’ (1604-7),
‘leave off’ and ‘save up money’ (1768).
It can be seen that the particle heibio has been forming PVs expressing notion of
passing in both special and temporal aspect, parallel to English by. There is little evidence
for direct calquing in this respect. Moreover, the meaning was extended in Welsh into
“putting aside” or “getting rid of something”, which is absent from English BY.
3.3.2.9. O GWMPAS, ROWND and (A)ROUND
The particle o gwmpas corresponding to English around (or round) occurred in 11 PV
types (14 tokens). Idiomatic PVs with o gwmpas appear to be rare. In most instances
O GWMPAS expresses primarily the basic spatial meaning of AROUND “location or
motion (in different direction) often viewed from a central point” (Rudzka-Ostyn
Although mynd heibio is not recorded in GPC, the dictionary contains a phrase wrth fynd heibio ‘in
passing (of comment)’ (20th c., “spoken”).
62
198
2003: 184). Most PVs in the corpus conveyed the largely figurative sense of something
being or happening “nearby”, “suggesting vague but close whereabouts” (Rudzka-Ostyn
2003: 185): bod o gwmpas ‘be around’, cadw o gwmpas ‘keep (someone) around’. PVs
suggesting “motion in no particular direction” or being “engaged in a continued or repetitious activity in different directions which is furthermore directed to no one or nothing
in particular” (Rudzka-Ostyn 2003: 188-189) were: hongian o gwmpas ‘hang around’,
llusgo o gwmpas ‘drag around’, mynd o gwmpas ‘go or set about something’ (GPC 1793),
prowlian o gwmpas ‘prowl around’ and possibly chwarae o gwmpas ‘play around’. There
was one example of PPV containing o gwmpas: chwarae o gwmpas â ‘play around with
someone’, ‘have a sexual relationship’, used in teenage slang. Except for mynd o gwmpas,
GPC contains only one PV with o gwmpas: troi o gwmpas ‘revolve around, be mainly
concerned with’ (GPC 20th c.),
Figurative PVs with o gwmpas generally appear to be calqued from English,
which is confirmed by the presence of directly transferred particle rownd. In the corpus
it occurred only once in a fictional dialogue: dod rownd ‘get around (someone)’, ‘persuade’. There is also only one PV with rownd in GPC: troi rownd ‘turn (a)round, revolve
around, be mainly concerned with’ (GPC “spoken”).
It is notable that all the items with o gwmpas in the corpus appeared in the conversation mode, mostly in novel dialogues and occasionally in press interviews. This indicates that PVs with o gwmpas are a fairly recent case of calquing and belong to colloquial registers of the language.
3.3.2.10. DRAW, DROS, DROSODD, and OVER, FOR, AWAY
Draw is an adverb, which according to various Welsh dictionaries can be translated into
English as: ‘yonder, there, beyond’ (GPC), ‘there, yonder’ (GWE), ‘over (there)’ (TSD),
‘yonder, off, beyond’ (GB). Draw has not been previously mentioned in the literature as
a particle in Welsh PVs. However in the corpus it occurred in 32 tokens with 7 types
(including 3 syntactic variants of cadw draw (o/oddi wrth).
Only one PV with draw was directly translatable into English over: dod draw
‘come over’, ‘visit’. Although the item was classified as DT, it could as well be attributed
to the N category due to the fact that three other PVs in the corpus galw draw, lit. ‘call
199
over there’, mynd draw ‘go over there’63 and picio draw, lit. ‘pick over there’ all convey
the meaning of visiting someone for a short time, calling by, popping over. Picio draw is
a particularly interesting case, in which an original Welsh PV is created with a verb borrowed from English. The remaining PV with draw was cadw draw (o/oddi wrth) ‘keep
away (from)’.
None of these items can be found in GPC. However, the dictionary contains the
following two examples: closio draw ‘move away’ (GPC “spoken Northern’) and sefyll
draw, ‘stand back, aside, or off, also fig.’ (GPC 1923), indicating the spoken character of
the PVs containing this particle. A brief search through historical corpora of Welsh available (Willis 2004; Luft et al. eds. 2013) has demonstrated, however, that the particle was
used in figurative PVs centuries ago. An interlude based on the story of King Lear,
Y Brenin Llur, dated to the first half of the 18th century, contains an example of sefyll
draw, two centuries earlier than the date provided by GPC for the first recorded use of
the item:
sa [~ saf] di draw fy merch Cordila
tra bwi 'n galw dy chwaer rodia64
‘stand aside my daughter Cordila
while I call your sister walking’
Although the meaning of sefyll draw here is transparent, the same text contains
another PV, dal draw lit. ‘hold there’, ‘keep away’, which is clearly figurative:
wel wel pa beth dybygwn
troi doe yn ol nis gallwn
ni godwn armi deg os daw
i 'w dal nhw draw os gallwn65
[well well, how could we suppose
we cannot turn yesterday back
mynd draw was included in the corpus due to the fact it preceded the adverb yno ‘there’ and therefore
was considered pleonastic.
64
Y Brenin Llur, Cwrtmawr ms. 212, p. 11, (Willis 2004).
65
Y Brenin Llur, Cwrtmawr ms. 212, p. 59 (Willis 2004).
63
200
we will not raise a fair army if he comes
to keep them away if we can]
It should be noted, that the text originates from North Wales. Most examples in the corpus
also occurred in Northern dialects.
In conclusion, the above evidence supports a claim that DRAW has evolved from
a spatial adverb indicating that an entity is “over there” to a native Welsh particle without
an English equivalent, used in PVs to indicate that an entity is away from sight, as in sefyll
draw, dal draw. The other figurative meaning of DRAW is connected with visiting, going
“there”; this meaning appears to be more recent. Constructions with draw may have
emerged from Northern dialects of Welsh.
Dros (or tros) is a simple inflected preposition with multiple meanings. Drosodd
(trosodd) is a form of dros which functions as an adverb. Both particles are used to form
PVs in Welsh and in most contexts are equivalent to English over, indicating “movement,
position, direction, etc., above or across, on (covering), across and down (over a border
or from direct position), also figuratively)” (GPC, “tros” sense 1.) In the corpus there
were 6 types of PVs with dros (15 tokens) and 9 types with drosodd (11 tokens).
According to Rudzka-Ostyn in its basic literal sense, OVER implies “being or
moving higher than and close to something or from one side to the other” (Rudzka-Ostyn
2003: 160). This meaning can be extended figuratively e.g. to communication, passing
knowledge or skipping, avoiding something. In the Welsh corpus, the meaning was contained in the PVs: rhoi drosodd ‘put over’, ‘make understand’ and dod dros rhywun ‘get
over (someone/something)’, ‘forget’. A similar PV found in GPC are edrych dros ‘overlook, fail to observe, pass over’ (1677). Gwylio dros ‘watch over, guard’ (1567) would
also belong to this network.
Another meaning of OVER implies that “an entity has to cover some distance –
spatial or mental – to get closer to another object or goal” (Rudzka-Ostyn 2003: 162).
This is the case with visits, lending money temporarily to somebody, getting to know
a person, replacing a person in a certain job, etc. Examples with DROS found in the sample are: dod drosodd ‘come over’, ‘visit’, cymryd dros/drosodd (o) ‘take over’, ‘replace’,
cymryd drosodd ‘take over’ ‘take control’, troi drosodd ‘turn over’, ‘start’ (of an engine).
As mentioned above, the meaning of visiting someone, implying “going there” was also
expressed in 3 PVs with DRAW: mynd draw, dod draw and picio draw.
201
The third meaning of OVER refers to “motion viewed as covering completely or
even in excess” (Rudzka-Ostyn 2003: 162). Three Welsh constructions in the corpus covered this meaning: bod drosodd ‘be over’, (GPC 1758), dod dros ‘come over someone
(an emotion)’, golchi dros ‘wash over someone (feeling)’. One more example found in
GPC is rhedeg drosodd lit. ‘run over’, ‘overflow’ (1604).
Next, OVER may imply “examining thoroughly from all sides” (Rudzka-Ostyn
2003: 167). This sense also occurred in the corpus in two PVs with DROS: meddwl dros
‘think (something) over’ and mynd dros ‘go over (something), ‘check, practice’. Although
these items are not in GPC, the dictionary contains three other examples: rhedeg drosodd
‘run over, study’ (1604-7), darllen trosodd ‘read through’ (GPC 1767) and edrych dros
‘look over’, ‘survey’ (1778) and puts this sense under sense b.: “all over, throughout,
through, also fig.”
The final meaning of OVER distinguished by Rudzka-Ostyn is connected with
“reflexive motion or being completely bent” (Rudzka-Ostyn 2003: 168). None such examples were found in the corpus, while GPC contains only an example of a transparent
PV of this type with DROSODD: troi drosodd ‘turn over, overturn, upset’ (1632).
There was one PV with dros when the particle was not equivalent to English over,
but for, meaning ‘for the sake of someone’: sefyll dros ‘stand up for, defend’
(GPC 16th c., but meaning ‘stand in for, take the place of, deputize’ – 14th c.). This is
connected with sense 3a of tros distinguished in GPC: “for the sake of, on behalf of, in
support of”.
As shown above, many, but not all concepts inherent to the English particle over
are present in Welsh dros and drosodd. The figurative meanings appear to be fairly wellestablished in the language and should not necessarily be considered calques, despite being directly translatable into English in most cases. In particular, the case of draw demonstrate the ability of Welsh to form its own PVs.
3.3.2.11. DRWODD, TRWY and THROUGH
202
The English through occurring in PVs corresponds to two Welsh particles: the preposition
trwy and adverb drwodd (trwodd), with trwy forming transitive APVs and drwodd intransitive ones. There were 6 types of PVs with drwodd, with only 2 verbs, mynd and dod
(7 tokens) and 7 types with the preposition trwy (11 tokens).
According to GPC, drwodd (or trwodd) is an impersonal form of the preposition
trwy functioning as an adverb. Out of the 7 tokens with drwodd, 6 appeared in the novel
Adenydd glöyn byw, mostly in dialogues, and 1 in a press interview translated from English. This might indicate that PVs with drwodd belong to dialectal spoken register, however the sample is too small to make general conclusions.
The PVs which were identified all expressed the main figurative meaning of
THROUGH, referring to “activities viewed as complete(d) motions” (Rudzka-Ostyn
2003: 195-196): dod trwy/drwodd 1) ‘get through’, ‘pass’ 2) ‘go through’, ‘endure’, or 3)
‘get through’, ‘analyse’; edrych trwy ‘look through’, ‘search’; gweld trwy ‘see through’;
mynd trwy/drwodd 1) ‘get through’, ‘finish’, 2) ‘go through’, ‘experience’ and 3) ‘go
through’, ‘search’.
GPC contains no examples of PVs with trwy or drwodd, except for gweld trwy
‘see through’ (GPC “common spoken”).
3.3.2.12. AR DRAWS and ACROSS
The particle ar draws corresponds to English across. 6 types of PVs with ar draws were
found in the corpus, with 68 tokens in total. The high number of tokens stemmed from
high frequency of two PVs: dod ar draws ‘come across’ and torri ar draws ‘interrupt’
(30 and 36 items, respectively).
In 4 PV types, AR DRAWS expressed a sense which does not occur in English
ACROSS and is related to interruption. Thus, the most frequent of these PVs, torri
ar draws lit.’ break across’, ‘interrupt’ (GPC 1788), conveys the sense of “breaking into”
someone’s speech66. The other two PVs with this meaning in the sample were siarad
ar draws ‘interrupt’ (GPC “spoken”) and llefaru ar draws. In one case, dod ar draws also
expressed the concept of interrupting, something “coming in our way”. GPC gives a 14thGPC mentions also another, older meaning of torri ar draws (1567), ‘split open’, ‘burst’, probably conveying the idea of metaphorically “cutting across”.
66
203
century manuscript Ystorya de Carolo Magno as the source of the first recorded use of
dod ar draws. The dictionary does not provide a citation but the example has been found
in a historical corpus of Welsh. Although the item is rather transparent, it already contains
the idea of interruption, as the knight metaphorically “breaks” into the flow of enemies,
literally: “ac y deuth Rolant ar draws yn y erbyn”67 [and Rolant came across against
them].
One more expression of this type found in the dictionary is taro ar draws meddwl,
lit. ‘hit across thoughts’, ‘occur to one, strike one’ (GPC 1811).
The other 3 types of PVs corresponded to the meaning of English across formulated by Rudzka-Ostyn as “figurative motion crossing to a human receiver” (2003:193).
The meaning was found in two Welsh PVs: dod ar draws ‘come across’, ‘find’, rhoi ar
draws ‘put (something) across’, ‘communicate something’. It is not impossible that this
network of senses of ar draws is calqued from English.
All in all, AR DRAWS appears to convey two metaphorical senses, one of which
is native to the Welsh language, while the other may be attributed to calquing.
3.3.2.13. AR ÔL and AFTER
Ar ôl is a complex preposition corresponding to English after. The sample contained 9 PV
types with ar ôl and 48 tokens. The semantic structure of AFTER has not been included
in Rudzka-Ostyn (2003), however it has been described by Boers (1996: 184–187), whose
model will be used here. According to this author, the first figurative meaning of AFTER
in PVs is related to following or pursuing an entity. This fairly transparent metaphor occurred also in PVs with AR ÔL: bod ar ôl ‘be after’, dod ar ôl ‘come after’, mynd ar ôl
‘go after’, rhedeg ar ôl ‘run after’. These items are not in GPC, but the sense can be found
in cerdded ar ôl ‘walk after’ (GPC 1588).
Within the next figurative meaning of AR ÔL, an entity moves along a path to the
purpose which is the destination. Within this metaphor a caretaker is viewed as a guide
along the path, which gives rise to expressions such as look after (Boers 1996:186). GPC
mentions this sense under sense 4 of ar ôl “after, about (used with words expressing care,
67
Ystoria Carolo Magno: Rhamant Otfel, NLW MS. Peniarth 5 (The White Book of Rhydderch, part 1), p.
87 (Luft et al. eds. 2013)
204
concern, etc.)”. The date of the earliest recorded use given is 1672. This sense was contained in the most frequent PV type with ar ôl (35 tokens), edrych ar ôl ‘look after’ (GPC
1853) and its variant gwatsio ar ôl ‘watch after’, which occurred only once and is
a loanblend. One more example found in the corpus was mynd ar ôl lit. ‘go after something’, ‘talk about’, ‘be concerned’, which is not directly translatable into English.
Finally, AFTER may involve a metaphor of the guide being an example (Boers
1996: 187). The PVs with AR ÔL expressing this sense were enwi ar ôl ‘name after’ and
tynnu ar ôl, lit. ‘pull after’, ‘take after’ (in appearance or temperament) (GPC “spoken”).
As can be seen, there is close correspondence between the semantic structures
AR ÔL and AFTER but the parallels may be at least partly coincidental due to the transparency of metaphors.
3.3.2.14. Other prepositional particles
The remaining particles in the corpus were single instances of other prepositions in idiomatic PPs: â (gyda, efo), am, at, yn erbyn, heb and oddi wrth.
Am is an inflected preposition with multiple meanings, one of which is equivalent
to English for (cf. sense 1d and 1e of “am”, GPC). There were 5 types and 9 tokens of
PPs with am in the corpus. The only one which was not directly translatable into English
was tynnu am, lit. ‘pull for’, ‘get on for, turn (a particular age)’. In the remaining examples, am corresponded to English for: edrych am ‘look for something’, mynd am ‘go for
something’, cwympo/syrthio am ‘fall for someone (in love)’. All of these examples occurred in fiction and the latter 3 examples belonged to teenage language in the novel
Adenydd glöyn byw, which indicates that they are recent calques from English.
The preposition â and its variants gyda and efo, corresponding to English with,
occurred in 4 PV types, with 7 tokens. Only one of the types was directly translatable into
English: mynd â ‘go with someone, ‘have a relationship’. The other examples appeared
to be native Welsh constructions: galw gyda, lit. ‘call with’, ‘call someone’, gwneud â,
lit. ‘do with’, ‘get on with someone, ‘have a good relationship’ and crafu efo, lit. ‘scratch
with’, ‘flatter someone and try to get something from them’. The latter PP, not registered
in dictionaries, seems to be a dialectal expression from the Isle of Anglesey. Two common
205
native PPs: mynd â, lit. ‘go with’, ‘take’ and dod â, lit. ‘come with’, ‘bring’ were also
considered non-compositional and not included in the corpus.
The preposition at occurred in only two idiomatic PV (7 tokens): troi at (6 tokens),
with two senses: ‘turn to (something)’, ‘change’ or ‘turn to someone for help’ (GPC 13th
c.) and tynnu at lit. ‘pull towards’, ‘turn (a particular age)’. It should be noted that at is
also used in the constructions with ati, which were excluded from the corpus, such as:
mynd ati, dal ati, gwneud ati (see 2.4.3.3).
The preposition yn erbyn ‘against’ occurred in only two idiomatic PVs (2 types
and tokens): dal yn erbyn ‘hold something against someone’, mynd yn erbyn ‘go against
someone (of facts)’. Finally, two prepositions occurred in the corpus only once: heb ‘without’ in gwneud heb ‘do without’ and oddi wrth ‘from’ in tynnu rhywun oddi wrth, ‘pull
someone away from’, ‘distract’. These items appear to be relatively transparent metaphors and therefore were not necessarily calqued.
3.4. Corpus study – summary of the findings and discussion
The analysed data have shown that idiomatic phrasal verbs are a widespread phenomenon
in standard Welsh, not confined to the spoken language. Idiomatic PVs are used extensively in semi-formal varieties of Welsh, including quality press and narrative parts of
award-winning novels. The proportion of items used in narrative and conversation modes
was almost even and only a small number of PVs was stylistically marked.
The corpus contained a total of 1134 PV tokens with 271 types conveying 398
different senses. Due to lack of digitalised data, a rough frequency count was performed
only for the fiction subcorpus, where the amount of PVs per 1000 words varied from
0.9 to 2, depending on the style of the novel, in particular the amount of colloquial language used. This can be compared to some extent with findings of Bieber et al.
(1999: 408), whose subcorpus of English fiction contained 2 PVs per 1000 words. It
should be remembered, however, that the authors’ analysis did not include prepositional
verbs. If one excluded PPs from the present corpus the number of PVs per 1000 words
would be between 0.7-1.868 that is the mean of 1.1. Although the size of the sample is
68
Afallon – 1.80, Adenydd glöyn byw – 1.39, Craciau – 0.92, Y Llyfrgell – 0.74 Tair rheol anhrefn – 0.67.
206
insufficient to make definite or general conclusions, it could be hypothesised that at least
in fiction PVs in Welsh are less frequently used in comparison with English. Further studies on larger corpora would be required to confirm these observations.
The productivity of PVs in Welsh has been found to be high. New constructions
are constantly formed, and the existing ones contain not only basic verbs of movement
but also borrowings from English, both recent and well-established. The most productive
verbs in the corpus were mynd and dod. They were also the most frequent ones, after
edrych, which had an exceptionally high number of tokens due to popularity of a single
expression edrych ymlaen. The analogous English verbs go and come have also been
found to be the most frequent ones in English corpus studies (Gardner and Davies
2007: 349). The third and fourth position in the Welsh corpus were occupied by troi ‘turn’
and tynnu ‘pull, draw, drag, take’, while on Gardner and Davies’s list, it was take and get,
which do not have obvious Welsh equivalents.
A total of 28 particles – 15 adverbial and 13 prepositional ones – forming PVs
was found in the corpus, excluding 25 dialectal and colloquial variants. The list contains
all the adverbial particles mentioned by Rottet (2005) and three additional ones: drwodd,
draw and on. Drwodd corresponds to adverbial through and was used mostly in conversation. It combined with only two verbs: mynd and dod. Draw was found to be a wellestablished and productive ‘native’ Welsh particle (see below). On occurred only in two
set phrases and can hardly be seen as productive. The particle rownd in the present corpus
occurred only once, as a preposition, and was classified as such, while Rottet provides
examples of the particle being used as an adverb (Rottet 2005:54).
The most frequent and productive particles in the corpus were allan, i fyny and
ymlaen. According to Bieber et al. (1999:413) in English the most productive adverbial
particles forming PVs are: up, out and on. The parallel sets of most productive/frequent
particles as well as verbs in Welsh and English may point out to the role of language
contact, but they also could be attributed to high salience of these verbs and particles in
the two languages.
The most frequent syntactic type of PVs, constituting around two thirds of the
items in the sample, were APVs, consisting of a verb and an adverbial particle. Intransitive constructions were much more frequent than transitive ones, with 511 and 165 tokens, respectively. PPVs, constructions with two particles, were mostly intransitive and
half of them were variants of APVs.
207
As regards the word order in transitive APVs, the usual pattern observed in the
corpus was Verb-Object-Particle. This agrees with the results of Hirata’s study
(2012:166), in which this order was generally preferred by speakers of Welsh. The unusual Verb-Particle-Object order appears to be archaic and therefore stylistically marked
in contemporary texts. On the other hand, it seems that the Verb-Particle-Object order has
re-emerged recently in less formal registers of Welsh, probably due to contact with English. However, examples of this were rare in the sample.
Although the majority of items in the corpus were directly translatable into English, the study has demonstrated that the phenomenon of PVs in Welsh cannot be attributed solely to language contact, confirming Rottet’s (2005:50-52) observations that
Welsh is capable of forming its own idiomatic PVs. A number of PVs in the sample were
not directly translatable, with at least 20 APV or PPV types and 15 PP types. A crosscheck with the historical dictionary of Welsh, GPC, has shown that some of these items
are centuries old (taro ar, gosod allan); others were dialectal expressions (crafu â, codi
allan) and semi-idiomatic, ad hoc metaphors (brasgamu ymlaen, diferu heibio). Moreover, the analysis of the semantics of Welsh particles revealed cognitive senses unique to
Welsh, for example AR DRAWS conveying the sense of interruption which is not found
in English ACROSS, and HEIBIO used in the sense of “putting aside”, “getting rid of
something”, absent from English BY.
Another important example is the adverbial particle draw ‘(over) there’, which
has been overlooked by researchers thus far. It is claimed here that draw should be considered a “native” Welsh particle as most items containing it are not directly translatable
into English and in some cases compete with other directly calqued variants. Available
diachronic corpus data show semantic extension of this adverb into a productive particle
with two figurative senses of DRAW: “moving away” and “visiting” someone, which
emerged over the centuries. The particle is productive in that it combines with the most
frequent verbs, such as mynd and dod, but may also enter new constructions with borrowed verbs such as picio.
The naturalness of PVs in Welsh is also proven by the presence of ad hoc metaphorical constructions, containing particles such as ymlaen, allan, heibio and
i ffwrdd/ymaith, whose figurative senses have been present in the Welsh language for
centuries. Therefore, the correspondence between some of the most salient particle senses
208
between Welsh and English, e.g. ALLAN/OUT for exclusion, YMLAEN/ON for continuation or progress, may not necessarily result from language contact. It might as well be
attributed to cognitive mechanisms of metaphorical extension of meaning from spatial to
abstract common to the cultural zone. The answer to this question would require further
studies.
Still, despite the fact that PVs are deeply rooted in the Welsh language, there can
be no doubt that mechanisms of transfer play a crucial role in the widespread use of these
constructions. The degree of direct word-to-word correspondence between Welsh and
English PVs cannot be coincidental. Thus, the present study confirms Jones’s (1979:113)
and Rottet’s (2005) assumptions that contact with English reinforces the elements already
present in Welsh due to extensive calquing and other mechanisms of transfer.
The mechanisms enumerated by Rottet (2005:52-60) have all been found in the
corpus. The first of them was loan rendition, visible e.g. in the pattern of calquing English
get as dod or rendering prefixed verbs with a verb-particle construction as in the case of
upload translated as llwytho i fyny. Pleonastic calquing occurred as well, both in spatial
and aspectual constructions.
The most frequent phenomenon was word-for-word calquing, the mechanisms of
which is fairly complex and by no means random. The semantic analysis has demonstrated that the naturalness of certain particle senses in Welsh greatly facilitates borrowing. On the other hand, restrictions on calquing have been noticed as well, regarding both
verbs and particles. English verbs which do not have simple Welsh equivalents are transferred to a limited degree. The most obvious example is the verb have which does not
have a Welsh equivalent as the concept of possession is expressed by prepositional constructions. Consequently, none of the English PVs with have occurs in Welsh. Some English particles, such as about and along, are also not calqued.
The analysis of transfer of meanings for each of the particles found in the corpus
has identified a group of particle senses which can be assumed to be calqued from English
due to their absence in the historical dictionary, low frequencies and/or markedness. The
most important of these calqued senses are:
ALLAN indicating maximum boundaries, completeness of an action;
209
I FYNY in most of its figurative senses: reaching a goal, an end or a limit; becoming
more visible, accessible or known; completeness, reaching the highest limit of something or covering an area completely;
I MEWN (I) in its figurative senses, where situations and circumstances are viewed
as containers;
I LAWR in most of its figurative senses: decrease in intensity, quality, quantity, etc.;
reaching a goal, completion, extreme limit down the scale;
O GWMPAS in figurative senses of: nearby, suggesting vague but close whereabouts;
motion in no particular direction; being engaged in a continued or repetitious activity
in different directions which is furthermore directed to no one or nothing in particular;
AM in prepositional verbs rendering English FOR;
AR for translating the adverbial ON;
OFF used in loanblends and wholesale borrowings.
Characteristically, three of the particles in the above list: ALLAN, I FYNY and
I LAWR may convey the sense of completeness. The fact that this notion can be expressed in a variety of ways in Welsh, PVs of this type may be viewed as pleonastic.
Similarly, O GWMPAS in many figurative senses may appear redundant and in some
cases as almost semantically empty.
The above findings agree with the dispersed, mostly intuitive opinions of researchers, such as Fowkes’s (1945: 244) view that particles such as i mewn and i fyny are “entirely foreign” to the Welsh language, the “striking” items mentioned by Jones (1979) or
examples of particles “to be omitted” mentioned by P.W. Thomas (1996:561).
It has been stated by Rottet (2000: 115) that in contemporary Welsh a large proportion of calqued PVs is well-established and the items are no longer seen as tokens of
English influence. The analysis of the stylistic markedness has confirmed this observation. Spatial pleonastic constructions seem to occur naturally in Welsh and, in contradiction of P. W. Thomas’s claims (1996:561), they did not appear to be marked as informal
in the sample.
A few observations can be made about the motivations behind the calquing phenomenon. One of them is to fill lexical gaps and name new concepts, for instance in the
context of technological developments. Such is the case with the extension of meaning of
the particles ymlaen ‘on’ and i ffwrdd ‘off’ in PVs related to switching mechanical devices
210
or light. Other examples included the loan rendition llwytho i fyny, ‘upload’ or calques
logio/optio/clocio i mewn/allan ‘log/opt/clock in’. Such items seem to blend very naturally into Welsh as they convey figurative meanings which are fairly transparent and wellestablished in the language. On the other hand, as already mentioned by Rottet (2005:61),
some recent constructions are occasionally flagged with quotation marks, which indicates
that they are still not well-established. Such was for example the case of dod allan ‘come
out’ in the context of revealing homosexuality.
The phenomenon of calquing certainly goes beyond the need to fill lexical gaps
and stems from the bilingual reality in contemporary Wales. Translation between Welsh
and English may certainly affect the amount of calquing occurring in written texts. For
instance, the present study has indicated that journalists translating interviews with speakers of English or authors who try to portray English-speaking characters in Welsh prose
may tend to use more PVs alongside other borrowed constructions. The results also suggest that calquing is, or at least is thought to be, more widespread among members of the
young generation, who have less control over calquing. In her study of two language
communities, Jones has noticed the prevalence of PVs in younger speakers’ speech (1998:
86). Similarly, many calques and especially loanblends in the corpus were associated with
teenage language.
There is also some evidence of connecting PVs with nonstandard varieties of
Welsh. Such was the case of the novel Afallon which uses heavily Anglicised dialect to
portray a deteriorating urban community. Calqued PVs were also used as tokens of ‘bad’
Welsh in magazines for the purpose of satire or drawing attention to lazy, incorrect translations.
In general, however, there are few instances of a prescriptive norm against verbparticle constructions. The less marked PVs are freely used in press and fiction. Their
number depends on the level of formality – quality magazines and papurau bro were
more conservative in that respect than the popular magazine Golwg or the newspaper
Y Cymro. It should also be remembered, however, that journalists working in weeklies as
a rule work under a greater pressure of time while prepare texts than authors writing for
magazines. This, rather than a deliberate policy, may account for more instances of direct
translation. Only one magazine, Y Faner Newydd appeared to follow a policy of avoiding
PVs. This rather purist approach can be explained by a rather strongly nationalist character of the publication.
211
3.5. Conclusions
This corpus study presented in this chapter study attempted to fill the gap in the literature
on phrasal verbs by providing a basis for a detailed linguistic description of the phenomenon. The study has a number of limitations, such as the small size of the sample, underrepresentation of some registers and dialects and manual annotation of the corpus
which is naturally subject to errors and omissions. Moreover, the manual method of analysis did not allow the detection of a possible thematic bias.
This corpus study has provided abundant data for a detailed description of verbparticle constructions in Welsh, regarding their productivity and frequency of verbs and
particles, the syntax of transitive construction and stylistic markedness. In general, the
analysis has confirmed the findings of Rottet (2005), the only paper written specifically
on the subject thus far. It has been demonstrated that PVs are extensively used in semiformal varieties of Welsh and their presence cannot be explained solely by language contact. Idiomatic phrasal verbs, including prepositional verbs, have been present in Welsh
for many centuries and new items which are not directly translatable into English still
emerge.
The findings of the corpus study provide strong evidence to refute the claim that
PVs belong only to informal or non-standard registers of Welsh. It has been shown that
PVs are by no means a marginal phenomenon in contemporary Welsh. By metaphorical
extension of particle meanings, idiomatic phrasal verbs blend naturally into the language’s vocabulary, enriching it with new concepts. On the other hand, some of the novel
particle senses may be perceived as redundant and consequently certain PVs might become tokens of careless direct translation from English. However, there was little evidence of a prescriptive norm against phrasal verbs in the corpus.
The results of the study call for including verb-particle constructions in future
grammars of Welsh, as well as dictionaries and pedagogical materials. As noticed by
Rottet (2005: 62-66), the latter tend to be fairly inconsistent in the representation of PVs
or simply ignore them. Dictionaries and teaching materials will be investigated in more
detail in the next chapter of the thesis.
212
Chapter 4: Phrasal verbs in Welsh dictionaries and
teaching materials
4.1. Introduction
The findings presented in Chapter 3 suggest that phrasal verbs are extensively used in
contemporary written Welsh and although there is little evidence of a prescriptive norm
against them, some constructions are stylistically marked and may be perceived as nonstandard. This chapter aims to answer the question whether such norms can be observed
in metalinguistic contexts, primarily in textbooks, guidebooks and dictionaries of Welsh.
The question is crucial in the light of issues concerning learners and new speakers (see
1.1.2, 1.2.7), who are likely to use such materials to make conclusions about what is correct and acceptable in the Welsh language. It is worth exploring whether and to what
degree the two dimensions, namely the linguistic description and the corpus findings,
correspond.
A brief investigation of selected dictionaries has been conducted by Rottet (2005:
60-67), who analysed a number of printed dictionaries, a reference book (Jones 1976) and
a book for learners (Gruffudd 2000), concluding that the presentation of PVs tends to be
ambiguous and often put within the wider context of transfer from English. This chapter
will expand on and update Rottet’s findings, by investigating available course books,
teaching materials, printed and online dictionaries and other online resources. The final
section draws from the corpus study described in the previous chapter by discussing the
representation of the most frequent PVs in contemporary dictionaries of Welsh.
213
4.2. Teaching materials and courses
4.2.1. Analysis
For the purpose of the present study, an investigation of textbooks and teaching materials
for learning Welsh has been undertaken. The sample consisted of the following materials
available to me at the time the study was conducted:
a) course books and guidebooks, currently in sale or available in the National Library
and the Town Library in Aberystwyth;
b) free courses and materials for students and teachers of Welsh available online;
c) teaching materials obtained from Welsh for Adults Centre in Bangor,Canolfan
Bedwyr in Bangor and the National Centre for Teaching Welsh.
The texts were analysed for the presence of phrasal verbs and the way they are described.
Attention has also been paid to extracts concerning the phenomenon of calquing. Although the sample is not exhaustive, it contains a number of illustrative examples which
may serve as the basis for making general observations regarding the standardness of PVs
in Welsh pedagogy. The examples are described according to the type of text they represent.
In contrast to English-language teaching, the term phrasal verb/berf ymadroddol
is rarely, if ever, used in Welsh teaching materials. PVs are usually contained in sections
related to idioms and described accordingly as idioms or simply phrases. Since idioms
are as a rule introduced at more advanced stages of language learning, there were few
PVs encountered in textbooks for beginners and intermediate students. The books investigated were the popular 3-level WJEC (Welsh Joint Education Committee) series for
adults: Cwrs Mynediad [Entry course] (Meek 2005), Cwrs Sylfaen [Basic course]
(Stonelake and Davies 2006) and Cwrs Canolradd [Intermediate course] (Davies and
Conlon 2007) (South Wales version), and two textbooks created by Heini Gruffudd: Welcome to Welsh – A Complete Welsh Course for Beginners (Gruffudd 2006) and intermediate Cymraeg Da (Gruffudd 2000) and two parts of yet unpublished books prepared by
National Centre for Teaching Welsh which will be introduced into Welsh for Adults
courses starting from September 2018: Mynediad A1 [Entry A1] (Y Ganolfan Dysgu
214
Cymraeg Genedlaethol 2018a) and Uwch B2 [Upper B2] (Y Ganolfan Dysgu Cymraeg
Genedlaethol 2018b).
The WJEC series used in Welsh for Adults courses contain only a small number
of PVs. The first part of the course introduces only two rather transparent items bwyta
ma’s ‘eat out’ and mynd ma’s ‘go out’. The second textbooks adds two idiomatic PVs:
galw heibio ‘call by’ and torri lawr ‘break down’, while the intermediate level presents
several other idiomatic PVS: dod ymlaen ‘come on’ (about lights), dod draw ‘come over’,
edrych ar ôl ‘look after’, dod ymlaen ‘get on’ (with someone), torri ar draws ‘interrupt’
and edrych ymlaen at ‘look forward to’. The last PV appears most frequently in the book
as the phrase is recommended to be used in correspondence (“I am looking forward to
your reply”). Torri ar draws ‘interrupt’ is presented as a model phrase to explain complex
prepositions.
Gruffudd’s books (2000, 2006) do not have a separate section on PVs. However,
several idiomatic PPs are included in chapters on prepositions in Cymraeg Da: tynnu at
‘to get on for’, lladd ar ‘criticise’, torri ar, ‘interrupt’ (2000: 96, 98). Ymlaen is also
mentioned as an adverbial used in phrases: “we can use to translate on in some phrases,
especially in speech”69 (2000:103). The given examples are: bod ymlaen ‘be on’, troi
ymlaen ‘turn on’ (lights) and mynd ymlaen ‘go on’.
The unpublished books prepared by the National Centre for Teaching Welsh represent PVs in a similar way. The first part of the course, intended for beginners, contains
only two PVs: the quite transparent bwyta allan ‘eat out’ and ‘mynd allan’ ‘go out’ (for
entertainment), and also pleonastic torri i lawr ‘break down’. The upper intermediate introduces more idiomatic phrases: torri i mewn ‘break into’, enwi ar ôl ‘name after’, dod
ar draws ‘come across’, torri ar draws ‘interrupt’ and siarad i lawr ‘talk down’. The
latter is put in inverted commas, which may signal it is calqued from English70.
The analysed textbooks for beginner and intermediate learners did not include any
comments on Anglicisms in general and presented some frequently occurring PVs as elements of the spoken language, but also, in case of edrych ymlaen, used in writing. This
stands in rather sharp contrast to the remaining textbooks analysed, which were aimed at
students at the advanced level.
“Rydyn ni’n gallu defnyddio ‘ymlaen’ i gyfieithu on mewn rhai ymadroddion, yn enwedig ar lafar.”
“Mae pobl weithiau’n osgoi defnyddio ‘ti’ rhag ofn iddyn nhw ‘siarad i lawr’ â rhywun.” ‘People sometimes avoid using [informal] ‘you’ not to ‘talk someone down’(Y Ganolfan Dysgu Cymraeg Genedlaethol
2018b: 34).
69
70
215
The overview of these works will begin with a rather outdated but interesting example of a textbook Gloywi Iaith 3 (Department of Welsh Language and Literature, University of Wales 1988), a course for students who wish to improve their knowledge of the
literary language. The book contains three exercises with idioms “translated literally”
from English and the student is asked to provide a “proper” way to say them in Welsh
(1988: 36-38). Among them are the following PVs: dod i fyny ‘come up’ (of time),
gwneud allan ‘make out’, ‘invent’, mynd allan ‘go out’ (of fire), sefyll allan ‘stand out’
(be special), dwyn i fyny ‘bring up’, ffeindio allan ‘find out’, mynd ymlaen ‘continue’.
The third exercise puts PVs in the context of traditional prescriptive norms, referring to
the authority of the 19th-century author Emrys ap Iwan and his article “Plicio gwallt yr
hanner Cymry” (1939 [1889]) (see Introduction and 1.2.4.1). Students are asked to analyse examples of calqued idioms “condemned” by ap Iwan and match them with the
equivalents he proposed (1988:38). Among the examples are the following PVs: rhedeg
i lawr ‘run down’, llosgi i fyny, ‘burn up’ and gwneud am ‘make for something’, ‘go in a
particular direction’.
Similar judgements are to be seen in more recent publications. To begin with
books used in schools, many examples of PVs have been found in Seren iaith (Breese and
Clement 2011), a revision exercise book for young people studying Welsh to GCSE or
Advanced level. PVs are contained in two sections of the book: “Idioms and direct translation” (2011:78-83) and “Common mistakes” (2011:98). The first one opens with the
statement “Sometimes we translate English phrases into Welsh and forget the correct way
to say them in Welsh.” In the accompanying exercises students have to replace a number
of PVs with one-word equivalents or other phrases. Among these are: gwneud feddwl
lan/i fyny ‘make up one’s mind’, dwyn i fyny ‘bring up’, rhedeg allan/mas ‘run out of’,
troi golau bant/i ffwrdd ‘turn the lights off’, ffeindio allan/mas ‘find out’, rhedeg rhywun
i lawr ‘run someone down’, sefyll allan ‘stand out’, and dod ar draws ‘come across’. The
latter three are to be replaced not with single words but native idiomatic PPs or idioms:
lladd ar, ben ac ysgwydd yn well na’r gweddill, taro ar. The exercises in the book feature
imaginary characters such as a teenage girl Lisa “who is strongly influenced by English”
and her mother who tries to convince the girl that the Welsh she uses is wrong. Another
exercise mentions a teacher who “doesn’t have a good grasp on Welsh” using phrases
216
such as amser wedi dod i fyny ‘time has come’, lit. ‘time has come up’71 instead of mae’n
bryd ‘it’s time’ or ffeindio allan instead of darganfod. The second section containing PVs
is called “Common mistakes”. The introduction states: “Many English phrases contain
the words ‘up’, ‘down’, ‘out’, ‘in’ ‘off’. Do not translate them literally. Use Welsh phrases
e.g. troi i fyny > dod/ymddangos” (2011:98). The exercise contains the following PVs
and their equivalents: troi lawr, troi i ffwrdd, gwneud feddwl i fyny, cario ymlaen, pigo
lan, edrych i fyny, helpu allan and ysgrifennu i lawr.
Another set of PVs was found among resources for the Welsh Language Skills
Certificate prepared by Welsh National College (Coleg Cymraeg Cenedlaethol 2013).
The online course contains five sets of exercises in which students are asked to fill gaps
in a Welsh text by replacing English PVs with “correct” Welsh words. PVs are grouped
according to prepositions: off, on, over, in, out, up and down. In each exercise there is
a clickable explanation box informing about a tendency to translate English PVs into
Welsh word-for-word rather than use native vocabulary. The box on up and down additionally draws attention to pleonastic constructions, stating that it is unnecessary to use
prepositions in phrases such as eistedd i lawr, ysgrifennu i lawr, ffonio i fyny, torri i fyny.
Among the examples of PVs in this course are loanblends: ffeindio allan, sortio allan,
setlo mewn, cario ymlaen, pleonastic constructions such as helpu allan, and direct translations: rhoi mewn ‘give in’ and cymryd drosodd ‘take over’. In a text-correction exercise,
PVs are described as mistakes and placed along spelling and grammar errors such as
wrong mutations.
Similarly, a grammar course for teachers of Welsh available on the digital learning
platform Hwb (Roberts 2011) contains a section called “direct translation” in which students are to correct phrases containing PVs. These are also grouped according to particles:
i fyny (with verbs rhoi, mynd, troi, edrych, dwyn, goleuo, golchi, gwneud), allan (mynd,
rhedeg, darllen, gweithio, helpu, colli), i lawr (ysgrifennu, troi, gadael) and i mewn
(edrych, rhoi). Other two phrases to be replaced are directly translatable rhoi’r golau
i ffwrdd ‘put the lights out’ and rhoi dillad ymlaen ‘put clothes on’. The exercise also
includes examples of preposition stranding and direct translation of English verbs have
and go.
71
This appears to be a rather unusual example of a loan rendition and could, in fact, be influenced by the
expression in English “time’s up”.
217
Another type of resource referring to PVs are proficiency courses for native speakers and advanced learners known as Gloywi Iaith [Polishing the language]. Llawlyfr
gloywi iaith, textbook for the course issued by Bangor University, devotes one of its six
chapters, “Writing English in Welsh”, to the issue of translating from English (Bangor
University 2008: 44):
With so much English around us everywhere it is natural that it influences the kind of Welsh
we use. It is easy to hear English words used instead of Welsh words but we do not always
notice how English changes the patterns of speaking and writing in Welsh. These days we
translate directly from English without noticing. (…) Below we will focus on weaknesses
which are fairly easy to find and that you can try to avoid. 72
The above passage presents the inconspicuous influence of English as an undesirable
phenomenon, which can be deliberately avoided. The authors proceed to enlist major
“weaknesses” stemming from unconscious translation. The first section “translating idioms” is fully devoted to PVs. It mentions the tendency to translate phrases with English
adverbs up, down, out, in, off, on and around literally into Welsh. According to the authors, idiomatic PVs, although natural in English, are something “unwelsh”: “Unwelsh
sentences like the ones below have become considerably popular by now (…) Yet, it is
possible to make them Welsh in a completely natural way with a smallest amount of effort” (2008: 45)73. The examples provided of PVs and their equivalents are: edrych fyny
– chwilio, rhoi fyny – ildio, dwyn i fyny – magu, troi i lawr – gwrthod, dod allan – cael ei
gyhoeddi, edrych i mewn – ymchwilio, galw i ffwrdd – gohirio, cario ymlaen – parhau.
There is also an example of an expression mae si o gwmpas ‘rumour is around’ that should
be replaced with mae si ar led. The rest of the chapter refers to other contact-related phenomena in Welsh, such as preposition stranding, translating English prepositions and using conditional os instead of the relative particle a.
Another course for adults of the Gloywi Iaith type has been found on an online
platform, Y Bont (Davies 2016). It contains a section on “English idioms in Welsh”, consisting of tasks where students have to replace PVs with other phrases. Among the PVs
“Gyda chymaint o Saesneg o’n cwmpas ym mhob man, mae’n naturiol bod hynny yn dylanwadu ar y
math o Gymraeg yr ydym yn ei defnyddio. Mae’n hawdd clywed geiriau Saesneg yn cael eu defnyddio yn
lle geiriau Cymraeg, ond nid yr ydym bob amser yn sylwi sut mae’r Saesneg yn newid patrymau siarad ac
ysgrifennu Cymraeg. Erbyn hyn, byddwn yn cyfieithu’n syth o’r Saesneg heb sylweddoli hynny. (...) Isod,
byddwn yn canolbwyntio ar y gwendidau sy’n weddol hawdd eu canfod ac y gallwch chi geisio eu hosgoi.”
73
“Bellach, mae brawddegau anghymreig fel y rhain yn ddigon cyffredin (…) Eto, mae’n bosib Cymreigio’r
brawddegau hyn yn gwbl naturiol gyda’r mymryn lleiaf o ymdrech.”
72
218
and their equivalents are: gwneud ei feddwl i fyny – penderfynu; dal i fyny – rhwystro,
rhedeg lawr – lladd ar, dwyn i fyny – magu, mynd i lawr – gostwng, troi off – diffodd.
The section also mentions direct translation of English idioms (dros y lleuad – wrth ei
fodd, allan o wynt – a'i wynt yn ei ddwrn) and two other cases of calquing (mae gen i ofn
– mae arna i ofn, eistedd yr arholiad – sefyll yr arholiad).
Cymraeg Clir (Williams 1999: 49), a textbook helping users to become familiar
with different registers of Welsh (see 1.2.4.2) also recommends sensitivity to borrowings:
One of the biggest problems in the Welsh language is that everyone who speaks it is bilingual. A further problem is that the second language of most of us happens to be one of the
strongest languages in the world, that is English. In consequence English has an effect on
the way we think and the way we write. Although we do not translate sentences from English all the time, we often think in English and then make the same mistakes some people
make while translating.74
This is exemplified by a sentence containing tyfu i fyny ‘grow up’, where, according to
the author, the preposition is unnecessary. He proceeds to take special notice of PVs,
suggesting that it is only the non-idiomatic ones that are correct:
You need to be careful with any phrase which contains the English up e.g. look up, grow
up, wash up, put up etc. the only time we will use i fyny in Welsh is when we can to the
same thing i lawr too, e.g. we can look up at the moon or look down at our feet, so it is
acceptable in this sense. (..) Be careful with down, out, in, off, on and around as well when
they occur in English phrases which are translated. Ask yourself if you need a Welsh word
to render them.75
The use of PVs is advised against in other guides to writing as well. In Ymarfer
ysgrifennu Cymraeg [Written Welsh practice] (Thomas 2012) the chapter on idioms discusses a number of native Welsh expressions, showing their richness and explaining their
origins. This is followed by a section on “Translating” which consists of a table showing
Un o broblemau mwyaf yr iaith Gymraeg yw bod pawb sy’n ei siarad yn ddwyieithog. Problem bellach
yw bod ail iaith y rhan fwyaf ohonom yn digwydd bod yn un o’r ieithoedd cryfaf yn y byd, sef y Saesneg.
Mae’r Saesneg wedyn yn cael effaith ar y ffordd yr ydym yn meddwl a’r ffordd yr ydym yn sgrifennu. Er
nad ydym bob tro yn cyfieithu brawddeg Saesneg, rydym yn aml yn meddwl yn Saesneg ac wedyn yn
gwneud yr un camgymeriadau ag y bydd rhai’n eu gwneud wrth gyfieithu.
75
Rhaid bod yn ofalus gydag unrhyw ymadrodd sy’n cynnwys yr up Saesneg e.e. look up, grow up, wash
up, put up, a.y.y.b. Yr unig amser y byddwn yn defnyddio i fyny yn Gymraeg yw pan fyddwn yn gallu ei
wneud i lawr hefyd, e.e. gallwn edrych i fyny ar y lleuad neu edrych i lawr ar ein traed, felly mae’n dderbyniol yn yr ystyr yma. Yma, gwell fyddai ysgrifennu...wedi iddyn nhw dyfu... Byddwch yn ofalus gyda
down, out, in, off, on ac around hefyd pan fyddant mewn ymadroddion Saesneg sy’n cael eu cyfieithu.
Gofynnwch i chi’ch hun a oes angen unrhyw air Cymraeg i gyfateb iddyn nhw.
74
219
“acceptable” and “unacceptable” phrases with the introductory statement: “It is important
that we do not translate English idioms or expressions translated literally into Welsh since
we have our own way of saying them”76 (2012: 162). Out of 15 examples in the table,
there are 8 PVs: ffeindio allan, gweithio allan, rhoi ar, troi i lawr, mynd i fyny, gwneud
ei feddwl i fyny, edrych i fyny, ysgrifennu i lawr.
One of the most recent publications mentioning PVs is Iawn bob tro [Right every
time] (Jones 2016), another guidebook for writing in Welsh. The author’s attitude towards
borrowings is balanced:
It is easy to become neurotic about idioms. The truth is that there are enough sayings and
phrases which are word-for word identical in Welsh and English, and surely in other languages. The borrowing pattern settles in a language gradually so that we do not worry too
much about it; on the other hand, there are patterns to be heard that are second-hand and
still unnatural. It is probable that ‘rhoi i fyny smocio’ will not be eliminated anytime, although ‘rhoi’r gorau’ is still equally useful and familiar. But it is still clumsy to say ‘rhoi
rhywun i fyny dros nos’ [put someone up for the night] where it means give bed or accommodation to someone.77 (Jones 2016: 82)
The author does not prescribe against PVs as such but draws attention to those that, in his
view, are less well-established and natural. He proceeds to give more examples which
“have been with us for long and are probably to stay. They are not very painful, but undoubtedly one can offer more Welsh things” (2016: 83), among them dwyn i fyny (instead
of magu), cario ymlaen (instead of dal ati), gwnaeth ei feddwl i fyny (instead of
penderfynu), dangos rhywun o gwmpas (instead of dangos). However, in the author’s
opinion, the most “painful” token of calquing is translating full idioms.
“Y mae’n bwysig nad ydym ni ddim yn defnyddio idiomau Saesneg, na dulliau Saesneg o fynegi wedi
eu cyfieithu’n llythrennol i’r Gymraeg gan fod gennym ni ein ffordd ni ein hunain o ddweud.”
77
Hawdd yw mynd yn niwrotig ynghylch priod-ddulliau. Y gwir yw fod digon o ddywediadau ac ymadroddion air-am-air yr un fath yn Gymraeg ac yn Saesneg, ac mewn ieithoedd eraill hefyd mae’n bur sicr.
Mae patrwm benthyg yn cartrefu mewn iaith o dipyn i beth fel nad ydym yn poeni rhyw lawer yn eu gylch;
ar y llaw arall mae patrymau sydd i’w clywed yn ail-law ac yn annaturiol o hyd. Tebyg na ddisodlir ‘rhoi i
fyny smocio’ bellach, er bod ‘rhoi’r gorau...’ yn dal mor hwylus a chyfarwydd. Ond chwithig o hyd yw
‘rhoi rhywun i fyny dros nos’[, lle golygir rhoi gwely neu lety iddo.
76
220
4.2.2. Summary
of findings
The above overview demonstrates that, as far as Welsh pedagogical materials are concerned, idiomatic PVs are not presented as a distinctive class of verbs in the Welsh language or discussed in great detail. Popular courses for less advanced learners include single examples of PVs which are in common use without distinguishing between native
constructions and loan translations. In contrast, when it comes to sources for advanced
learners and/or native speakers, they place much focus on selected idiomatic PVs, presenting them as undesirable calques from English alongside other transfer phenomena,
such as preposition stranding. Since transfer phenomena are prevalent in the speech of
young people, it is unsurprising that the strictest guidelines against PVs have been found
in a textbook for schools, Seren iaith. It can be assumed that authors of such books see
a need to set clear rules of what is correct or not for the sake of young people. The analysed materials for teachers were less condemning in their approach, presumably aiming
to make teachers more language-conscious and sensitive to potential calquing. Similar
tendency can be seen in Gloywi Iaith-type courses, where authors endeavour to promote
natural, idiomatic Welsh, rather than forbid specific words.
The underlying idea behind the representation of PVs in Welsh teaching materials
is that transferred constructions pose a potential thread to native idiomatic items. The
analysis has demonstrated that many Welsh-language courses for proficient speakers aim
to retain native idiomaticity of Welsh, encouraging users to invest effort to identify
calques from English in their language, a task more difficult than avoiding code-switching
or using English vocabulary. The vast majority of “condemned” PVs are recommended
to be replaced with either one-word equivalents (mynd ymlaen – parhau) or in some cases
native PPs (taro ar preferred over dod ar draws, lladd ar over rhedeg i lawr), while in
the case of pleonastic PVs (tyfu i fyny, ysgrifennu i lawr) students are advised to drop the
“unnecessary” particle.
The PVs mentioned in the studied sources were often described in negative terms,
such as “incorrect”, “unnatural” and “unwelsh” or less negatively as “weaknesses”. The
main division line is drawn between the concept of a “Welsh” and “unwelsh word”, equating correctness with lack of English influence. Proficiency in the Welsh language is thus
associated with the ability to keep the two languages apart, according to the traditional
221
monolingual norm. According to Robert (2011:140) this is the currently the most prevalent standard language ideology in relation to Welsh. The tradition of condemning particular PVs appears to be rooted in norms dating back to the late 19th century, with the same
examples being passed from generation to generation of scholars.
Among the PVs prescribed against in the teaching materials the vast majority were
APVs. By comparing them with the findings of the previous chapter (see 3.4), which
identified marked particle senses, it can be seen that the two lists largely coincide. The
essential items occurring in teaching materials include the following particles:
allan indicating maximum boundaries, completeness of an action (see 3.3.2.1). The
particle was considered redundant (helpu allan – helpu) or equivalents of PVs were
proposed (ffeindio allan – darganfod, rhedeg allan – gorffen);
i fyny in most of its figurative senses (see 3.3.2.2). The particle was considered redundant (tyfu i fyny – tyfu) or equivalents of PVs were proposed (dwyn i fyny– magu,
edrych i fyny – chwilio);
(i) mewn (i) in its figurative senses, e.g. setlo mewn (see 3.3.2.5);
i lawr in most of its figurative senses (see 3.3.2.4). The particle may be considered
redundant (ysgrifennu i lawr – ysgrifennu) or equivalents of the PV are proposed (troi
i lawr – gwrthod);
o gwmpas (see 3.3.2.9) in its figurative senses e.g. dangos o gwmpas;
off, as in troi off. (see 3.3.2.7).
A few other examples could not be assigned to the above categories. Among them
were rhoi i ffwrdd and troi i ffwrdd, both referring to the concept of 'switching something
off, to be replaced with diffodd. There were very few cases of PVs containing ymlaen, the
most productive and frequent particle forming PVs. The most notable of those was cario
ymlaen; the fact that the PV is a loanblend with borrowed cario may possibly explain
why it has been seen as unacceptable.
A final point to be made is that most of the investigated materials do not explicitly
mention register variation or distinguish between written/spoken or formal/colloquial
Welsh. Authors of teaching materials and guidebooks do not make allowances for what
is acceptable in casual, informal speech. This partly stems from the fact that the majority
of the investigated sources are specifically concerned with standard language and aimed
222
at professionals, illustrating the existence of a linguistic norm which encourages the
avoidance of translation of English idioms. There is evidence that similar norm is introduced in schools, at the advanced stages of language teaching. The observed difference
between spoken-orientated and written-orientated materials illustrates the extent of diglossia in Welsh.
4.3. Books and dictionaries of idioms
A special type of reference book relevant for this chapter are dictionaries of idioms. Rottet
(2005: 62-63) has investigated two of them, by Cownie (2001) and Jones (2013 [2001])
and noticed that both authors overtly state in their introductions that they had intentionally
omitted idioms directly translatable into English. Jones justifies it by the fact that their
meaning is immediately clear to any Welsh speaker (2013: ix), while Cownie states that
many idioms borrowed from English are either unnecessary or “on the vague border between acceptable and unacceptable” (2001: viii). The latter author refers directly to PVs:
Welsh has taken over many English idiomatic phrases, often to the enrichment of the Welsh
language when the borrowed phrase extends the range of ideas that can be expressed succinctly, and sometimes to its detriment when the borrowed phrase does nothing more than
supplant or distort a perfectly good Welsh word or phrase. The phrases gwneud eich ffordd
(to make your way to) and gwneud i ffwrdd â (to do away with) are just two examples of
many English idiomatic phrases expressed in Welsh with words that, while not importing
an alien construction into Welsh, are at present on the vague border between acceptable
and unacceptable Welsh”. (Cownie 2001: viii)
By stating that the constructions are “not alien”, Cownie suggests that Welsh has its own
PVs. Indeed, although the dictionary generally omits PVs, a fairly large number of them
was included: bwrw allan, bwrw ymlaen, cadw draw, canu allan ohoni, codi allan, cymryd at, dal allan, dal at/wrth, dod ar draws, dod draw, dod dros, dod ymlaen, edrych
ymlaen, gafael allan, galw heibio, gyrru ymlaen, lladd ar, mynd ymlaen, palu ymlaen,
rhedeg ar, rhoi heibio, sefyll draw, sefyll dros, taro i mewn, torri ar draws, torri i ffwrdd,
torri i lawr, troi heibio, troi lan, tynnu ar ôl, and a few others. The vast majority of those
items belong to N/LR category, in particular items with the very productive particles
ymlaen and allan, as well as native draw and PPs. However, there is also a small number
of DT constructions, some of which are marked as dialectal e.g. “troi lan (S)/i fyny (N)”.
223
A similar representation of PVs is found in Jones’s Dweud eich dweud: A guide
to colloquial and idiomatic Welsh (2013). The small number of PVs included are mostly
N/LR items, among them constructions with: bwrw (ymaith, ymlaen), dod (â, at rywun,
ymlaen â), rhoi (am, heibio), mynd (â, ar, rhagddo, ymlaen, yn, yn ôl), torri (ar draws)
and troi (ar, heibio). Despite the focus on the colloquial spoken idiom, common PVs such
as edrych ymlaen, gweithio allan, ffeindio allan are not included. This policy has been
aptly summarised by Rottet (2005:63): “These sources on Welsh idioms were intended
by their authors (according to their prefaces) to help preserve authentic Welsh idioms.
This documentation is intended, it is said, to help counter a (perceived) modern tendency
in which Welsh speakers today, particularly young ones, increasingly dress English idioms in Welsh words and use few authentic Welsh idioms in their speech.”
Another reason might be the tradition behind compiling these dictionaries and the
sources the authors used. Both refer in their introductions to previous works, notably dictionaries of idioms by Cule (1971) and Jones (1975, 1987) and a book on prepositions
Y geiriau bach (Davies 1994). These works also tended to exclude PVs, except for those
which are not translatable into English. Jones’s books of idioms (1975, 1987) contain a
small number of native PV, such as torri ar (1975), codi allan, dal allan, darfod am, dod
ymlaen (1987). The entries are accompanied with examples from works such as the Four
Branches of the Mabinogi (codi allan), the Bible (darfod ar) and Gweledigaethau y bardd
cwsg (1703) (dal allan)78, thus proving the native status of the constructions. Regarding
the book on prepositions by Davies (1994), it includes a small number of idiomatic PPs,
among them tynnu at, mynd rhagddo, dod trwyddi. In one section, the author warns
against translating English prepositions (1994:160), which would justify the omission of
presumably calqued PPs.
One more publication worth mentioning is a recent collection of idioms and
phrases Ar flaen fy nhafod (Lewis 2012). The PVs included in the book are: bwrw ymlaen,
cadw draw, dal ar, dod dros, dweud ar, dwyn i fyny, galw heibio, gyrru ar, lladd ar, mynd
rhagddo, mynd ymlaen, rhedeg ar, rhoi i fyny, sefyll allan, sefyll i fyny, sefyll dros, sefyll
i lawr, taro ar, taro i mewn, torri allan/mas, torri i mewn, troi heibio, tynnu ar ôl, tynnu
ymlaen. Similarly to the books mentioned above, these are mostly N/LR items. However,
The author cites Pedair Cainc y Mabinogi edited by Ifor William, William Morgan’s Bible and Gweledigaethau y bardd cwsg edited by J. Morris Jones. Other bibliographical details are not provided.
78
224
some of them: dwyn i fyny, rhoi i fyny, sefyll allan/dros/i fyny/i lawr are directly translatable.
To sum up, those Welsh dictionaries of idioms which are currently available contain a relatively small number of PVs. The authors focus on native constructions, aiming
to preserve and popularise native Welsh idioms79. Cownie (2001) is the only author who
directly refers to PVs, recommending being sensitive to them and the few DT items in his
dictionary are marked as belonging to the spoken register. The most recent book of idioms
(Lewis 2012) appears to present a slightly more liberal approach, as it contains a couple
of PVs which, as shown above, are commonly ‘condemned’ in Welsh teaching materials,
e.g. dwyn i fyny ‘bring up’, sefyll allan ‘stand out’. Generally speaking, however, dictionaries of idioms are written within the normative rather than descriptive approach, trying
to encourage the users to draw from the richness of Welsh vocabulary to convey notions
that must not necessarily be expressed with constructions borrowed from English.
4.4. Cysill
Some restrictions against using PVs have been introduced into the spelling and grammar
checking tool Cysill. It is part of the Cysgliad software package with a compendium of
electronic dictionaries, which is purchased by numerous institutions in Wales (Prys 2016:
3261). Since 2009, the tool has been available online for free (Wooldridge 2011: 20) as
Cysill ar-lein (Bangor University 2018). It is one the most popular of the electronic resources produced by the Language Technologies Unit at Bangor University and as of
September 2015 the online version had over 600 daily users (Prys et al. 2016: 3263),
while over 10 000 licences of Cysgliad had been sold by 2008 (Prys 2009: 39).
In her thesis, Woolridge (2011) has investigated the possibilities of developing
Cysill to correct cases of interference from English which are particularly likely to occur
in translated texts. Referring to idiomatic phrases, she has found that the tool already
marks selected constructions as incorrect. As can be expected, among those are PVs, such
as rhoi lan and dwyn i fyny, to be replaced with rhoi’r gorau i and magu, or pleonastic
79
It is worth mentioning that, apart from the constructions listed above, all three dictionaries contain plenty
of Welsh prepositional constructions of the types which were excluded from this analysis (e.g. dal ati, wedi
canu arno, dod ato ei hun, see 3.1.1).
225
codi lan/i fyny to be replaced with codi (Woolridge 2011:18-19). In the course of the
present research, only one other example has been found: cario ymlaen to be replaced
with parhau; the checker will also not accept the borrowed particle off. None of the PVs
mentioned as ‘incorrect’ in the previous section is being marked by Cysill. Possible reasons for this are potential difficulties in distinguishing between idiomatic PVs and nonidiomatic items or free combinations by the programme. What is more, Cysill might not
be able to extract transitive PVs when the verb and particle are separated by an object.
On the whole, however, the choice of only a handful of PVs is rather inconsistent in view
of the fact that other evident loanblends such as sortio allan, ffeindio allan are accepted
by the checker.
4.5. Dictionaries – case studies
As discussed in 1.2.5., dictionaries play an important role in language standardisation on
the lexical level, first and foremost by including or excluding lexical items, tagging them
according to register, style and providing notes on usage.
Due to the bilingualism of Welsh speakers, Welsh-English and English-Welsh
dictionaries are characterised by asymmetry regarding their function. While Welsh-English dictionaries are used for both production and comprehension, the primary function of
English-Welsh ones is encoding rather than decoding, as they can be of assistance only
in composition and communication. In the context of PVs, these differences have been
pointed out by Rottet (2005:63) in his investigation of PVs in Welsh dictionaries:
Terms included in English-Welsh dictionaries inevitably reflect, to one degree or another,
the lexicographer’s views on what constitutes “good” or “appropriate” Welsh for the dictionary user to produce, whereas inclusion of a term in a Welsh-English dictionary simply
reflects the lexicographer’s views on the likelihood that the dictionary user will encounter
the term in question and will need information about it.
This policy seems to be true concerning most medium-sized dictionaries of Welsh,
aimed at learners. Regarding Geiriadur Prifysgol Cymru, the large historical dictionary
of Welsh, Rottet (2005: 64-65) has suggested that the policy towards less standard lexical
items has changed during decades of the dictionary’s compilation: early volumes of GPC
226
are more likely to omit PVs, while the last instalments reflect the idea that PVs are acceptable in spoken styles or informal registers, which is marked with the label ‘spoken’.
The same phenomenon has been observed in the course of analysis in 3.3. Similarly,
Rottet has also investigated the English-Welsh Geiriadur yr Academi and concluded that
there is no consistent policy towards PVs. When translating English verb-particle constructions, the dictionary might offer only non-PV Welsh equivalents, flag PVs with register tags, put the particle in brackets as optional or give a PV without any comment or
tag (2005: 63). Rottet notices that the criterion of how long a given item has been used in
Welsh may play a role in its inclusion or exclusion, but no consistent rule can be found.
He also pays attention to the fact that Welsh lexicographers might have not included these
PVs which are easily comprehensible due to space limitations (2005:67). Regarding the
last point, it should be noted that the economy criterion lost importance as the last decade
has brought an abundance of online Welsh dictionaries and digitalisation of the two largest printed ones. This is one of the reasons why it was considered worthwhile to reinvestigate Rottet’s observations on the lexicographic representation of Welsh PVs.
The investigation has been carried out in the form of case studies of selected PVs.
The analysis was concerned with the most frequent PVs in the corpus, i.e. occurring in
more than 10 concordances, counting syntactic and dialectal variants of the same construction. There were 23 such PVs, with 30 senses distinguished. The PVs have been
searched for in all major dictionaries of Welsh currently available: the historical semibilingual Welsh-Welsh-English Geiriadur Prifysgol Cymru (GPC) – online version; English-Welsh Geiriadur yr Academi (GA) – online version, large bilingual Geiriadur Cymraeg Gomer and its online counterpart Gweiadur (Gwe/Gom); online dictionaries geiriadur.net of University of Wales Trinity Saint David (TSD) and Geiriadur Bangor (GB);
printed medium-sized Geiriadur Mawr and its concise version Geiriadur Bach (GM);
small printed dictionaries for learners Collins Spurrell (CS), Welsh Learner’s Dictionary
(WLD) and Pocket Modern Welsh Dictionary (PMWD); Geiriadur Idiomau by Cownie
(2001) (GI). For a detailed description of the dictionaries see 1.2.6.
Table 18 gives the number of PVs found in each dictionary. As can be seen, three
of the investigated printed dictionaries Geiriadur Mawr, Geiriadur Bach and Collins
Spurrell do not include any of the selected PVs which is unsurprising as these dictionaries
generally do not contain phrases. This was also observed by Rottet (2005:63). The same
policy is followed by the online GB which had only 6 PVs from the list. The highest
227
number of PVs was found in the English-Welsh GA; it included 19 out of 23 selected
PVs and 21 out of 30 distinguished senses. The monumental GPC fared slightly worse as
it included approximately half of the items. Other, small or medium sized dictionaries
contained around one third of the studied PVs.
Table 18. Number and percentage of the most frequent PV in the corpus included in
Welsh dictionaries.
With senses distinguished
30
%
21
70%
17
57%
12
40%
11
37%
11
37%
10
33%
9
30%
6
20%
0
0%
0
0%
Total no. of PVs
GA
GPC
Gwe/Gom
TSD
GI
PMWD
WLD
GB
GM
CS
Main sense only
%
83%
65%
43%
48%
43%
35%
30%
26%
0%
0%
23
19
15
10
11
10
8
7
6
0
0
Table 19 presents the selected PVs and their representation in dictionaries, excluding those which did not contain any PVs (GM, CS). The five most frequent PVs were
included in the majority of dictionaries, except for sense (2) of mynd ymlaen. However,
no PV was included in all dictionaries and only two, torri i lawr and torri ar draws, were
found in 7 out of 8 dictionaries. The majority of other PVs (20) were included in half or
less of the dictionaries.
Table 19. The representation of selected PVs in dictionaries of Welsh.
Dictionary
Phrasal verb
English translation
edrych ymlaen
mynd ymlaen (1)
mynd ymlaen (2)
edrych ar ôl
torri ar draws
dod ar draws
mynd heibio
cario ymlaen
dod draw
edrych yn ôl
look forward
continue, progress
go on, happen
look after
interrupt
come across, find
go by, pass (of time)
carry on
come round, over
look back
GPC
x
x
x
GA
Gwe/
Gom
x
x
x
x
x
x
x
x
x
x
x
x
x
TSD
x
x
GB WLD
PMWD
GI
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
228
symud ymlaen
mynd allan
mynd allan (2)
eistedd i lawr
dod ymlaen (1)
dod ymlaen (2)
dod ymlaen (3)
dod ymlaen (4)
troi i fyny/lan
torri i lawr
torri i lawr
torri i mewn
bod ymlaen
bwrw ymlaen
rhoi ymlaen (1)
rhoi ymlaen (2)
sortio allan
tyfu i fyny
ffeindio allan
taro ar
TOTAL/30
move on, progress
go out (for enjoyment)
go out (date someone)
sit down
come on (progress, improve), get on (do well),
hurry
get on (have a good relationship)
come on (start working)
turn up
break down of machine
break down, lose control
break in(to)
be on
continue, make progress
put on (start machine,
lights)
put on weight
sort out
grow up
find out
come across, bump into
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
17
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
11
x
11
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
21
x
12
6
9
10
In order to disambiguate the above data, the studied PVs will be now described
separately to investigate their representation in the dictionaries in more detail, comparing
it with the data from the corpus study.
4.5.1. Edrych ymlaen
Edrych ymlaen ‘look forward’ was the most frequent PV in the corpus occurring in
161 concordances, usually as a PPV with at as the second preposition (125 tokens). It
occurred 14 times as an APV without the second preposition and 22 times as a PPV with
other prepositions: i (14), am (7) and ar (1). With regard to mode, 120 tokens occurred in
narrative and 41 in conversation.
The verb appeared in various registers, from dialectal informal variants such as
edrych 'mlaen, edrach ymlaen, dishgwl mla'n to literary conjugated forms such as edrychaf ymlaen. The high frequency of this verb in the corpus should be attributed to the fact
that it was used as a stock phrase in papurau bro while reporting about local events.
229
The variant edrych ymlaen i appeared only in press, with 13 out of 14 examples
coming from local bulletins and 1 example of a tweet quoted in a magazine. This points
to the informal, non-standard character of the variant, which may be a case of calquing
English to as i rather than standard at. The other non-standard variant with the particle
am is a more ambiguous case as it was used in a variety of texts: local papers, fiction –
both narrative and dialogues – and a newspaper interview translated from English.
The treatment of this most frequent PV by Welsh dictionaries shows considerable
differences. Quite surprisingly, there is no entry for edrych ymlaen in GPC, which may
be explained by the fact that the entry for edrych is rather old, dated to 1965. However,
the phrase edrych ymlaen (at) is used as a definition of synonymous verbs: gwrthrychu
‘look forward, hope’ (1975 entry), rhagdremio ‘look forward’ (1998), rhagsyllu ‘look
forward’ (1998) and ysbïo ‘look (forward)’ (2002).
However, edrych ymlaen is included in most of the other dictionaries: GA,
Gom/GW, TSD, WLD and PMWD. The latter two contain example sentences with the
verb in formal registers, for instance “Edrychaf ymlaen at glywed gennych” ‘I look
forward to hearing from you’ used in correspondence (PMWD). GA gives edrych ymlaen
as a translation of look forward but proposes it as second after the phrase disgwyl yn
eiddgar ‘ardently expect’.
4.5.2. PVs with ymlaen conveying progress
The second most frequent PV in the corpus was mynd ymlaen, with 94 concordances. Out
of 4 senses of the PV distinguished: 1) ‘continue, proceed’ (69 tokens), 2) ‘happen’ (23),
3) ‘play’ (of music) (1), 4) ‘pass, advance’ (1), only sense 1) and 2) were included in the
analysis. The PV occurred 7 times as a PPV with the preposition â and, similarly to edrych
ymlaen, verb and particle forms and grammatical forms used varied depending on the
register, e.g. formal aethpwyd ymlaen, dialectal mynd mlân, mynd mla’n.
Mynd ymlaen in sense 1) was found in TSD (‘proceed’, phrase “sy’n mynd
ymlaen” – ‘ongoing’), GB (‘proceed’) and PMWD (‘carry on’, ‘progress’). It is noticeable that the dictionaries do not translate the item directly into English as ‘go on’. In GA
mynd ymlaen is given as the translation of carry on, go ahead/along/before/forward/on,
proceed and progress. The construction was not included in GPC or Gom/GW. However,
230
in Gom/GWE the PV is used in the definition of the verb parhau ‘continue’ and GPC
contains an idiom mynd ymlaen yn wndwn ‘to get on extremely well’ (first recorded use
1761) in the entry for gwyndwn.
As regards sense 2) of mynd ymlaen ‘go on, happen’ (23 tokens), it occurred primarily in conversation in fiction, mostly in the context of a question “what’s going on?”
In several cases, the characters speaking were teenagers or members of the working class.
None of the dictionaries studied included this sense of mynd ymlaen. This points to the
nonstandardness of this PV in relation to standard digwydd ‘happen’. Indeed, GA proposes digwydd as the translation of the sentence “what's going on here?” – “be sy'n digwydd yma?”.
Sense 1) of mynd ymlaen competes with other PVs, cario ymlaen and bwrw
ymlaen. Cario ymlaen occurred 17 times in the corpus with 8 C and 9 N tokens. However,
it should be noted that two of the N tokens referred to a title of a popular novel
Cario ‘Mlaen, and 3 were part of the colloquial narrative of Afallon. This suggests that
cario ymlaen is generally associated with the informal spoken. The PV has been found in
GPC (‘carry on’) and TSD (‘carry on, carry forwards’), however it is not quite clear
whether the entry in GPC with the date 1793 refers to the idiomatic or transparent sense
of the PV. Other dictionaries propose alternative translations for carry on, including other
PVs: GA – parhau, dal ati, canlyn arni, mynd ymlaen, GWE/Gom – bwrw ymlaen, gyrru
ymlaen, PMWD – mynd ymlaen, parhau. GA includes cario ymlaen only in its more
transparent meanings, as one of the options for translating carry along, carry forward,
carry across.
Bwrw ymlaen, probably a native Welsh PV, occurred in 12 concordances, 8 times
as a PPV with the second particle â. Ten out of 12 tokens were used in narrative and the
PV occurred in various registers. It is not included in the GPC entry for bwrw dated 1953.
It is included in GA under the phrase “continue to work” in the entry for continue, alongside other equivalents (parhau, dal i, mynd/dal ymlaen â). It is also to be found in
Gwe/Gom (‘carry on’) and TSD (‘knock on, forge ahead, go ahead, proceed with’). GI
marks bwrw ymlaen as dialectal Southern and translates it as “to be getting on a bit in
age; to make progress”.
Another PV with the particle, symud ymlaen, was found in 14 concordances
(6N and 7C items). Although the PV is directly translatable into English move on, this
translation was not included in the dictionaries: in GA and GB it was found in the entry
231
for ‘progress’ (in GA together with mynd ymlaen). In GB the entry symud ymlaen is linked
to a dictionary of terminology (Termiadur Addysg), treated as a specialist term.
Out of the three PVs with ymlaen suggesting progress, mynd ymlaen in sense
1) and bwrw ymlaen were the ones most accepted by the dictionaries. Although cario
ymlaen was more frequent in the corpus than bwrw ymlaen, the latter was found in a larger
number of dictionaries. This can be explained by the native status of bwrw ymlaen in
comparison to cario ymlaen, which, as a loanblend, appears to be discouraged, despite
the fact that it is not a recent construction. Curiously, none of the 3 PVs is present in GPC
as a separate entry, although mynd ymlaen was found as part of an idiom. Mynd ymlaen
in sense 2) appears to be marked and was not included in dictionaries presumably as a
calque.
4.5.3. Other construction with ymlaen: dod ymlaen, bod ymlaen, rhoi ymlaen
Ymlaen was the most productive particle in the corpus, and several other PVs containing
this particle ranked high on the frequency list. To begin with dod ymlaen, it appeared in
13 concordances (7 N, 8 C) primarily in fiction and informal contexts, with colloquial
variants such as dod 'mlaen, dwad/dŵad ymlaen, dod mlân. The PV is highly polysemous;
the following main senses were distinguished within the sample: 1) ‘get on (do well)’,
‘come on (progress, improve)’ 2) ‘come on (hurry)’, 3) ‘get on (have a good relationship)’, 4) ‘come on (start working)’. Sense 1) of the PV is recorded in GPC (‘make progress, get on’ 1735), GA (‘get along’, ‘progress’), Gwe/Gom (‘to get on, develop, succeed’), TSD (‘make progress’) and GI ‘get on, to do well, to succeed’. The examples in
GA include both informal and formal registers: “dw i’n dod ymlaen” (‘I’m getting along’),
“mae'r claf yn dod yn ei flaen yn dda” (‘the patient is progressing favourably’). Sense 2)
of the verb – the phrase “come on” used to tell someone to hurry has been found in
Gwe/Gom, GA, WLD, and also PMWD in the entry for ymlaen. The less frequent senses
3) and 4) are not recognised by dictionaries.
Ymlaen was also used with the verb bod to indicate the state of something being
“on”, “happening”, “working”. This construction has been found only in GPC entry for
ymlaen ‘be on, of clothes, switch, light, event’ (marked as spoken, entry dated 2002) and,
232
less directly, in GA, which gives bod ymlaen as an option to translate come on: “the film
comes on at eight o clock” – “mae'r ffilm [ymlaen] am wyth o'r gloch”.
Rhoi ymlaen ‘put on’ was found in 11 concordances, primarily in conversation
(10 tokens). The main sense of the PV was to ‘switch on a device’ (10 tokens); once it
was used in the sense of ‘put on weight’. Both senses are included in GPC (entry dated
1998), marked as spoken: ‘put light, television, kettle on’, ‘put on weight’’. GA contains
only sense 1) (translating the phrase ‘put the light on’). Interestingly, the entry contains
a note prescribing against using preposition ar or arno to convey on. It also proposes
cynnau’r golau as the second equivalent. This is an interesting case of the dictionary recommending a directly translatable PV (rhoi ymlaen) over another one which is an evident
calque (rhoi ar). Other dictionaries do not contain rhoi ymlaen; however, WLD has a similar PV troi ymlaen ‘turn on’, which is also in GA together with other equivalents. TSD
and Gwe/Gom suggest cynnau to translate English turn on.
As for sense 2), it is absent from the vast majority of dictionaries; Gwe/Gom and
GI suggest a number of other equivalents for put on weight (enill/magu pwysau, magu
bloneg/bol, rhoi pwysau, dodi pwysau ymlaen).
4.5.4. Edrych ar ôl
The third most frequent PV, edrych ar ôl presents an interesting case. There were 35
concordances of the PV, about two third of which (22) appeared in conversation, with
dialectal variants of edrych such as: edrach, ‘drych, disgwyl, shgwl. In narrative, it appeared in novels, papurau bro and the satirical magazine LOL, but not in quality press.
This evidence from the corpus would put edrych ar ôl in the middle of the formality
continuum.
The GPC entry dated 1965 contains edrych ar ôl: “to look after, mind, keep an eye
on”, giving 1853 as the date of the first recorded use. WSD and Gom/Gwe give edrych
ar ôl to translate look after, the latter along with gofalu am and gwarchod. However, two
other dictionaries advise against using the PV. The entry in GA proposes a set of other
expressions: gofalu, ymorol am rywun, gwarchod rhywbeth, cymryd gofal o rywbeth and
dialectal Southern carco, and ends with a note: “the Anglicism edrych ar ôl rhth is in
common use but not recommended”. Similarly, PMWD has no entry for edrych ar ôl in
233
the Welsh-English part and suggests gofalu am as the equivalent of look after with a note:
“edrych ar ôl is used but some consider it wrong”. Also TSD and GB translate look after
only as gwarchod and/or gofalu am.
It appears that despite high frequency of edrych ar ôl there exists a prescriptive
norm against it visible in some dictionaries. There is no such prescription, however, in
the most recent GI and Gom/Gwe.
4.5.5. Native PPs: torri ar draws, dod ar draws and taro ar
Two PVs with the particle ar draws: torri ar draws and ddod ar draws ranked fifth and
sixth in terms of frequency in the corpus. Torri ar draws, ‘interrupt’, lit. ‘cut across’ had
35 concordances in the corpus. Being a stock phrase used in novel dialogues (“he interrupted”, etc.), it appeared almost exclusively in narrative (34 concordances). This native
PP occurred in all the studied dictionaries except for GB. GPC dates its first recorded use
to 1788, but also marks it as “spoken”. GA gives torri ar draws as the first translation of
interrupt, next to ymyrryd, siarad ar draws; TSD contains an example phrase in a considerably formal register. This evidence suggest that the dictionaries consider the phrase
well-established and accepted.
Dod ar draws ‘come across’ occurred 30 times in the corpus, primarily in narrative (24 tokens). In one case it conveyed the sense of ‘cross’ rather than ‘find’. The item
is included in GPC, with its semi-idiomatic sense going back to the 14th century (see
3.3.2.12). GWE/Gom contains the PV in its English-Welsh part together with other equivalents of ‘come across’: dod ar warthaf, darganfod, cael hyd i, taro ar. The PV is also
included in TSD, GI and the English-Welsh part of PMWD. This demonstrates that although dod ar draws is directly translatable into English it is not regarded as an Anglicism,
which proves the aforementioned native status of PVs with ar draws (see 3.3.2.12).
The PP taro ar conveying a similar sense of ‘finding’, coming across’ had 11 concordances in the corpus, mostly in conversation in the sense of ‘bump into someone’. In
narrative, the PV usually referred to inanimate objects. The PV occurred in both formal
and informal registers. It is included in GPC (first recorded use 1603, date of entry 2001),
translated as ‘come across, chance upon, meet with, hit upon, find’ and marked as ‘spoken’. It has also been found in GA (‘come across’, ‘chance upon’, ‘bump into’, ‘meet’),
234
TSD (‘come across’, ‘chance upon’), Gwe/Gom (‘come across’, ‘hit upon’) and
GI (‘bump into’, ‘meet by chance’).
4.5.6. Mynd heibio
As can be seen, the three PPs – torri ar draws, dod ar draws and taro ar – were included
in the majority of Welsh dictionaries, presumably due to their native status. Another item
that can be considered native despite being directly translatable is mynd heibio ‘go by’,
’pass’. There were 29 tokens of mynd heibio, with 26 used in narration. The PV was
always used in the context of time and occurred in both higher and lower registers. Despite relatively high frequency, the PV has been found in only two dictionaries:
GB (‘elapse’, ‘pass’) and GA (‘go (of time)’). PMWD and TSD included the PV only in
its transparent sense of ‘go past’(someone).
4.5.7. Semi-idiomatic PVs: dod draw, edrych yn ôl, mynd allan gyda, torri i mewn
Four other PVs on the list: dod draw, edrych yn ôl, mynd allan gyda and torri i mewn are
also rather transparent, semi-idiomatic constructions. This probably accounts for the fact
that all four were generally absent from dictionaries. Dod draw is mentioned only in
TSD (‘come across’) and GA (‘come round’). In the latter it is placed alongside dod
heibio to translate an informal sentence “come round and see me” (“tyrd (dewch)
draw/heibio i’m gweld”). Edrych yn ôl, occurring 16 times in the corpus (9 N and 5 C tokens), was found only in GA under ‘look back’. The next PV, mynd allan, appeared in 16
concordances in the corpus with two senses: ‘go out’ (for enjoyment) and mynd allan (â)
‘go out (with), ‘date, have a relationship’. The first sense was not explicitly recorded in
any dictionary, although GA, WLD and TSD translate mynd allan as go out without specifying the meaning. The second sense was found only in Gwe/Gom, together with canlyn,
bod yn gariad i. Both TSD and GA contain the phrase ‘go out with someone’, but suggest
canlyn as the Welsh equivalent. Finally, torri i mewn ‘break into’ was found in 12 concordances (5 C and 7 N), with 11 referring to robbery and 1 in the sense of appearing
(“heddwch yn torri i mewn i'r presennol” – ‘peace breaking into the present’). The main
235
sense of the PV was included only in the two large dictionaries: GPC (‘break or burst
in(to)’, first recorded use 1567, entry dated 2000) and GA ‘break into a house’. The item
has not been found in GWE, TSD, PMWD or GI, but it is included in GB as a verbal noun
‘housebreaking’ and in WLD as the equivalent of break in, without specifying the meaning.
4.5.8. Pleonastic PVs: eistedd i lawr, tyfu i fyny and torri i lawr
There were 15 tokens of eistedd i lawr ’sit down’, in the corpus, occurring in both conversation and narrative modes (7 and 8 tokens, respectively) and in various degrees of
formality. In the dictionaries, however, there is evidence of prescriptive norm against
the PV. GPC gives 1567 as the first recorded use of the item, but also tags it as “common
spoken”. In GA, in turn, eistedd i lawr is marked as informal and an example sentence in
higher register “please sit [yourself] down” is translated as “eisteddwch os gwelwch yn
dda; a wnewch chi eistedd?; dewch i eistedd”, omitting the particle. Similarly, WLD and
the English-Welsh side of PMWD proposes eistedd as the equivalent of sit down. These
dictionaries suggest that the use of i lawr is informal and pleonastic.
Tyfu i fyny ‘grow up’ presents a similar case. There were 11 concordances of the
verb in the corpus. The dialectal variant of the particle, lan, occurred 7 times. The PV was
used mostly in conversation (8 tokens), in informal contexts. The PV is included in the
largest dictionaries, GPC, GA and Gwe/Gom. The historical dictionary gives 1691 as the
date of the first recorded use (entry dated 2002) and also provides spoken informal examples. In GA the PV is also marked as informal and dod i oed ‘come to age’ is chosen as
the main translation of grow. Gwe/Gom gives both dialectal variants of the PV
(i fyny/lan). The PV is absent from TSD, GB and GI in both Welsh-English and EnglishWelsh parts, while PMWD and WLD propose tyfu, without the particle, as the equivalent
of grow up.
Another supposedly pleonastic PV in the corpus was torri i lawr, which occurred
in 11 concordances in two senses: ‘break down (of a machine)’ and ‘have a nervous
breakdown’. It occurred only in fiction, in informal contexts. The informality is also
acknowledged by GPC, in which these senses of torri i lawr are marked as spoken without
the date of the first attestation; other more transparent senses date back to the 17th and
236
18th centuries. The PV is also included in GA, Gwe/Gom, GB and in the English-Welsh
parts of WLD and PMWD. WLD, however marks the particle as optional. Only
TSD avoids the PV, suggesting other equivalents of break down: datgymalu, diffygio,
methu’n llwyr, mynd i’r gwellt. The PV also appeared twice as a transitive construction
conveying the sense of ‘destroying something completely’. This sense was not recorded
in any dictionary.
4.5.9. Calques and loanblends: troi i fyny, sortio allan, ffeindio allan
Finally, the list of most frequent PVs contains 3 constructions which are evident calques
from English: troi i fyny ‘turn up’, sortio allan ’sort out’, ffeindio allan ‘find out’. Troi
i fyny ‘turn up’, ‘appear’ occurred in 14 concordances, 8 C and 6 N, mostly in informal
context, similarly to its English counterpart. The PV, together with its dialectal variant
troi lan, is to be found in GPC, marked as spoken (first recorded use 1858, entry dated
2001). In the entry turn up in GA the PV is evidently avoided and a number of other
equivalents is proposed (dangos eich wyneb, ymddangos, cyrraedd). Two dictionaries include only a dialectal variant, troi lan: TSD – ‘to appear on the scene’ and Gwe/Gom –
‘turn up’. However, both indicate the informality of the construction – Gwe/Gom tags it
as “mainly informal”, while TSD suggests the verb cyrraedd to translate the more formal
sentence “he turned up ten minutes late” – “cyrhaeddodd ddeng munud yn hwyr”. GI and
PMWD include both variants of the verb.
There were 11 concordances sortio allan ‘sort out’ in the corpus, with variants
sortio mas, sorto mas. The PV appeared almost exclusively in conversation and in several
instances was stylistically marked, used by teenagers and members of the working class
in novels as well as in parody in the satirical LOL. The PV is included in GPC, marked
as spoken (entry dated 1999), but it is absent from all other dictionaries. GA and PMWD
give datrys ‘solve’ as the equivalent of sort out.
Ffeindio allan, ‘find out’ present a very similar case. All the 11 tokens of this PV
were used in conversation, mostly in fiction in the speech of teenagers or characters ‘really’ speaking English. Dialectal variants in the sample were ffeindio mas, ffindo mas,
ffindio mas. As with sortio allan, ffeindio allan is recorded only in GPC (entry dated
1967), which distinguishes two senses of the verb: 1) ‘find out’, 2) ‘discover’, both first
237
recorded in the 18th century. Other dictionaries propose other equivalents of English find
out: darganfod (GA, WLD, PMWD) and datgel, datrys (GA).
4.5.10. Summary of the findings
The above findings allow to make some general conclusions regarding the representation
of PVs in dictionaries of Welsh. The majority of Welsh dictionaries contain a relatively
small number of frequently used PV, due to a small number of phrases in general. Older
printed dictionaries, such as Geiriadur Mawr and Collins Spurrell do not include any
phrasal verbs at all. The more transparent verb-particle constructions are likely to be excluded apart from the largest dictionary, GA and GPC. This can be explained by the fact
that such items are easily comprehensible for dictionary users.
The historical dictionary GPC recorded about two third of the selected PVs. The
criteria for the inclusion of PVs in this dictionary have been found to be unclear. The date
of compiling the entry may play a role, such as in the case of excluding native bwrw
ymlaen, but there were cases which suggest otherwise, for instance quite inexplicable
exclusion of edrych ymlaen and mynd ymlaen, the two most frequent PVs in the corpus.
The dictionary that included an exceptionally high number of PVs was GA. This
stems from the dictionary’s policy of representing a wide spectrum of registers and styles.
On the other hand, the same dictionary occasionally marks PVs as incorrect. Such instances were rare, however. The policy followed more often by GA and other EnglishWelsh dictionaries (WLD, PMWD, TSD) is that of exclusion or proposing other equivalents. This happened mainly in the case of loanblends and pleonastic constructions. In
some instances, such PVs were included in the wordlist but marked as informal. This
agrees with the findings of the previous chapter regarding stylistic markedness of certain
items. Another group of constructions which may be excluded from dictionaries are semantic loans, i.e. new senses of well-established PVs which appear to be modelled on
English, such as mynd ymlaen ‘happen’, mynd allan gyda ‘go out with’. The above case
studies suggest that the main criterion of exclusion of a PV from dictionaries is the existence of an alternative Welsh phrase to express the same concept. The case of edrych ar ôl
shows that even a well-established and immensely popular phrase can be prescribed
238
against in order to retain native Welsh expressions, such as gofalu am, gwarchod. Similarly, extended senses of the very productive particle ymlaen, such as rhoi ymlaen, are
discouraged due to the fact that they can be expressed in another way.
The study has also revealed a group of well-established items. A number of native
PVs, such as mynd ymlaen, torri ar draws are generally accepted by the dictionaries,
which coincides with the findings of the corpus study. With regard to registers, although
many PVs are rather informal in character, some constructions, in particular those with
the most productive particle ymlaen (symud and edrych ymlaen) were shown to belong to
formal contexts (official correspondence, specialist vocabulary), similarly to English.
4.6. Conclusions
The investigation presented in this chapter has demonstrated a twofold approach to idiomatic phrasal verbs reflected in Welsh pedagogy and lexicography. On the one hand, it is
clear that idiomatic phrasal verbs as a class are not entirely outside the linguistic norm
and standard vocabulary of Welsh. The study has indicated that certain frequent constructions, such as edrych ymlaen at, symud ymlaen are shown to occur in formal and official
as well as informal registers of Welsh. Moreover, a number of native PVs – in particular
PPs, but also APVs – were accepted in the investigated sources, although occasionally
marked as informal. The analysis has thus confirmed the observations made in Chapter 3 in that the investigated materials contained a meaningful number of accepted, wellestablished PVs, including both native and borrowed constructions. The latter are most
often accepted in the cases when they enrich Welsh vocabulary with new concepts, especially when the meaning of the particle is natural to the Welsh language. Moreover, teaching materials for beginners which concentrated on spoken Welsh contained a small number of popular PVs and in general did not include any prescriptive comments regarding
transfer from English. It can be assumed that authors of these sources strive to avoid
manifesting ideologies which could discourage non-native speakers from learning the
Welsh language.
On the other hand, sections on translation from English with particular focus on
PVs were found in a considerable number of sources which focus on standard written
language, created for proficient users, such as tutors and native speakers or advanced
239
learners who want to improve their Welsh. Many of the analysed pedagogical materials
manifested an approach which could be described as protectionist. Calqued PVs are described as lexical errors or tokens of ‘bad’ Welsh, associated with the language of teenagers and learners. Consequently, authors of the studied sources strive to make Welsh
speakers more aware of the influence English might have on their speech and present
expressions alternative to the borrowed ones. These recommendations generally do not
take a form of negative, forbidding prescriptions, but rather positive encouragement to
draw from the richness of Welsh vocabulary.
This tendency of presenting PVs as the most prevalent case of transfer phenomena
and warning against them is visible also in Welsh lexicography, in both general dictionaries and dictionaries of idioms. It has been found that major Welsh dictionaries attempt
to preserve native Welsh idioms by suggesting alternative ways of expressing English
PVs in Welsh. This policy can be described as normative rather than prescriptive as there
were few instances of dictionaries explicitly forbidding the use of PVs. They occurred in
dictionaries for learners and Geiriadur yr Academi, which have a clearly educational purpose. Similar was the case of dictionaries of idioms, which deliberately exclude many
common PVs if the authors consider them to be calqued from English. It is slightly more
surprising that omissions of very frequent PVs were also found in the historical GPC,
which could be expected to record all Welsh phrases in use. It has also been noted that
works compiled recently by the lexicographer G. Lewis (Gwe, Gom, Ar flaen fy nhafod)
appear to be more liberal towards borrowed constructions than other investigated dictionaries. Similarly, the attitude to PVs in guidebooks for writers varied depending on individual attitudes of authors, from deeming some constructions “unacceptable”
(Thomas 2012) to advising against being oversensitive to borrowings (Jones 2016).
These differences show that many directly translatable items are, to quote Cownie
(2001: viii) “on the vague border of acceptable and unacceptable”, which in some cases
gives rise to discrepancies between norm and usage. Generally, however, such cases were
relatively rare as many items discouraged in the analysed materials, such as loanblends,
were also stylistically marked in the corpus. The greatest difference between the findings
of the corpus study and the analysis of lexicographic and pedagogical materials occurred
with constructions such as tyfu i fyny ‘grow up’, torri i lawr ‘break down’ and eistedd
i lawr ‘sit down’. While in the corpus these PVs were among the most frequently used
240
and not stylistically marked phrases, they were often deemed incorrect in teaching materials and dictionaries as pleonasms. It can be assumed that while such constructions may
appear illogical to lexicographers and teachers with high metalinguistic awareness, ordinary users do not perceive them as anything striking. All in all, in spite of some attempts
to discourage the use of certain PVs in Welsh, the findings of this chapter confirm Rottet’s
conclusion that the phenomenon of PVs is “clearly too well entrenched at present for
anyone to realistically think of eradicating it altogether from the language” and therefore
“the policy (…) of accepting PVs as part of Welsh speech but urging that people strive
for better Welsh idiom in writing is a level-headed approach” (Rottet 2005:66).
The analysis presented in this chapter has demonstrated that despite some differences in attitudes of different authors, there exists a rather consistent linguistic norm regarding the status of PVs in standard written Welsh to which speakers are exposed
throughout educational process (school and higher education) and also in their professional lives (proficiency courses, dictionaries and other language tools, guidebooks of
style). The question that arises is whether professional users of semi-formal varieties of
Welsh find it manageable to adhere to these standards in the present bilingual reality and
the growing influence of English. This issue will be investigated in the empirical study
presented in the next chapter.
241
Chapter 5: The acceptability of phrasal verbs among
professional speakers of Welsh – a field study
5.1. Introduction
One of the primary aims of this dissertation is to investigate the integration of idiomatic
phrasal verbs within the linguistic norm in Welsh, considering various registers of the
language with focus on the semi-formal varieties, which, as shown in Chapter 1, appear
to be the most salient source of the current linguistic norm. Chapter 2 has presented the
current state of research on PVs in Welsh, demonstrating that these constructions have
been largely absent from any linguistic description of the language or treated as a marginal phenomenon, with some scholars describing PVs as redundant borrowings from
English or nonstandard elements of colloquial varieties. These claims were verified in the
course of a corpus study in Chapter 3, which showed that numerous idiomatic PVs are
common also in semi-formal varieties of Welsh and stylistically unmarked. Furthermore,
the analysis of transfer mechanisms based on the cognitive model showed that Welsh
possesses a number of native verb-particle constructions. Moreover, there are grounds to
believe that a substantial number of PVs have emerged as a result of natural processes of
metaphorical extension of particle meanings. Nevertheless, contact with English, in particular direct calquing, has been found to significantly reinforce the usage and productivity of verb-particle construction. As demonstrated by the semantic analysis, these mechanisms are clearly not random – while some particle senses are highly productive, others
appear to be marked as calques from English. It is the latter constructions that are associated with non-standard varieties of Welsh, as proven by their stylistic markedness in the
corpus.
Chapter 4 has examined the representation of PVs in teaching materials and dictionaries. The study has revealed that numerous teaching materials aimed at advanced
users of Welsh strive to sensitise Welsh native speakers and learners to the phenomenon
242
of calquing, presenting many PVs as incorrect, careless translations from English. This
prescriptive norm was less visible in Welsh dictionaries, which in most cases simply
avoided PVs by excluding them from wordlists and – in case of English-Welsh dictionaries – by proposing alternative equivalents. In fact, the existence of another Welsh phrase
which could express the same concept appeared to be the main criterion for exclusion of
a PV, reflecting lexicographers’ endeavours to retain native expressions. Verb-particle
constructions were usually accepted when they enriched Welsh vocabulary with new concepts, whilst loanblends and pleonastic constructions were most frequently avoided or
prescribed against. In the case of pleonastic constructions in particular, the lexicographic
norm disagreed with the observed usage. It has been suggested that this is due to high
language awareness of specialists who perceive such constructions as ‘illogical’.
Despite the incongruities, Chapters 3 and 4 have demonstrated a relative consistency of the linguistic norm regarding the status of PVs in standard written Welsh, to
which speakers are exposed in the process of education and possibly later on in their
professional lives. However, the question whether proficient speakers of today are aware
of such norms remains open. The study presented in this chapter endeavoured to investigate the homogeneity of the linguistic norm in Welsh among its recipients and encoders
to see how deeply monolingual-oriented ideologies permeate. Measuring speakers’
awareness concerning borrowings is a way of investigating the effects of contact processes such as calquing (Nelde and Weber 2000: 188). Bearing in mind that direct translations of PVs occur at a deeper level of transfer than obvious lexical borrowings, it was
considered worthwhile investigating whether proficient users of Welsh do indeed perceive PVs as non-standard tokens of interference from English, as would be suggested by
the resources analysed in Chapter 4.
This chapter, therefore, aims to investigate the matter drawing from the results of
a field study conducted in Wales in the years 2015-2016, which comprised 55 interviews
with professional speakers of Welsh, accompanied by questionnaires on the acceptability
of verb-particle constructions.
243
5.2. Study background
5.2.1. Lexical acceptability
The study presented in this chapter is mainly concerned with the integration of verb-particle constructions in Welsh by investigating the declared acceptability of selected phrasal
verbs among language-aware speakers. Acceptability, is understood here as “the extent to
which linguistic data would be judged by native-speakers to be possible in their language”
(Crystal 2008:4). Due to the fact that informants’ judgements can be influenced by a number of factors, such as linguistic intuitions, differences between dialects and the degree of
prescriptivism they have been exposed to, one can expect lack of agreement among speakers regarding the acceptability of given items (Crystal 2008:4). In the study presented in
this chapter the participants were asked to assess the acceptability of selected PVs in a
two-part questionnaire in 1) semi-formal writing when compared to their equivalents and
2) across different registers of the language.
Language awareness is defined here as “informed, sensitive and critical response
to the use of language by oneself and others, including the awareness of relevant terminology (metalinguistic awareness)” (Crystal 2008: 266). The term, used primarily in educational linguistics, in the present context will denote linguistic sensitivity of speakers
who may or may not accept Welsh verb-particle constructions in a given context. Linguistic awareness of the speakers regarding the use of borrowings in Welsh has been
investigated in semi-structured interviews in order to draw a profile of the investigated
group and identify speakers’ beliefs which might influence their lexical choices.
5.2.2. Previous research
Studies of acceptability have been conducted in fields such as theoretical linguistics, L1
and L2 acquisition, communicative disorders and L1 attrition (Altenberg and Vago 2004:
105) and have been mostly concerned with grammaticality judgements and syntax. Such
studies generally do not investigate speaker’s conformity to prescriptive norms, but their
linguistic competence; grammaticality is commonly measured with scales of acceptability
244
, as (Bard et al. 1996). Research on lexical acceptability is much scarcer, although the last
decade saw a number of corpus-based studies on the integration of lexical borrowings
from English in various European languages, which extended the scope of investigation
from single items to phraseology (Furiassi et al. 2012). As well as emphasising the quantitative, corpus-based analysis, recent studies on lexical borrowing gathered in Zenner
and Kristiansen (2014) have adopted an onomasiological, concept-based approach. In
contrast to traditional semasiology, which takes the borrowed word as point of departure,
the onomasiological approach focuses on alternative lexicalisations of a particular concept; in other words not only the loanword but also possible recipient language equivalents are taken into consideration, which allows for more in-depth, variationist analysis
(Zenner and Kristiansen 2014: 5-6). Qualitative empirical studies on the acceptability of
lexical borrowings are extremely rare (Zenner and Kristiansen 2014: 10) and, as in the
present research, have been used to supplement corpus-based studies (e.g. survey on borrowings by Soares da Silva (2014) comparing Standard and Brazilian Portuguese).
The acceptability of borrowings in relation to linguistic norms and ideologies has
been little investigated in Welsh linguistics. As described in 1.3.3, research on lexical
borrowings in Welsh has been generally very scarce, with few empirical studies touching
on the subject. A major study related to the issue of ideological stances behind the use of
borrowings is Robert’s (2013) thesis on the implementation of lexical planning initiatives
for Welsh, i.e. attempts to codify and standardise Welsh vocabulary. As part of her research, the author investigated naturally-occurring talk in a media institution (the S4C
television channel) and a workplace (a Housing Association), observing the linguistic
ideologies manifested by speakers. The results in both places were mixed: purism was a
prevalent ideology in the television channel, only inasmuch as it did not “clash with other
institutional priorities, such as the need to inform and entertain” (Robert 2013:190); in
the investigated workplace borrowings were seen as appropriate to some extent and used
alongside ‘pure’, codified vocabulary, which the author attributes to intense language
contact and the functional differentiation of Welsh and English. Thus, as the author
claims, the ideology of purism was in competition with other priorities stemming from
everyday needs and interactions (2013: 242). Drawing from her own observations, Robert
also states that recent demographic and social developments in Wales might not only have
brought changes in speakers’ linguistic resources, but also “challenged the norms and
authority that had previously upheld the authority of the standard register” (2013: 255).
245
This, in the author’s opinion, may result in increasing tolerance towards variable language
practices, since purism and standard language ideology are bound to exercise much less
influence in vernacular practice outside of institutional settings. These observations are
relevant to the present thesis, which has demonstrated tensions between normative practices, in which purism plays an important role, and the practice of speakers reflected in
the investigated corpus.
Robert’s investigation on the issue of planned language concerned mainly speakers’ choices between loanwords and native Welsh terminology, including coinages. The
phenomenon of calquing has been paid much less attention in the literature (see 1.3.2).
Specifically with regard to calquing PVs, the issue of acceptability has been only briefly
mentioned in two recent theses by Woolridge (2011) and Hirata (2012).
Woolridge (2011), who investigated the use of the spell-checker Cysill to recognise English interference in Welsh texts, has conducted a study on a parallel corpus in
which two professional translators evaluated texts translated from English into Welsh.
Some PVs have been mentioned as a salient token of word-for-word unconscious translation of English idiomatic phrases and marked as incorrect (2011: 64). The author has
also shown that the process of translation is a multi-layered one and the emergence of
Welsh PVs might not necessarily be a result of direct calquing80.
Hirata (2012) is the only study found which has directly referred to the subject of
the acceptability of PVs. While investigating preposition stranding, the author checked
the acceptability of a small number of PVs among native speakers of Welsh. He concluded that the speakers were inclined to accept one-word equivalents rather than PVs
and this might be correlated with their exposure to literary language (2.3.6).
5.2.3. Ideologies of standardisation
This scattered evidence found in a small number of studies suggests that standardisationrelated ideologies play an important role in creating and maintaining the linguistic norm
For example, the PV diweddu i fyny ‘end up with’ appeared as a translation of eventually, which means
that the translator incapable of finding a Welsh equivalent, first “turned” eventually into alternative expression ‘end up with’ and then calqued it into Welsh (Woolridge 2011:65).
80
246
concerning English borrowings in Welsh, including verb-particle constructions. However, one should not expect consistency in speakers’ views on this subject as they are
influenced by conflicting ideologies. As described in Chapter 1, the ongoing process of
standardisation of Welsh is a very complex one, due to the multitude and diversity of
environments that create the linguistic norm. This is not an unusual situation for a minority language, where, as stated by Costa et al. (2017: 12), standards are “subject to negotiation, debate, contestation and appropriation by various types of social actors in very diverse circumstances.” In the case of Welsh, a pivotal role is played by universal
bilingualism and the resulting change of the traditional monolingual norm. The latter is
still embodied in prescriptive or purist ideologies permeating different channels of language transmission, such as education, despite the fact that on average speakers of Welsh
seemingly have less and less contact with the traditional standard literary language (see
1.2.4.1). On the other hand, popular media appear to emerge as the most powerful source
of the current standards and their inclusive policy of “normalising less perfect Welsh”
(see 1.2.7) can be expected to increase tolerance for the growing amount of interference
from English in Welsh. As noticed by Costa et al. (2017: 12), “desire for a pure, monoglossic norm may emerge in minoritised language communities as it has in national ones,
but it does not (or cannot) translate into the same monolingual outcome.” The emerging
Welsh standard, in particular the written standard, appears to be in the process of negotiation between maintaining purist ideologies and catering for needs of speakers who view
bilingualism as a natural thing (cf. Musk 2012) and who may have little confidence regarding their proficiency in Welsh.
As a result, phrasal verbs in Welsh may be considered a case of contact-induced
variation, where the calqued constructions exist alongside native items, a common phenomenon in the Welsh lexicon (1.3.1.3). Thomas states that calqued constructions are less
likely to be noticed as borrowings and in some cases can even be preferred as an alternative to a loanword; however, for some purists they might be seen as a violation of the
native grammar, restructuring the linguistic system on the level of word formation, phraseology, syntax or semantics; for this reason, Thomas sees them as a useful test of the
level of purism (Thomas 1991: 70-72).
The study presented in this chapter, although focusing on the perceived acceptability of verb-particle constructions themselves, aims to explore this linguistic feature
within the wider context of the ongoing language standardisation process and competing
247
norms pertaining to different domains of social life. The question is whether the monolingual norm prescribing against the use of some PVs is prevalent among language-aware
professionals or whether they are bound to accept loan elements as part of the standard.
As discussed in 1.5. the investigation is not designed as a full-fledged study of individual
attitudes but endeavours to identify some ideological stances which might influence
speakers’ lexical choices stemming from ideologies of standardisation presented in 1.2.
5.3. Study design
The choice of methodology of the study was dictated by several factors, related to the
complex nature of verb-particle constructions. Firstly, as previously mentioned, the main
focus of the study is on the status of the linguistic feature within a constructed linguistic
norm rather than social aspects of its usage. The chief problem in investigating phrasal
verbs in natural production – be it speech or writing – is that they are extremely difficult
to elicit from speakers, in contrast to, for example, concrete nouns, verbs or adjectives,
especially due to the existence of alternative equivalents. Therefore, a study measuring
actual linguistic behaviour of speakers with regard to the use of PVs in standard written
or spoken Welsh would require large samples of material obtained through participant
observation, which was unfeasible for the researcher. However, bearing in mind that the
focus of the study is on the linguistic norm, it was decided that a direct approach may
yield valuable results with regard to the degree of standardisation of PVs and proficient
speakers’ awareness of the linguistic norm. Some methodological difficulties regarding
measuring linguistic attitudes towards phrasal verbs have been already discussed in 1.5.
Especially in view of the fact that phrasal verbs are not easily identifiable objects in the
minds of the speakers, it was considered appropriate to investigate the acceptability of
PVs along a simpler, binary criterion of acceptable/inacceptable. This also allowed to
investigate a larger number of constructions within available time limits.
An attempt to identify the prevalent linguistic ideologies among the interview
groups was made based on the speakers’ explicit statements as well as data from questionnaires. This necessarily means that the results do not reflect speakers’ behaviour in
real life, but merely their declared lexical choices. The fact that Welsh speakers’ actual
behaviour might not match their grammaticality judgments was observed for example by
248
Parafita Couto et al. (2015) in their study on code-switching. However, what is of interest
to the thesis, is the awareness of the monolingual norm identified in previous chapters
and speakers’ conscious decisions rather than covert attitudes and beliefs. Therefore, the
applied methodology is believed to bring some valuable and valid data.
5.3.1. Aims of the study
The present study was intended to complement the analysis regarding the integration of
PVs in standard Welsh conducted in Chapters 3 and 4. The study has the following specific aims:
-
AIM 1: to measure the acceptability of selected PVs in semi-formal registers of
Welsh against alternative expressions, making use of the onomasiological approach described in 5.2.2.;
-
AIM 2: to measure the declared acceptability of selected PVs in semi-formal registers compared with the spoken informal;
-
AIM 3: to identify some possible motivations for accepting or rejecting phrasal
verbs and tie them to standardisation-related ideologies and other extralinguistic
factors.
5.3.2. Research hypotheses
Based on the findings presented in the previous analytical chapters of the thesis, the following research hypotheses were formulated for this study:
Hypothesis 1 (Aim 1)
According to the current linguistic norm, native lexical items synonymous to PVs will be
more accepted than PVs in general. However, with regard to individual PVs the degree
of acceptability will depend on their status in relation to contact with English:
249
-
native phrasal verbs, in particular PPs, semantic extensions of well-established
items and calques which fill a semantic gap will be the most accepted ones;
-
pleonastic items are expected to be in the middle of acceptability continuum;
-
replaceable direct calques will be generally avoided;
-
loanblends and dialectal variants of common PVs will be rejected in the semiformal register.
Hypothesis 2 (Aim 2)
Selected high-frequency PVs will be widely accepted in informal conversation and their
acceptability will diminish the more formal the register. Acceptability levels will be also
modulated by the type of PV in relation to English, as in Hypothesis 1.
Hypothesis 3 (Aim 3)
Professional speakers will be aware of the linguistic norm prescribing the use of PVs,
however, they will not perceive such a norm to be valid in speech. Speakers’ linguistic
choices might be tied to multiple, often conflicting ideologies.
5.3.3. The sample
In view of the aims outlined above, the study focused on linguistic norms in standard
varieties of Welsh. Due to the fact that these are associated primarily with semi-formal
registers and rooted in the written language, investigation of the acceptability of verbparticle constructions required finding participants who would be familiar with the written language. As shown by statistical data, this is not the case for all Welsh speakers
(1.1.1). Moreover, it was considered worthwhile searching for informants who, as well as
being the users of standard language, contribute to encoding the linguistic norm themselves. Such speakers’ judgements would be valuable markers of the norm and standardisation-related ideologies due to the fact that one may expect them to display high levels
of linguistic awareness and sensitivity to language contact.
250
Bearing in mind the current sociolinguistic situation in Wales, the growing number of L2 Welsh speakers and complex environments of acquiring or learning the language (1.1.2), it was considered important not to restrict the choice of participants according to a “native” or “new” speaker criterion. Therefore, the sampling of participants
involved selecting groups working through the medium of Welsh in fields such as education, art, academic research, politics and media. Such informants will be named here “professional speakers”. The term is adapted after McEwan-Fujita (2008) who applied it in
her study to describe middle-class, white-collar workers using Scottish Gaelic in their
professional environment. The linguistic practice in the two languages is, of course, very
different since, in contrast to speakers of Gaelic, professional Welsh speakers are more
likely to be first-language speakers and live in a Welsh-speaking environment, using the
language also outside the workplace. Notwithstanding this difference, using the term
“professional speaker” emphasises the fact that the participants of the study were chosen
on the basis of their expected high level of linguistic awareness and contact with standard
varieties of Welsh on an everyday basis.
Altogether, seven groups were recruited, with a total of 55 speakers. Each group
was assigned a letter code: A, B, C, G, L, O and Y, and members of each group were
numbered consecutively in the order they were interviewed. Research was carried out in
two periods: 32 interviews were conducted in July and August 2016 in Ceredigion, and
during the National Eisteddfod in Abergavenny, Monmouthshire; the remaining 22 interviews took place in March 2017 in Gwynedd and Denbighshire.
Group A (Athrawon ‘teachers’) consisted of 8 teachers employed in a secondary
school in Denbighshire. The school is a Welsh-medium one, where all subjects are taught
through the medium of Welsh. However, the students’ command of Welsh can vary due
to the fact that a significant proportion of them come from English-speaking families.
Group A was the most homogenous one in terms of linguistic background in that all the
teachers were first-language speakers of Welsh living in North Wales. This group was
selected with the purpose of examining issues related to standards in teaching in a bilingual educational environment.
Group B (Bedwyr) consisted of 8 staff members of Canolfan Bedwyr, a centre for
Welsh language services, research and technology located in Bangor. The centre is a key
Welsh institution responsible for standardising terminology and developing digital tools,
251
such as the spellchecker Cysill, dictionaries and mobile applications. It also offers translation services, language courses and engages in the Cymraeg Clir project, which encourages public institutions to use simple, natural language (see 1.2.4.2). Members of group B
who took part in the study were professional translators and/or teachers conducting language skill courses for improving written and spoken Welsh. For this reason the group
was expected to be highly sensitive to issues of language contact.
Group C (Cymdeithas) were 9 members of Cymdeithas yr Iaith, an activist group
campaigning for the rights of the Welsh language (see 1.1.1.). As well as engaging in
language activism, the participants in this group dealt with Welsh in their professional
lives, for example by teaching, creating public websites and engaging in creative and
journalistic writing. Group C was the most diversified with regard to the speakers’ place
of residence, due to the fact that members from different parts of Wales were interviewed
during the National Eisteddfod. Group C was also the most varied when it came to linguistic background as had the highest proportion of new speakers or speakers coming
from bilingual families. The purpose of selecting this group was to discuss issues related
to language policy and the official status of Welsh.
Group G (Golwg) were 11 journalists and staff members of Golwg, a company
which publishes the Welsh-language weekly Golwg, a magazine for learners Lingo
Newydd and children’s magazine Wcw a’i ffrindiau; it also runs an on-line news platform
Golwg360. The interviews were conducted at the company’s offices in Lampeter and
Caernarfon and at its stand at the National Eisteddfod. The purpose of selecting this group
was to discuss issues related to journalism and media language.
Group L (Llyfrgellwyr ‘librarians’) consisted of 10 librarians working in two libraries in Aberystwyth: the National Library (L1-L6) and Town Library (L7-L10). The
National Library sponsored by the Welsh government is the largest library in Wales and
a key centre for research and education, whose staff are expected to have a good command
of Welsh. The Town Library located in Alun R. Edwards Centre is a lending library financed by the county council and an important location for the local community life. At
the time the research was conducted, not all of the employees in the institution were able
to communicate in Welsh. However, in my own experience, a Welsh-speaking member
of the staff was always available. Nearly all participants of Group L were native speakers
of Welsh. Some them worked in the library part-time and engaged in other professional
activities, including academic research, translation and work at local government bodies.
252
The group was expected to give insight into bilingual community life in the area of Ceredigion, where Welsh-speakers constitute around half of the population.
Group O (Oedolion ‘adults’) consisted of 6 teachers working at Welsh for Adults
(Cymraeg i Oedolion) centre in Bangor, which offers Welsh courses at different levels of
proficiency in Gwynedd, Conwy and Anglesey. Most members of group O came from
North Wales and were native Welsh speakers. Due to the first-hand contact with learners
of Welsh and linguistic resources used for teaching, the group was expected to be able to
discuss matters related to bilingualism and the use of standard language in teaching.
Group Y (Ysgrifennwyr ‘writers’) were members of an informal creative writing
club in Aberystwyth. The club consists of about a dozen writers who produce prose and
poetry in the Welsh language and meet irregularly to present and discuss the outcomes of
their work. The group was expected to inform about their views on using standard language in fiction. Unfortunately, due to practical difficulties, only 3 members of the club
were able to take part in the interviews. Nevertheless, it was considered worthwhile to
include the elicited data in the study.
The groups were recruited in several ways, following the triangulation of data
approach for qualitative studies, i.e. the application of more than one sampling method
for data collection (Angouri 2010: 34). Group L was recruited using simple random sample i.e. approaching librarians directly at their workplace. Groups A, B and O were recruited via a group member or leader by contacting heads of the institutions who prepared
lists of staff willing to take part in the study on a given date. Finally, groups G, C and Y
were contacted via my private networks as a previous trainee at Golwg and a member of
Cymdeithas yr Iaith, whose members introduced me to the writing club. This allowed for
a relatively wide range of professional speakers, from language specialists to lay people
who are actively engaged in language maintenance.
5.3.4. Methods
The method chosen for examining speakers’ preferences and conscious lexical choices
was a judgement test in the form of a questionnaire. The questionnaire consisted of two
parts, A and B, corresponding with Aims 1 and 2 of the study (5.3.1.), respectively. In
253
addition, the participants also filled a personal questionnaire and were interviewed in order to investigate their linguistic habits and awareness and identify extralinguistic factors
which may influence their lexical choices (Aim 3).
5.3.4.1. The procedure
Prior to the study, each participant signed two copies of a written consent in which they
were informed that the interview would be recorded and that they were entitled to keep
their anonymity and withdraw from the research at any time (see Appendix B). The interviews were conducted in various environments: institutions, offices, private homes and
at the National Eisteddfod. As the intention was to interview the speakers in a possibly
monolingual environment, all the instructions in the questionnaires were in Welsh and
the researcher spoke only Welsh during the interviews. None of the speakers attempted
to switch to English during the conversation. A single study, including the interview and
filling the questionnaire, took approximately 20-30 minutes and consisted of four stages.
In the first stage, the participants were asked to fill a personal questionnaire (Appendix
B). The second part of the study was a one-to-one semi-structured interview (Appendix
C). Following the interview, the participants were given a two-part questionnaire evaluating the acceptability of selected PVs (Appendix B). The procedure was concluded with
a follow-up interview, in which participants were asked to share their thoughts and comments on the questionnaire (Appendix C).
The whole procedure was recorded. The data from the questionnaires were transported into Excel sheets and analysed. The interviews preceding and following the filling
of questionnaire were transcribed using the VOICE Transcription Conventions (VOICE
Project 2007), while oral comments to questionnaires were noted down and inserted in
the datasheets. The interview transcriptions were analysed and organised around themes,
following the topic-oriented approach to analysing metalinguistic discourse (cf. Rodgers
2017: 84).
254
Questionnaire A (Aim 1)
In order to measure the acceptability of selected PVs in semi-formal registers of Welsh
against alternative expressions (Aim 1), Part A of the questionnaire approached the standardness of PVs from an onomasiological perspective, checking the speakers’ preferences
for PV or non-PV items to express a given concept. The task consisted of 12 gapped sentences extracted from printed and online press articles or websites of public institutions
and organisations written in semi-formal registers. Every sentence originally contained a
PV. Some sentences were slightly modified by abbreviation or replacing proper names
with general nouns. The informants were given between 2 and 4 expressions – at least
one of them being a PV – to fill a gap in a sentence and were asked to tick all the expressions they would find appropriate in the context. It was explained that ‘appropriate’ meant
what they would expect to see in press or on official websites. The participants were
informed that there were no ‘correct’ answers and that they could choose any number of
phrases or none at all. Since the task was to establish whether the selected PVs are seen
as part of the standard language, speakers were given a binary choice between appropriate/inappropriate, yet were not restricted to choosing a single item to fill the gap. They
were encouraged to propose alternative expressions or share any comments either by writing them in a “Comment” box below each sentence or orally as the whole procedure was
recorded.
The aim of Part A was to draw individual onomasiological profiles for each concept expressed by a PV (cf. Speelman et al. 2003: 319) and determine the acceptability
and stylistic markedness of a selected item against semantically equivalent expressions.
It should be emphasised that one cannot consider the selected expressions to be true synonyms as such synonymy hardly ever exists in natural language. Following Zenner et al.’s
methodology (2012: 760) in their study of Anglicisms in Dutch, the questionnaire contained near-synonyms which are maximally equivalent with a selected PV on four levels:
denotational, stylistic, expressive and structural; the authors made use of Edmonds and
Hirst’s (2002) theoretical model of granularity of meaning according to which “at the
conceptual, coarse-grained level, near-synonyms have the same meaning, and are hence
clustered under one shared concept.” The non-PV items in the present study were chosen
255
on the basis of equivalents of the selected PVs proposed by dictionaries, primarily Geiriadur yr Academi. The sentences in the questionnaire contained the following PVs and
their equivalents:
Sentence 1 – dod yn ôl ‘come back’ against dychwelyd ‘return’. As discussed in section 2.4.5 although the present study generally focuses on idiomatic constructions, the
transparent PV dod yn ôl was considered worth investigating as it is given as an example of “replaceable” PV by P. W. Thomas (see 2.3.3);
Sentences 2, 6 and 10 – tyfu i fyny ‘grow up’ against tyfu ‘grow’; helpu allan and
dialectal helpu mas ‘help out’ against cynorthwyo ‘help’ and helpu ‘help’; eistedd
i lawr ‘sit down’ against eistedd ‘sit’; these PVs are common examples of pleonastic
calques;
Sentence 4, 7 and 11 – idiomatic PVs directly translatable into English which appear
to be stylistically marked – gweithio allan and dialectal gweithio mas ‘work out’
against darganfod ‘discover’; pigo i fyny and dialectal pigo lan ‘pick up’ against
dysgu ‘learn’ and cael crap ar ‘get a grasp on’; troi i ffwrdd ‘turn off’, troi bant ‘turn
off’, troi off ‘turn off’ against diffodd ‘turn off’;
Sentence 5 – edrych ymlaen at ‘look forward to’ against non-standard variant edrych
ymlaen i and disgwyl yn eiddgar am ‘expect ardently’; this sentence examined the
case of a very frequent PV which does not have an obvious one-word equivalent;
Sentence 12 – eistedd yn ôl ‘sit back’, ‘make no effort to do something’, against sefyll
o’r neilltu ‘stand aside’; this sentence examined the case of a relatively rare PV, which
does not have an obvious one-word equivalent and whose English equivalent is also
less frequently used;
Sentence 3 – mynd ymlaen ‘go on’ against digwydd ‘happen’, a case of semantic extension of a well-established item (see 4.5.2);
Sentence 9 – bwrw ymlaen ‘continue’, ‘carry on’ against cario ymlaen ‘carry on’ and
parhau ‘continue’; this sentence examined the acceptability of competing native
and borrowed PVs (see 4.5.2);
Sentence 8 – dod ar draws ‘come across’ and taro ar ‘hit on, come across’; this sentence checked the acceptability of two frequent and well-established idiomatic PPs
one of which is directly translatable into English (see 4.5.5). Both items are proposed
256
by Geiriadur yr Academi as equivalents of come across (something), with taro ar
given as the first option.
As the informants were allowed to suggest their own words or phrases to express a given
concept, data from Part A of the questionnaire were analysed in the form of case studies
for each of the selected PVs (5.4.2.).
5.3.4.2. Questionnaire part B (Aim 2)
In order to measure the declared acceptability of selected PVs in semi-formal registers
compared with the spoken informal, in Part B of the questionnaire the speakers were
asked to judge the appropriateness of using selected PVs in four different situations. The
participants were given four sets of sentences. The sets were introduced by putting them
in the context of four imaginary situations, representing different registers and levels of
formality:
Situation 1 – a conversation with a friend on everyday matters;
Situation 2 – a public meeting with a politician discussing local issues;
Situation 3 – writing a popular book about environmental issues;
Situation 4 – preparing a leaflet on local health service.
Situation 1 was intended to represent an informal, spoken mode, while Situation
2 represented careful, formal speech. Situation 3 was to direct participants’ attention towards written semi-formal variety by emphasising that the book they are writing is targeted towards larger audience yet touches on serious, sometimes quite technical subjects.
Finally, Situation 4 pertained to the official variety of Welsh, which could be expected to
be more formal than popular literature yet understandable to the public (see 1.2.4.2).
Each set contained six sentences with the same six PVs marked in bold font (Appendix B). The six selected PVs represented the following types of verb-particle constructions:
257
edrych ar ôl ‘look after’ – a commonly used prepositional verb, discouraged by some
dictionaries and pedagogical materials (see 3.3.2.13);
ffeindio allan (dial. ffeindio mas) ‘find out’ – a frequently occurring loanblend excluded from dictionaries and prescribed against in many teaching materials (see
4.5.4). It was also found to be stylistically marked in the corpus (see 3.3.2.13);
rhedeg allan o (dial. rhedeg mas o) ‘run out of’ – a phrasal prepositional verb probably directly calqued from English (see 3.3.2.1) and generally excluded from Welsh
dictionaries, apart from GPC which tags it as “spoken”;
torri i lawr ‘break down’ – a common pleonastic calque, which is included in most
Welsh dictionaries but marked as informal (see 4.5.8);
troi i fyny (dial. troi lan) ‘turn up, come’– a direct calque from English, considered to
be rather informal if not dialectal (cf. Cownie 2000);
troi ymlaen ‘turn on’, ‘switch on’ – a recent calque from English which does not have
an obvious Welsh equivalent and hence fills a lexical gap; it is not commonly found
in dictionaries, it was included, however, in a teaching book by Gruffudd (2000) (see
4.2.1.).
The sentences in the questionnaire were modelled on authentic sentences containing
the selected PVs found in printed texts or online, including tweets and Internet forums.
The participants were asked to imagine themselves in the situation described before each
set of sentences and tick a box next to the sentence if they would use a PV in the context.
If a PV had a dialectal variant, both were given and the participants could choose the
option they would use by underlining it (see Appendix B). The participants were not
obliged to propose alternative expression if they would not use a PV; however, many of
them did so of their own accord.
It should be noted that several of the selected PVs were polysemous and in two cases
– torri i lawr and troi i fyny – the PVs were used in two different senses. In Situations 1,
2 and 4 they were used in their most frequent senses: ‘stop working’ for torri i lawr and
‘arrive, appear’ for troi i fyny, while in Situation 3 they were used in more technical senses
of ‘separating into parts’ and ‘increase power’ respectively. Similarly, rhedeg allan o was
used in an idiomatic phrase ‘run out of steam’, ‘losing energy’ in Situation 2, while in the
other three situations it referred to lacking something material (battery, oil, medicines).
This was done at the cost of quantitative reliability of the data in order to ensure the
258
naturalness of a PV in a given context and to investigate whether a need to use the item
in a sense which is more technical will be reflected in marked differences between the
acceptability of these PVs across registers due to the fact that the speakers will find it
more difficult to replace the PV by another item.
5.3.4.3. Personal questionnaire and interviews (Aim 3)
Other than filling the judgement test questionnaires, the speakers filled a personal survey
and took part in a one-to-one semi-structured interview. This was done in order to gather
information about the speakers, considering the following aspects:
-
Demographical factors – sex, age, education, place of residence, the age of acquiring/learning Welsh;
-
Use of Welsh in various spheres of everyday life;
-
Use of written Welsh;
-
Exposure to Welsh-language literature and media;
-
Confidence in using Welsh vocabulary;
-
Use of lexicographic resources;
-
Use of borrowings in speech and writing.
This data were considered useful for investigating speakers’ linguistic habits and awareness and identifying factors which may influence their linguistic choices in order to list
possible motivations for accepting or rejecting phrasal verbs and tie them to standardisation-related ideologies (Aim 3). The interviews were semi-structured in that they included
a number of regular questions asked to every participant as well as questions adapted to
a given group and individual speakers. The regular questions were:
a) Are you aware of using English words or English borrowings when you speak Welsh?
b) Do you speak Welsh in formal/official situations? Do you try to avoid borrowings in
such situations?
c) Do you write in Welsh? If so, how often?
d) Do you try to avoid borrowings in writing?
259
e) When you write, does it happen that you cannot think of a Welsh word you need? Do
you search for it then? If so, where?
f) Have you done the following in the last month: 1) Read a book in Welsh, 2) Read a
Welsh-language newspaper or magazine, 3) Listen to a Welsh-language radio 4)
Watch Welsh-language television, 5) Use Welsh-language software (Microsoft Office, Cysill, Cysgeir), 6) Watch an online video in Welsh (YouTube, S4C Clic, BBC
iPlayer, 7) Used a dictionary of Welsh – if so, which one?
The original questions in Welsh are shown in Appendix C. Additional questions were
asked in the course of the conversation depending on the profession of the participant; for
instance, teachers were asked about the use of borrowings in classroom, journalists about
the influence of English on their writing, writers about the use of borrowed words in
prose.
In a follow-up interview the participants were asked to share their thoughts and
comments about the questionnaire. If time allowed, they were also asked about their views
on standard language and the influence of English on Welsh. Questions in the interview
were adapted to the group and the speaker (Appendix C). Unfortunately, due to time constraints, it was not always feasible to cover all the themes with each of the participants.
The loose structure of interviews was aimed to encourage the participants to freely
express their beliefs in order to identify recurring themes in the discourse about phrasal
verbs and, more broadly speaking, issues of borrowing, calquing and the influence of
English on Welsh. These results, organised around themes are presented in section 5.4.4.
5.3.5. Qualitative and quantitative aspects of the study
While the study combines qualitative and quantitative aspects, its nature is primarily qualitative. The results of Part A of the questionnaire are presented in the form of percentages,
however the analysis examines each PV qualitatively, taking the onomasiological perspective. This approach was adopted due to the complex nature of the constructions described in previous chapters and the fact that the participants could propose their own
260
alternative expressions, which makes the number of native equivalents of PVs non-defined. Therefore, the analysis of the results is oriented towards the description of individual constructions, highlighting the differences in their acceptability.
Part B of the questionnaire relies more on quantitative data as the participants
were offered a binary choice of accepting/not accepting the item in a given register.
Therefore an attempt was made to examine several extralinguistic variables derived from
the collected data on the speakers against the results using statistical analysis. It should
be stressed that these results should not be generalised as the selected group is definitely
not representative of the whole population of Welsh speakers, but was aimed to represent
language-aware influentials.
Generally speaking, quantitative data from the study should be treated with caution since the issues studied are bound to be related to speakers’ attitudes and ideologies
(see 1.5.). Bearing this in mind, the data from the questionnaire will be set against the
data retrieved from the interviews regarding the speakers’ linguistic awareness, perceived
competence, use of borrowings and opinions on various aspects of language contact and
standardisation.
5.4. Results
5.4.1. The profile of the professional speakers
The data from the survey and interviews allow to draw a profile of the speakers, providing
evidence for their status as professional users of Welsh. Demographical information regarding the linguistic background of the sample and some possible extralinguistic variables to be examined will be now described.
5.4.1.1. Extralinguistic data
Sex and age
261
A total of 55 professional speakers of Welsh took part in the study, 19 men (35%) and 36
women (65%). The imbalance between sexes was caused by the fact that the two teacher
groups, A and O were strongly feminised (only one man in each) and the Y group (writers) consisted of females only.
With regard to age, the questionnaire assigned subjects to one of six age groups:
18-24, 25-34, 35-44, 45-54, 55-64 and 65+. Speaker L9, a part-time library worker just
under the age of 18 was included in the first group. The age distribution of the speakers
is shown in Figure 5.
Sex and age were considered as variables in the statistical analysis of the results
of Part B. Correlation between age and the acceptability of PVs had been suggested by
Hirata (2012).
16
25%
No. of speakers
14
22%
12
10
18%
15%
13%
8
6
7%
4
2
0
18-24
25-34
35-44
45-54
55-64
65 +
Age in years
Figure 5. Age of the speakers.
Place of residence
The geographical distribution of speakers is presented in Figure 6. Due to the location of
the institutions where the interviews were conducted, the majority of speakers lived in
North Wales (25 speakers – 45%) and Mid Wales (22 speakers – 40%). Speakers living
in South Wales constituted 15% of the informants (8 speakers). With regard to counties,
the highest number of participants lived in Ceredigion (19), Gwynedd (13) and Carmarthenshire (6) (had moved between North and South Wales, 9 speakers had moved within
different regions of North Wales, 1 speaker had moved within different regions of Mid
Wales, 1 speaker had moved within different regions of South Wales, and 2 speakers had
spent their childhood in England.
262
The above data show that the geographical distribution of speakers was uneven
and their mobility made it difficult to assign each participant to a particular area of Wales.
For this reason, geographical factors were not considered a variable in the analysis.
Table 20).
The speakers also indicated the area in which they lived at the time they had acquired or learned Welsh. More than half of the speakers, 27, learned Welsh in the same
area as their current place of residence, 7 speakers had moved between Mid Wales and
North Wales, 4 speakers had moved between Mid Wales and South Wales, 4 speakers
Figure 6. Geographical distribution of the participants according to their place of residence.
263
had moved between North and South Wales, 9 speakers had moved within different regions of North Wales, 1 speaker had moved within different regions of Mid Wales, 1
speaker had moved within different regions of South Wales, and 2 speakers had spent
their childhood in England.
The above data show that the geographical distribution of speakers was uneven
and their mobility made it difficult to assign each participant to a particular area of Wales.
For this reason, geographical factors were not considered a variable in the analysis.
Table 20. Geographical distribution of speakers according to their place of residence.
County
No. of speakers
Ceredigion
19
Gwynedd
13
Carmarthenshire
6
Denbighshire
5
Anglesey
4
Conwy
3
Powys
3
TOTAL
55
Linguistic background
The personal questionnaire asked about the age of acquiring or learning Welsh and the
researcher asked a follow-up question about the first language of the speaker’s parents.
The majority of participants, 85% (47) acquired Welsh from their parents at home, with
41 having two L1 Welsh-speaking parents, 1 having two L2 Welsh-speaking parents,
3 having one L1 parent and 2 having one L2 parent. One speaker learned Welsh as a child
older than 4 years, coming from an English-speaking family, but acquiring Welsh in the
neighbourhood. Two informants learned Welsh as teenagers and 4 as adults. One speaker
said she learned some Welsh as a child but due to a period spent abroad, she only mastered
it as an adult. This shows that within the studied sample, most professional users of Welsh
were L1 speakers who acquired Welsh at home. For this reason linguistic background
was not considered a variable in the statistical analysis.
Education
As regards the level of education, the vast majority of speakers in the sample, 44 (80%),
attended higher education, with 11 speakers holding a master’s degree or higher. Three
264
speakers (5%) had attended upper secondary and 5 (9%) lower secondary education.
Groups A, B, O and Y consisted only of speakers with higher education. The group with
the highest average level of education was Y, as all three speakers held master’s degree
or higher. The group with the lowest average level of education was L – it consisted of
4 speakers with lower secondary and 6 speakers with higher education (3 of them with
master’s degree or higher).
Whilst the majority of the sample had attended higher education, it was considered
worthwhile to consider university education or lack of it as a variable when analysing the
results of Part B since education is one of the major channels of transmission of the linguistic norm (1.2.4.1.).
Professional status
To confirm their status as professional speakers the participants were asked about their
professional status. The vast majority of the participants (51), including all speakers in
Groups A, B, G and O, were full or part-time employed. In Group Y (writers) one participant was retired and in group C (Cymdeithas yr Iaith) one participant was retired, one
was a full-time student and one was temporarily unemployed.
5.4.1.2. Use of Welsh in daily life
Regarding the use of Welsh on a daily basis, the participants were asked to determine the
amount of Welsh they spoke in four spheres of life: at home, at work, in the neighbourhood and with close friends, choosing from five options: “only Welsh”, “mostly Welsh”,
“Welsh and English equally”, “mostly English” and “only English”. The answers in each
component were then transposed into a 0-4 scale so that each speakers could score a maximum of 16 points. Each speaker was than assigned to one of five categories according to
the scale presented in Table 21.
Table 21. Speakers’ use of Welsh and English in daily life.
Use of Welsh
0-2
(almost) only English
No. of
speakers
1
3-6
mostly English
2
% of
speakers
2%
4%
265
7-10
13
24%
11-14
Welsh and English
equally
mostly Welsh
29
53%
15-16
(almost) only Welsh
10
18%
Welsh was the main language used in everyday life for the majority of speakers (71%),
while only 6% reported English to be the main language they used. It is also worth noting
that the scores for the component “in the neighbourhood” were markedly lower than for
the other three components, which probably illustrates the diminishing use of Welsh as
a community language.
Since the number of speakers who used mostly or only English in their daily life
was marginal (3 speakers), this variable could not be considered for a statistical test to be
valid.
5.4.1.3. Reading, writing and use of Welsh-language media
In order to examine the participants’’ exposure to standard written Welsh, they were
asked about the frequency of reading and writing in Welsh in their everyday life. Questions on readership concerned reading books and newspapers and/or magazines. Reading
books in Welsh was less common than reading the press: 38 (69%) speakers said they had
read a book and 49 speakers (87%) said they had read Welsh press in the last month. Only
5 speakers (9%) had not read either books or the press in Welsh.
With regard to writing, 40 (73%) speakers said they used written Welsh every day
and 9 (16%) said they did so often. The majority of speakers said they used written Welsh
at work and in personal correspondence. Five speakers (9%) declared writing in
Welsh only occasionally, mostly at school or work. One speaker said that, being unfamiliar with formal Welsh, he avoided writing in the language altogether, as it was difficult
and time-consuming for him (C5). Two speakers declared that they felt they did not write
enough in Welsh (L8, C9). Three participants in Group B were professional translators,
and another three speakers mentioned that their jobs involved plenty of translation between Welsh and English (A2, L1, G2). Five speakers wrote creatively (A1, G1 – poetry
and group Y – prose). All in all, the exposure to written language was generally high,
266
confirming the professional status of the speakers and their expected familiarity with
standards in writing.
Another set of questions in the interview concerned the use of Welsh-language
broadcast media, BBC Radio Cymru and the television channel S4C, in the previous
month. The participants were also asked about watching online video clips on websites
such as Youtube and using S4C video on demand service, Clic. This was considered important due to the fact that media appear to be one of the major sources of linguistic
standards as discussed in 1.2.
Radio and television were very popular among the participants, with 91% listening to Radio Cymru and 88% watching S4C; video clips were also popular, with 79%
watching them regularly (participants who said they did not use a given medium at all
were not taken into account). By adding up the number of positive answers to the questions on readership and use of Welsh-language media, it was established that 85% of
speakers declared they used all or most of the Welsh media and read books (score 4 or 5
on a 1-5 scale). Detailed results are presented in Table 22. Similarly as with language use,
the homogeneity of the group made this variable invalid for statistical analysis.
Table 22. Declared readership and the use of Welsh media among the participants.
Score
Readership and the use of Welsh-language media
No. of Speakers
% of Speakers
1
no contact with Welsh-language media/books
1
2%
2
little contact with Welsh-language media/books
1
2%
3
some contact with Welsh-language media/books
6
11%
4
using most Welsh-language media and reading books
20
36%
5
using all Welsh-language media and reading books
27
49%
5.4.1.4. Speakers’ perceived competence and confidence
It is worth noting that although the informants were not asked directly about their perceived competence in Welsh, one fifth of the speakers expressed a lack of confidence in
their command of Welsh in the course of the interview of their own accord, in some cases
tracing this to the extensive influence of English in their area.
267
As the majority of participants in the study were native speakers living their lives
entirely or mostly through the medium of Welsh, they generally showed confidence about
their command of the language. Five speakers mentioned having a grade in Welsh and
four stated that their Welsh was quite pure/correct. On the other hand, 8 speakers suggested in the interview that their Welsh was not quite correct; half of this group were new
speakers of Welsh. Only one speaker, however, said explicitly that she did not feel confident in speaking Welsh (C3). This participant had learned Welsh recently and was also
the only speaker using almost only English in her everyday life. The speaker used “Wenglish” as a derogatory term for her language which, as she believed, had too many English
words in it.
Two other speakers used the term “Wenglish” in the interviews to describe one of
the languages they speak in everyday life. For these participants Wenglish was not a derogatory term, however, but a name for a mixture of Welsh and English. L9, a high school
student, said he spoke Wenglish with his friends at school, where it is common for young
people to mix the two languages and code-switch (for a different understanding of Wenglish see 1.1.2). The speaker said he usually did not think about which language he was
speaking at any given moment:
L9: yn enwedig yn yr ysgol (.) wi’n siarad lot o WENGLISH mae nhw’n galw e (...) [so
cymysgiad] o Gymraeg a Saesneg gyda ffrindiau (...) [achos] weithiau mae (.) geiriau sy’n
esceipo (...) a does dim (.) a mae rhaid defnyddio gair Saesneg wedyn (...)
R: [pawb yn siarad cymysgedd]
L9: [cymysgedd]
R: ie
L9: ond dw i ddim yn sylweddoli bodd i’n wneud e fi’n meddwl (...) achos dw i’n siarad
yn y ddau (.) (fi)’n jyst yn (.) yn newid rhwng y ddau yn hawdd a does neb yn sylweddoli
rili
‘L9: Especially at school I speak a lot of WENGLISH they call it (...) so a mixture of Welsh
and English with friends (...) because sometimes there are words which escape (...) a there
is no... you have to use an English word then (...)
R: Everyone speaking a mixture.
L9: Mixture.
R: Yeah.
L9: But I don’t realise I’m doing it I think (...) because I speak in the two (.) I just change
between the two easily and nobody realises really.’81
81
All translations of quotations from interviews in this chapter are given in the form of a narrative, while
original transcripts follow the VOICE conventions. Bracket ellipses indicate the omitted parts of the interview, usually irrelevant noises or meaningless interruptions of the researcher; capital letters denote words
pronounced with emphasis; in translations italics are used for Welsh phrases which need not be translated
or indicate that the participant used an English word; pauses marked by (1) are replaced in translations with
ellipsis.
268
Speaker O2 said she used this kind of language with her family who “throw Wenglish in
all the time” (taflu Wenglish i mewn trwy’r amser). The participant referred to ideologies
of purism, saying that mixing languages is something frowned upon by some people, but
it did not matter to her.
Some speakers indicated that their Welsh was not quite correct due to the fact they
lived in an area heavily influenced by English. Two of these speakers lived in Ceredigion
(L2, Y2) and three in North Wales (A1, B6, C9). L2, an elderly speaker coming originally
from Merionethshire but living in Ceredigion, believed that people in the North use “perfect” Welsh, devoid of any English borrowings.
An additional question concerning speakers’ competence was whether they experienced a situation when they could not think of a Welsh word they needed. Answers to
this question are presented in Figure 7. The majority of speakers, 39 (75%), said they
sometimes had trouble finding a Welsh word, while 13 speakers (24%) said they did not
experience such problems. In the latter group, eight speakers added they sometimes found
it difficult to think of an English word. The same was admitted by eight speakers in the
first group.
not sure
2%
I sometimes
find it difficult
to think of a
Welsh word
71%
I have no
difficulty to think
of Welsh words
24%
I often find it
difficult to think
of a Welsh word
3%
Figure 7. Answers to the question “Does it happen that you cannot think of a Welsh word
you need?”
5.4.1.1. Use of dictionaries and software
269
Following the latter question, the informants were asked what they did in a situation when
they were unable to find a Welsh word. In most cases the speakers thought the question
referred to writing and declared that they used a dictionary. All the participants were
asked whether they used dictionaries of Welsh and computer tools or software which can
be of help in writing in Welsh in the last month. This information was considered potentially relevant for investigating participants’ knowledge of the standards derived from
lexicographic resources.
Over 90% of speakers declared they used dictionaries of Welsh on a regular basis.
The majority of speakers (67%) said they used dictionaries when they could not remember a Welsh word. Many speakers noted that this happened not only in writing, but also
while speaking, as mobile applications make it possible to look up a word at any time.
Less than 10% of speakers did not use dictionaries or used them very rarely, although
some of those speakers admitted ‘googling’ words occasionally. Seven persons declared
they used dictionaries only to check technical terms – most of these informants were professional translators. Two participants (both aged over 65) used dictionaries for other purposes: the pleasure of discovering Welsh vocabulary or for solving crosswords.
In a follow-up question, the participants were asked to name the dictionaries they
used. The results are presented in Table 23. The most popular dictionary by far was Geiriadur yr Academi, in most cases its on-line version. The two other most popular dictionaries were online Geiriadur Prifysgol Cymru (15), Cysgair (13) and mobile application
apGeiriaduron (9) based on the latter. The most popular printed dictionary was Geiriadur
Mawr, used mostly for checking spelling. As regards other tools, the spell-checker Cysill
was mentioned by over a half of the speakers. Four speakers also mentioned using Google
Translate.
Table 23. Dictionaries of Welsh and other tools used by the informants.
General dictionaries of Welsh
No. of speakers
Geiriadur y Academi
28
Geiriadur Prifysgol Cymru
15
Cysgair
13
apGeiriaduron
9
Geiriadur Mawr
8
Trinity Saint David
6
Geiriadur Newydd/Cyfoes
3
Geiriadur Bangor
2
270
Gweiadur
1
Geiriadur Gomer
1
Collins Spurrell
1
Geiriadur Gomer i'r Ifanc
1
Other tools
Cysill
GoogleTranslate
29
4
5.4.1.2. Use of borrowings
It is rather striking that all speakers gave an immediate positive answer to the first question of the interview “Do you use English borrowings in your Welsh?” Forty-five speakers (81%) declared using some borrowings, while the remaining 10 said they used many
borrowings. Sometimes speakers gave specific examples of loanwords or specified the
type of borrowings they used: for instance well-established vocabulary items, phrases,
proverbs, interjections, borrowed verbs with -io ending. It should be noted at this point
that some of the participants occasionally seemed to refer to instances of code-switching
rather than using loanwords. As described in 1.3.1, the distinction between the two phenomena is a difficult one to establish. Therefore it should be born in mind that some of
the speakers’ statements might refer to situations of interference in general rather than
borrowings specifically.
The second question concerned using borrowings in formal context. Ten speakers
declared they did not use Welsh outside informal conversations. Only one speaker said
that he would not avoid borrowings in formal situations, as he believes that one should
speak naturally in every situation and his Welsh is “rather more informal”. The remaining
44 participants (80%) said they avoided using borrowings from English in formal situations. Examples of such situations given by the speakers were: at work (library, meetings
with colleagues in Canolfan Bedwyr, teacher’s meetings, classroom, trainings, meetings
of the local council), at school or university, during lectures, presentations, job interviews
and when appearing in the media (radio, television).
271
In the third question, the speakers were asked whether they avoided borrowings
in writing. Except for two participants who said they did not write much in Welsh, all of
the speakers answered positively. While 41 informants (75%) said they avoided borrowings in writing, 5 speakers specified that they avoided borrowings only in formal writing,
while 7 avoided only some borrowings. In answering this question speakers were generally much quicker and more confident in comparison with the previous question about
formal speech. Many indicated that monitoring one’s Welsh is much easier in the written
mode as there is more time to think, correct and polish. Moreover, one can rely on reference materials such as dictionaries.
G5: yn bendant. osgoi benthyciade wrth ysgrifennu bob amser a dweud y gwir. y:m yn fwy
na na mewn iaith lafar achos (.) chi’n gallu rheoli beth dych chi’n ysgrifennu (.) chi’n gallu
(.) mireinio (.) gwneud yn siŵr ei bod hi’n iaith lan o gymharu falle wrth sgwrsio.
‘Definitely. Avoid borrowings while writing every time to tell the truth. More than in the
spoken language because you can control what you write you can polish make sure it is a
clean language in comparison with maybe what is said.’
C6: ‘swn i’n byth yn sgwennu fel ‘swn i’n siarad ym ‘swn i ddim yn dweud SO (.) ‘swn i
ddim yn (2) ie (.) ‘sa’r ffordd dw i’n sgwennu yn hollol wahanol i’r ffordd dw i’n siarad
(...) os dw i ddim yn cofio gair faswn i’n sbïo fo i fyny a (.) ‘swn i ddim y dweud so: ac
petha sy’n hawdd i slipio fewn yn Saesneg.
‘I’d never write the way I speak, I wouldn’t say SO, I wouldn’t... yeah, the way I write
would be totally different from how I speak (...) if I don’t remember a word I would look it
up, I would say so and things which are easy to slip in in English.’
A couple of speakers mentioned using informal Welsh in the social media. For instance,
speaker G3 emphasised that she used more English words on social networking sites; in
contrast, B7 prided herself on avoiding loanwords in writing every time:
G3: os ydw i’n sgwennu’n ffurfiol (.) mae e’n bur. ond os ydw i’n sgwennu ar gyfryngau
cymdeithasol (1) <smile> (...) siwr bod ti’n clywed hyn (.) t’mod os wyt ti’n neud rhywbeth
facebook neu (.) y: rywbeth bach mwy ANffurfiol (.) mae’na fwy o eiriau Saesneg (.)
‘If I write formally, it’s pure. But if write on social media… (smile). Surely you’ve heard
it, you know if you do something, Facebook or something more INformal, there are more
English words.’
B7: bydda i’n edrych ar Facebook weithiau bydda i’n gwa gweld rhai pobol mae (.) mae’n
boitsh o Saesneg a Chymraeg a fydda i (.) ceisio sgwennu yn gywir (.) ar Facebook. mae
rhywun yn gallu sgwennu yn anffurfiol mewn cywair anffurfiol (.) ond defnyddio geiriau
Cymraeg.
272
‘I sometimes look on Facebook and I see some people who... it’s a hotchpotch of English
and Welsh and I try to write correctly on Facebook. One can write informally in the informal register but use Welsh words.’
Borrowings in speech
Borrowings in formal speech
2%
18%
avoiding
borrowings in
formal Welsh
not using
formal Welsh
18%
using some
borrowings
using many
borrowings
82%
80%
using
borrowings in
formal Welsh
Borrowings in writing
4%
avoiding borrowings in writing
9%
avoiding some borrowings in
writing
13%
74%
avoiding borrowings only in
formal writing
speaker does not write much in
Welsh
Figure 8. Declared use of borrowings in speech and writing.
G1, a young poet, made an interesting comment about using borrowings in poetry. Although he avoids them in traditional strict-metre poetry, they can occasionally embellish
the verse with a dialectal flavour:
G1:
(...) dw i’n sgwennu yn bennaf trwy gyfrwng y gynghanedd (.) y:m felly (.) bydden
ni’n defnyddio iaith eitha ffurfiol a cheisio osgoi y: geiriau (.) benthyciadau o Saesneg a
geiriau tafodieithol (.) ond mae’n dibynnu ar y cyd-destun (.) byddwn i’n ambell waith fel
sgwennu cerddi (.) sy’n cynnwys tipyn o dafodiaith oherwydd mae rai stamp unigryw arno
fe (...)
‘I write mostly through the medium of cynghanedd, so we would use quite formal language
and try to avoid words… borrowings from English and dialectal words, but it depends on
the context, sometimes I would, like, write poems which contain a bit of dialect because it
has some unique stamp (…)’
273
The answers to the questions regarding the use of borrowings are summarised in Figure
8. On the whole, speakers in all groups showed a high awareness of differences between
various registers of Welsh. The perceived diglossia between spoken and written language
was evident in that all speakers admitted using English vocabulary in speech and at the
same time avoiding it in writing.
5.4.1.3. Summary of data
The data from the personal questionnaire and interviews provide some interesting insights
into the linguistic background, practices and awareness of the professional speakers of
Welsh who took part in the study. Most informants were native speakers with higher education, who declared they used Welsh on a daily basis both at work and privately, but
were less likely to use it in the neighbourhood. The majority of speakers declared that
they used Welsh in formal as well as informal situations. What is more, the speakers
declared being highly exposed to the standard by regular reading and contact with the
Welsh-language media.
Despite that, one fifth of the speakers felt the need to share their lack of confidence
regarding their command of Welsh. Except for a few cases of new speakers, the majority
of these participants pointed to language contact as a motivation of their perceived lack
of competence. Most of the speakers stated that they experience difficulties in finding
Welsh vocabulary and a high percentage of the informants stated that they regularly use
dictionaries for help. Other than illustrating the influence of English on the language of
professional speakers, this demonstrates their linguistic awareness in that many of them
declared striving to look for Welsh words rather than code-switch. Interestingly, a meaningful percentage of speakers mentioned Geiriadur yr Academi as the major authority
when it comes to Welsh lexicon, which confirms the popularity and importance of this
dictionary.
274
5.4.2. Questionnaire Part A
This section presents the results of Part A of the questionnaire, in which the participants
were asked to indicate the expressions they found appropriate in order to fill a gapped
sentence. The percentages of positive answers for each PV or alternative expression are
summarised in Table 24.
Table 24. Percentage of speakers accepting the selected lexical items in Part A of the
questionnaire.
dod yn ôl
% of positive
answers
40%
Alternative
item
dychwelyd
% of positive
answers
89%
grow up
tyfu i fyny
24%
tyfu
84%
go on, happen
mynd ymlaen
15%
digwydd
96%
gweithio allan
6%
darganfod
98%
gweithio mas
4%
edrych ymlaen at
62%
disgwyl yn
eiddgar
73%
edrych ymlaen i
4%
helpu allan
7%
cynorthwyo
69%
helpu mas
10%
helpu
53%
pigo lan
9%
dysgu
95%
pigo i fyny
15%
cael crap ar
11%
dod ar draws
76%
taro ar
44%
bwrw ymlaen
62%
parhau
44%
cario ymlaen
20%
eistedd i lawr
44%
eistedd
64%
troi i ffwrdd
15%
diffodd
93%
troi bant
7%
troi off
0%
eistedd yn ôl
38%
sefyll o'r neilltu
67%
Concept
PV
come back, return
work out, discover
look forward to, eagerly await
help out
pick up, learn
encounter, come across
continue, carry on
sit down
switch off, turn off
sit back
The results of the study have confirmed Research Hypothesis 1 (5.3.2.), in that non-PV
equivalents were more accepted than PVs in general, while the degree of acceptability
varied depending on their status in relation to contact with English. A cursory look at the
table demonstrates that the informants expressed preference for non-PV items over PVs
in all the investigated cases, except for the native PV bwrw ymlaen which was preferred
over single-word parhau. Dialectal variants of PVs: gweithio mas, troi bant, helpu mas
275
and edrych ymlaen i were rejected by the vast majority of speakers, and none of the informants accepted the loanblend troi off. The results of Part A will be now presented in
more detail, considering additional comments made by the informants and patterns in the
acceptability of PVs depending on their category.
In Sentence 1, the speakers were to complete a sentence about a dead soldier’s
body “returning” to Britain. The choice offered was between a PV dod yn ôl and single
verb dychwelyd. As discussed in 2.4.5, although the present study generally focuses on
idiomatic constructions, the case of dod yn ôl was considered worth investigating as it is
given as an example of “replaceable” PV by P. W. Thomas (2.3.3).
Sentence 1. Mae corff swyddog diogelwch o Gymru gafodd ei ladd yn Irac wedi _____ i
wledydd Prydain.
‘The body of a security officer from Wales that was killed in Iraq _____ to the British Isles.’
Answers to Sentence 1 are presented in Figure 9. The majority of participants, 60%, chose
dychwelyd as the best and only option for filling the sentence, while 29% chose both the
PV and non-PV. The remaining 11% thought that dod yn ôl was the only suitable phrase.
Four speakers pointed out directly that dychwelyd was more formal or indicated the formality of the item by saying they would expect to see it in the news. For others, however,
the preferred option was dod yn ôl as the “more common, like Cymraeg Clir” (B8), “more
natural” (C1), “more homely” (G9) or the one they would use it in speech (A6, C9). Three
speakers commented there was no difference between the two verbs. The context of the
sentence played a role as well. Some speakers believed that it is illogical to talk about a
dead body “returning” anywhere, therefore the verb would be suitable only in the passive
voice and dod yn ôl was preferred (482).
82
Throughout the chapter numbers in brackets indicate the number of speakers who chose a given option
or expressed the same opinion.
276
SENTENCE 1
dychwelyd
60%
dychwelyd,
dod yn ôl
29%
dod yn ôl
11%
Figure 9. The verbs chosen by the speakers in Sentence 1 (dod yn ôl/dychwelyd).
All in all, the PV was accepted by 40% of the speakers, which does not suggest a strong
prescriptive norm against using it. The speakers generally indicated that the choice of the
verb would depend on the level of formality. However, some informants perceived a semantic difference, in that the PV was more passive than dychwelyd, which for these
speakers required an agent. There were no comments pointing out that dod yn ôl might
be borrowed from English, which confirms the naturalness of transparent PVs of this type
in Welsh.
All the other sentences in Part A focused on idiomatic PVs which might be seen
as influenced by contact with English. The first group – Sentences 2, 6 and 10 – were PVs
of the pleonastic type: tyfu i fyny ‘grow up’, eistedd i lawr ‘sit down’ and helpu allan
‘help out’. The first two occurred very frequently in the corpus, yet, as shown in 4.5.8 are
also frequently prescribed against. Helpu allan is a less well-established example of an
extended meaning of the particle allan (see 3.3.2.1). The results of the survey for these
constructions are presented in Figures 10-12.
The most accepted of the three pleonastic PVs was eistedd i lawr. It was chosen
by 44% of participants, compared with 23% for tyfu i fyny and 17% for helpu allan/mas.
Sentence 10. Nid yw'r ateb mor syml â dweud wrth y ddwy ochr i _____ a thrafod y
sefyllfa.
‘The answer is not as simple as to tell both sides to _____ and discuss the situation.’
277
SENTENCE 10
both
11%
eistedd i lawr
33%
none
3%
eistedd
53%
Figure 10. The verbs chosen by the speakers in Sentence 10 (eistedd/eistedd i lawr).
As seen in Figure 10, more than half of the speakers chose eistedd as the only acceptable
option, while one third accepted only eistedd i lawr. Two speakers did not choose any
option as they believed it would be better to use other expressions in the context (sefyll
‘stand’ or omit the phrase altogether). Four speakers, who preferred the non-PV item,
commented that eistedd i lawr is a borrowing from English. Others, however, preferred
eistedd i lawr as common in speech (C9, O5), “better” (L1), “making more sense” (Y2),
“natural although it comes from English” (G11). The answers indicate that eistedd i lawr
is a relatively well-established PV, which did not strike the speakers as a borrowing.
A much stronger prescriptive norm was suggested by the answers to Sentence 2
concerning the concept of growing, getting older, expressed with tyfu ‘grow’ or tyfu i fyny
‘grow up’.
Sentence 2. Er gwaethaf y llawdriniaeth brys i'w wyneb bydd rhaid i’r bachgen gael sawl
llawdriniaeth arall, er mwyn ychwanegu croen newydd at ei wyneb wrth iddo _____.
‘Despite emergency face surgery the boy will need several other surgeries to add new skin
to his face as he will _____.’
278
SENTENCE 2
both
7%
tyfu i fyny
16%
tyfu
77%
Figure 11. The verbs chosen by the speakers in Sentence 2 (tyfu/tyfu i fyny).
As shown in Figure 11, the vast majority of participants (77%) chose tyfu as the
verb they would expect to see in the context. As many as 10 speakers pointed out that tyfu
i fyny comes from English and is therefore incorrect. One speaker said she had been taught
that tyfu i fyny is “wrong” (B4). Some speakers did not mention the English influence
directly but said that the particle was unnecessary as it did not add anything to the sentence (3). On the other hand, 16% of speakers considered tyfu i fyny to be the only appropriate option to fill the sentence. Two speakers noticed that the PV is very common and
fully accepted and one speaker commented that “although it is possibly translated from
English”, it “makes sense” (Y2).
The least popular pleonastic construction was helpu allan. The participants had
four options to choose from: cynorthwyo and helpu, both meaning ‘to help’, and PVs
helpu allan and helpu mas ‘help out’ to fill the following sentence:
Sentence 6. Os oes rhieni ar gael i _____ yng ngardd yr ysgol ar brynhawn dydd Iau,
Mawrth 17eg, byddwn yn ddiolchgar iawn.
‘If there are parents available to _____ in the school garden on Thursday afternoon, 17
March, we will be very grateful.’
The participants’ answers are presented in Figure 12. All in all, 84% of speakers
chose a non-PV option with cynorthwyo or helpu; many speakers said the choice between
the two would depend on the level of formality, cynorthwyo being more formal. Four
persons emphasised that helpu allan and mas were translations from English and one said
279
the particle was unnecessary. Helpu mas was accepted by only 10% of speakers, usually
together with other forms, as the particle mas is used dialectally rather than in standard
Welsh.
SENTENCE 6
helpu
18%
cynorthwyo,
helpu
31%
cynorthwyo
34%
helpu allan
7%
cynorthwyo, helpu
mas
4%
helpu mas
2%
helpu, helpu mas
4%
Figure 12. The verbs chosen by the speakers in Sentence 6 (helpu/ helpu allan/ helpu mas/
cynorthwyo).
Generally speaking, as far as pleonastic constructions are concerned, speakers evidently preferred non-PV options to PVs. Speakers who justified their choice often
claimed that in their view the particle was unnecessary. It was particularly noticeable with
helpu allan, where the particle carries relatively little meaning and, what is more, the verb
is already a borrowing from English. Eistedd i lawr, however, appears to be quite acceptable in semi-formal Welsh and for some speakers it carried an additional meaning in comparison with the verb eistedd on its own.
Sentence 3 presented the case of a well-established PV mynd ymlaen whose basic
idiomatic meaning of ‘progressing’, ‘continuing’ was extended to ‘happening’, presumably under the influence of English go on. In the sentence, the participants were offered
two verbs to fill the gap: mynd ymlaen or digwydd ‘happen’:
Sentence 3. Nid pawb, am wahanol resymau, sydd yn gwybod beth yn union yw eisteddfod
a beth sydd yn _____ mewn eisteddfod.
‘Not everybody, for different reasons, knows what exactly eisteddfod is and what _____ in
an eisteddfod.’
280
Answers to the question are presented in Figure 13. The vast majority of informants, 85%, chose digwydd as the only appropriate option, while only 2 participants chose
mynd ymlaen. One of them justified the choice by saying that it meant something else
than digwydd, referring to many things happening at the same time (O6). Only six participants accepted both verbs; B7 commented that the choice would depend on the register
and C8 said there was no difference in meaning; B3 said that both were acceptable, but
digwydd was better. While justifying the choice of digwydd over mynd ymlaen, the participants would either deem mynd ymlaen too informal and spoken (4) or a borrowing
from English (3). L3 commented at this point that digwydd was perhaps not used often
enough and “we should speak more standard”. Similarly, O5 commented that he would
expect to see digwydd but had certainly seen the other option in print.
The low acceptability of mynd ymlaen meaning ‘to happen’ points to the fact that
the semantic extension of the meaning of this PV is not well-established in written semiformal Welsh. The PV was generally viewed as colloquial and non-standard. However,
some speakers indicated that the PV is used in semi-formal registers as well.
SENTENCE 3
digwydd
85%
mynd ymlaen
4%
both
11%
Figure 13. The verbs chosen by the speakers in Sentence 3 (digwydd/mynd ymlaen).
Sentences 4, 7 and 11 concerned more striking cases of idiomatic PVs directly
translatable into English, gweithio allan ‘work out’, ‘discover’, pigo i fyny ‘pick up’,
‘learn’ and troi i ffwrdd/bant/off ‘switch off’, ‘turn off’.
In sentence 4 the speakers were given a choice between gweithio allan, dialectal
gweithio mas and darganfod ‘discover’.
281
Sentence 4. Chafodd neb ei anafu yn y digwyddiad, ac mae ymchwilwyr yn ceisio _____
sut yn union y dechreuodd y tân.
‘Nobody was injured in the incident but investigators try to _____ how exactly the fire
began.’
Answers to the question are shown in Figure 14. This sentence presented one of
the cases when the highest percentage of speakers preferred a non-PV over PV, with over
90% choosing darganfod as the only option. In justifying their choice some participants
said they rejected gweithio allan/mas as a calque from English (3). However, more often,
the PV was considered inappropriate only in the written mode as spoken and dialectal (6).
What is more, several speakers proposed alternative expressions to fill the gap: dyfalu,
canfod, dod o hyd i achos y tân. Similarly to English work out, gweithio allan appears to
belong to the spoken register, which is one possible reason for its rejection by so many
speakers. Moreover, with many other Welsh expressions to convey the idea of “working
something out” the PV does not fill a semantic gap.
SENTENCE 4
darganfod,
gweithio allan
2%
darganfod
93%
none
2%
darganfod,
gweithio allan,
gweithio mas
3%
Figure 14. The verbs chosen by the speakers in Sentence 4 (darganfod/gweithio
allan/gweithio mas).
A similar preference for a non-PV item was observed in Sentence 11 which examined verbs conveying the idea of switching off the lights: diffodd ‘switch off, extinguish’ and PVs troi i ffwrdd, troi bant, troi off:
Sentence 11. Caiff 6,185 o oleuadau eu _____ o ganol nos tan y wawr.
‘6,185 lights were _____ from midnight to the dawn.’
282
SENTENCE 11
troi i ffwrdd
5%
diffodd
82%
diffodd, troi i
ffwrdd
5%
diffodd, troi i
ffwrdd, troi bant
4%
troi bant
2%
diffodd, troi bant
2%
Figure 15. The verbs chosen by the speakers in Sentence 11 (diffodd/ troi i ffwrdd/troi
bant/troi off).
The vast majority of speakers preferred diffodd to the proposed PVs troi i ffwrdd, while
only 10% of speakers chose both the non-PV and PV options. Some speakers emphasised
that diffodd is suitable in the written mode being more formal (4). The dialectal troi bant
was accepted by only 4 participants, three of whom were new speakers, while the fourth
commented that the form was “on the verge” of being acceptable. None of the participants
chose the option troi off, which confirms that the particle off is highly marked and colloquial (see 3.3.2.7). The participants also drew attention to other PVs conveying the same
sense: rhoi off and switsho off. The latter was mentioned by O1 who said she tried to
persuade her little daughter to use diffodd, however the girl “brings” the loan switsho off
from nursery school. The case of diffodd and troi i ffwrdd exemplifies a situation when
the PV clearly belongs to spoken register and is considered too informal for writing.
The next case of a calqued PV gave similar results. In Sentence 7 the participants
were supposed to choose the verb conveying the idea of ‘picking something up’, ‘learning’:
Sentence 7. Does dim ots a ydych chi'n siarad Cymraeg neu beidio, mae'r dosbarthiadau
yn ddwyieithog a byddwch yn _____ yn sydyn yn y ddwy iaith.
‘It does not matter whether you speak Welsh or not, the classes are bilingual and you will
_______ the songs quickly in the two languages.’
283
The participants were offered the following options to complete the sentence:
dysgu’r caneuon ‘learn songs’, pigo'r caneuon i fyny ‘pick up the songs’, a dialectal
variant of the PV pigo'r caneuon lan and another idiomatic expression cael crap ar y
caneuon ‘get a grasp of the songs’, which is suggested as an equivalent of pick up by
Geiriadur yr Academi. The phrasal verb pigo i fyny was rejected by over 70% of the
speakers, three of whom commented that it was a borrowing from English. Only two
participants chose pigo i fyny as the only acceptable option; one of them stated that the
PV had a different meaning than dysgu. Pigo lan was chosen by 3 speakers, all of whom
came from Carmarthenshire, which reflects the local dialect. Only one speaker marked
all 4 options, commenting however that the PVs were “on the border” of being acceptable.
SENTENCE 7
dysgu, pigo i fyny
9%
dysgu, pigo lan
5%
dysgu, cael crap
ar
7%
dysgu
71%
pigo i fyny
4%
all options
2%
cael crap ar, pigo
lan
2%
Figure 16. The verbs chosen by the speakers in Sentence 7 (dysgu/ pigo i fyny/pigo lan/
cael crap ar).
The most commented on element of this example was the proposed idiom cael
crap ar ‘get a grasp on’. It was accepted by only 6 speakers, always together with other
options. Some speakers were entirely unfamiliar with the idiom and found it amusing (5)
because of the word crap ‘grasp’, evoking English crap. Others said that they knew the
idiom and heard it used but would avoid it themselves as it meant something else in English (3), was not understandable to everyone (4), or not suitable in the context (2). The
example provides interesting evidence of the influence of English on the use of Welsh
idioms by native speakers. It shows that alongside rejecting idioms that are borrowed
from English, speakers may also choose not to use a native Welsh idiom in the bilingual
environment because it may sound inappropriate due to the fact it is homogenous with a
284
negatively loaded English word. As regards the PVs pigo i fyny/lan, they can be regarded
as non-standard due to their English-looking form (loanblends) and the existence of a
simple equivalent dysgu.
Such equivalence is difficult to find for edrych ymlaen ‘look forward’, which was
the most frequent PV in the corpus. Sentence 5 examined the case of this PV, proposing
an alternative phrase disgwyl yn eiddgar ‘ardently expect’, given by GA as the equivalent
of look forward. The dictionary overtly discouraged the use of edrych ymlaen (see 4.5.1).
Speakers could also choose a variant of the PV with i as the second particle, which appeared frequently in the corpus (see 3.2.3.2)
Sentence 5. Yn dilyn ymateb gwych i'w nofel gyntaf, mae sawl un wedi bod yn _____
nofel nesaf.
‘Following the great response to the first novel, some people have been _____ (his) next
one.’
As shown in Figure 17, the answers for this PV were much more varied than in
the sentences previously described. Altogether, 62% of speakers accepted edrych ymlaen
at either as the only option (27%) or together with other verbs (35%). Some commented
that the expression disgwyl yn eiddgar sounded “unnatural” (O1), “too formal” (G9) or
“forced” (L3). On the other hand, a relatively high percentage, 38%, chose only disgwyl
yn eiddgar. Three speakers indicated that they liked the expression as a sophisticated one,
but appropriate only for writing; one of the speakers admitted he would never think of
using it himself. Another speaker noticed that although she did not mark edrych ymlaen,
she thought it was quite acceptable, despite it being a borrowing.
The variant edrych ymlaen i was accepted by only two speakers, one of whom was
a new speaker and expressed his lack of certainty over the choice. This points to the nonstandardness of this variant, which is supposedly a recent calque from English.
285
SENTENCE 5
all
4%
disgwyl yn
eiddgar, edrych
ymlaen at
31%
edrych ymlaen at
27%
disgwyl yn
eiddgar
38%
Figure 17. The verbs chosen in Sentence 5 (disgwyl yn eiddgar, edrych ymlaen at,
edrych ymlaen i).
All in all, it can be seen that in the case of a PV which cannot be replaced by a oneword equivalent, the acceptability of PV was much higher. A similar case was investigated in Sentence 12, in which the concept of “keeping away and doing nothing” could
be expressed with two verbs: a PV eistedd yn ôl ‘sit back’ and sefyll o’r neilltu ‘stand
aside’ proposed by GA as the equivalent of sit back.
Sentence 12. Mae angen cynnal mwy o ymarferiadau milwrol i roi neges glir nad ydym yn
_____.
‘One has to hold more military manoeuvres to give a clear message that we do not _____.’
SENTENCE 12
eistedd yn ôl
29%
sefyll o'r neilltu
58%
both
9%
none
4%
Figure 18. The verbs chosen in Sentence 12 (sefyll o’r neilltu, eistedd yn ôl).
286
As seen in Figure 18, the majority of speakers also in this case chose the non-PV
option, which was described by some as “more correct” (A7), or “more formal” (C5, O5).
One speaker rejected the PV eistedd yn ôl saying “it makes no sense” in Welsh (C1). The
PV was accepted by 38% of the speakers; two of them, however, pointed out that it is
borrowed from English. Four speakers stated that none of the options sounded well to
them and proposed alternative expressions: llaesu dwylo ‘stay indifferent’, lit. ‘relax
hands’, gwneud dim ‘do nothing’, gorffwys ar ein rhwyfau ‘be complacent’, lit. ‘rest on
our oars’, cadw draw ‘keep away’. Some speakers pondered whether eistedd yn ôl was
calqued from English as in this context it sounded like a translation of stand back rather
than sit back, parallel to other expressions, such as sit an exam, which in Welsh is sefyll
arholiad lit. ‘stand an exam’. These speakers perceived eistedd yn ôl as a kind of loan
rendition, which shows their sensitivity to contact phenomena. To the majority, however,
the phrase did not seem to be strongly associated with English. Despite the fact that the
usage of the particle yn ôl in an idiomatic sense is quite rare, as shown in 3.3.2.6, the PV
was not identified as an obvious calque.
Sentence 9 examined the acceptability of competing native and borrowed PVs
conveying the notion of “continuing” something: bwrw ymlaen and cario ymlaen alongside the single verb parhau.
Sentence 9. Er i’r cyhoeddwr gael cynnig grant gan y Cyngor Llyfrau ar gyfer cyhoeddi,
penderfynodd _____ â'r fenter heb y grant.
‘Although the publisher received an offer for a grant from the Book Council for publishing,
they decided to _____ with the project without the grant.’
As shown in Figure 16, the native PV bwrw ymlaen was chosen by nearly two
thirds of the speakers, usually together with other options, while one third of the speakers
accepted only parhau. Cario ymlaen was chosen by 21% of the informants, but only to
speakers who marked it as the only option. One speaker who marked all three options
commented that although cario ymlaen is acceptable, the other two would be her first
choice. Another speaker believed that both PVs are “very deeply rooted in the language”.
On the other hand cario ymlaen was rejected by 6 speakers due to the fact that it was a
borrowing Five speakers said they liked bwrw ymlaen as the most natural expression,
while one speaker believed that it is also a borrowing like cario ymlaen.
287
SENTENCE 9
parhau, bwrw ymlaen,
cario ymlaen
13%
parhau, bwrw
ymlaen
29%
cario ymlaen
3%
bwrw ymlaen
18%
parhau, cario
ymlaen
2%
parhau
33%
bwrw ymlaen,
cario ymlaen
2%
Figure 19. The verbs chosen in Sentence 9 (parhau/bwrw ymlaen/cario ymlaen).
Finally, Sentence 8 checked the acceptability of two PPs, both conveying the sense
of ‘encountering something suddenly’, one of which is directly translatable into English:
dod ar draws ‘come across’ and taro ar ‘strike, hit on’.
Sentence 8. Mae’n rhaid i ni herio peryglon gwahaniaethu ac eithafiaeth o bob math a
pheidio byth â sefyll o’r neilltu pan fyddwn yn _____ a chasineb.
‘We have to challenge the dangers of discrimination and extremism of any kind and never
stand aside when we _____ hatred.’
SENTENCE 8
both
22%
dod ar draws
54%
none
2%
taro ar
22%
Figure 20. The verbs chosen in Sentence 8 (dod ar draws/ taro ar).
The PV dod ar draws was accepted by the majority of speakers (76%), with 54%
selecting it as the only option. Some speakers stated that dod ar draws was more natural
288
and understandable “even though it is influenced by English” (4), while for B3 it was
more natural than “awkward” taro ar. For one speaker taro ar was more formal (C5) and
appropriate in the context, while for O6 it was too formal. B7 found it inappropriate in
the context, and C4 (a new speaker) said she had never heard taro ar. A few speakers
proposed alternative expressions: wynebu ‘face’ (C7, B8), gweld ‘see’ (B3), dod wyneb
yn wyneb â ‘come face to face’ (G11). Only one speaker (B8) rejected both items and
proposed wynebu as an appropriate choice. A summary of the results for Part A is presented in Table 26.
Table 25. Percentage of speakers accepting PVs and alternative expressions in Part A of
the questionnaire in descending order.
PVs
%
non-PVs
%
dod ar draws
76%
darganfod
98%
edrych ymlaen at
62%
digwydd
96%
bwrw ymlaen
62%
dysgu
95%
taro ar
44%
diffodd
93%
eistedd i lawr
44%
dychwelyd
89%
dod yn ôl
40%
tyfu
84%
eistedd yn ôl
38%
disgwyl yn eiddgar
73%
tyfu i fyny
24%
cynorthwyo
69%
cario ymlaen
20%
sefyll o'r neilltu
67%
mynd ymlaen
15%
eistedd
64%
pigo i fyny
15%
helpu
53%
troi i ffwrdd
15%
parhau
44%
helpu mas
9%
cael crap ar
11%
pigo lan
9%
helpu allan
7%
troi bant
7%
gweithio allan
6%
gweithio mas
4%
edrych ymlaen i
4%
troi off
0%
The results have confirmed the research hypothesis that the degree of acceptability would
depend on the PV’s status with regard to contact with English. The most acceptable constructions were PVs which are well-established and frequent, despite being directly translatable into English and where the particle sense is natural for the Welsh-language. The
289
third most acceptable PV was native bwrw ymlaen, the only PV which was chosen by
more speakers than the one-word equivalent parhau.
The three examples of pleonastic PVs in the questionnaire have confirmed
Thomas’s (1996) and Rottet’s (2005) observations that in more formal registers the equivalents without the aspectual particle, such as eistedd, tyfu, helpu, will be preferred over
pleonastic PVs. The degree of acceptability varied, however, in the three examined cases.
Eistedd i lawr was identified as a borrowing by fewer participants than tyfu i fyny and as
a consequence was accepted by a higher percentage of speakers; the much less well-established helpu allan/mas, which contains a borrowed verb and where the particle is almost semantically empty, scored much lower. The degree of acceptability of pleonastic
PVs may therefore depend on their perceived status as calqued constructions.
The questionnaire revealed a strong prescriptive norm against other calqued constructions examined. As expected, the most rejected PVs were dialectal variants of idiomatic PVs and a loanblend troi off with the particle borrowed directly from English. The
loanblend cario ymlaen was accepted by one fifth of the speakers, scoring much lower
than native bwrw ymlaen, although in the corpus the two items appeared with equal frequency. Noticeably, this loanblend was more accepted than calqued troi i ffwrdd/bant,
pigo i fyny/lan and gweithio allan/mas, which were considered inappropriate in the contexts by the vast majority of speakers. This is slightly surprising, bearing in mind the high
frequency of these items in the corpus. However, the PVs appear to be identifiable as
calques from English due to the fact that their particle senses are relatively new and not
well-established; moreover, there is a strong prescriptive norm against these specific construction, especially troi i ffwrdd (e.g. Thomas 2012:162). What is perhaps more surprising is the low acceptability of the calque mynd ymlaen ‘happen’, which is a sematic extension of a well-established mynd ymlaen ‘to continue’. This can be attributed to close
synonymy with a simple equivalent digwydd ‘happen’ and also the more informal character of the PV.
Contrary to expectation, the PV edrych ymlaen at was less accepted than the rather
sophisticated equivalent disgwyl yn eiddgar ‘ardently expect’, despite the high frequency
of the PV in the corpus and the fact it does not have a close native Welsh equivalent.
Nevertheless, the PV was the most acceptable one among all the directly translatable constructions, which demonstrates that it has its place in standard written Welsh. When
290
speakers had no choice of a single-word item, the more frequent and well-established
PP dod ar draws was preferred over native taro ar.
The PV eistedd yn ôl seemed to be somewhat perplexing for the participants. It
appears to be a little established item as for some informants it was unclear whether the
PV is a loan rendition of stand back or direct translation of sit back.
5.4.3. Questionnaire Part B
5.4.3.1. Results
In part B of the questionnaire the participants were given four sets of sentences. The sets
were introduced by putting them in the context of four imagined situations, representing
different registers and levels of formality: spoken informal (R1), spoken formal (R2),
written semi-formal (R3) and written official (R4).
Figure 21 and Table 26 present an overview of the results of Part B of the questionnaire, by giving the percentage of the participants who chose a PV in a given situation.
As expected, the perceived acceptability varied considerably not only across registers, but
also across the selected PVs, depending on their status and meaning.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
edrych ar ôl
ffeindio allan
Spoken informal
rhedeg allan o
Spoken formal
torri i lawr
Written semi-formal
troi i fyny
troi ymlaen
Written official
Figure 21. Part B – perceived acceptability of selected PVs in different registers.
291
Table 26. Part B – perceived acceptability of selected PVs in different registers
PV
Spoken
informal
Spoken
formal
Written
semi-formal
Written
official
torri i lawr
95%
89%
60%
78%
troi ymlaen
95%
89%
64%
75%
troi i fyny
85%
65%
67%
31%
edrych ar ôl
87%
58%
47%
45%
rhedeg allan o
76%
69%
53%
44%
ffeindio allan
87%
44%
27%
22%
The results of the study confirmed Research Hypothesis 2 (5.3.2.). All of the selected verb-particle constructions were declared as being used in informal conversation
by the vast majority of speakers. The results for spoken formal register were more mixed
and depended on the type of the PV. As had been predicted, in formal and semi-formal
registers the most accepted PVs were torri i lawr and troi ymlaen, a pleonastic construction and a PV filling a lexical gap. Similarly, troi i fyny in its technical sense was also
markedly more accepted than in the sense of ‘arriving, showing up’ (67% and 31%, respectively). However, the technical meaning of torri i lawr was less accepted than the
more popular sense of ‘stop working’ (see below). Relatively well-established calques of
three syntactic types (PP, APV, PPV) edrych ar ôl, troi i fyny and rhedeg allan o, while
highly accepted in speech, were avoided in writing by the majority of speakers. The
loanblend ffeindio allan, was accepted only by a small proportion of speakers in writing,
however, it was highly accepted in speech. A more detailed look at the results concerning
the six selected PVs provides some valuable data on their integration.
Table 27. Sentences in the questionnaire containing torri i lawr ‘break down’.
Register
Welsh sentence
English translation
R1
Torrodd y car i lawr83 a doedd hi
ddim yn gallu fforddio'r costau cynnal.
Mae cerbydau yn torri i lawr a dydy
casgliadau gwastraff bwyd ddim yn
digwydd.
The car broke down and she could
not afford maintenance costs.
R2
Vehicles break down and the food
waste collection doesn’t take place.
Acceptability
95%
89%
83
This is the only sentence in Part B that contains a synthetic rather than periphrastic form of the verb as it
was considered the most natural form in this context. Despite the fact that synthetic forms may be perceived
as more formal, in this case it has clearly not affected the speakers’ choices since the degree of acceptability
for this sentence was 95%.
292
Mae sylweddau nad yw’n bosibl eu
torri i lawr yn cael eu galw’n sylweddau hirbarhaus.
Mae gan staff yr ysbyty bŵer i
symud cerbydau sydd wedi'u gadael
neu sydd wedi torri i lawr.
R3
R4
Substances which are impossible to
be broken down are called highly
persistent substances.
The hospital staff have the power to
move vehicles which were left or
broke down.
60%
78%
Table 27 presents the sentences containing torri i lawr ‘break down’. The vast
majority of speakers (95%) declared that they would use the phrase in speech in the context of vehicles breaking down. The construction was also accepted by the majority of
speakers in spoken formal and written official registers (89% and 78%, respectively).
Speakers who rejected the PV in these contexts often proposed the verb without the particle, torri, as the expression they would use (6). One speaker accepted the PV on the
grounds that, since “so many people say it”, it could be used in a leaflet, although she
herself disapproved of it. Generally speaking, however, the vast majority of participants
did not see the PV as inappropriate. Markedly fewer speakers, 60%, accepted the PV in
situation 3, in which it was used in the sense of ‘separating into parts’. The sentence
referred to “breaking elements down” and was originally taken from an educational website for young people studying biology. It was observed that some speakers found it difficult to propose an equivalent for the PV in this sentence; some of them admitted they
could not think of anything (2) or stated that it should be “something technical” (2). The
proposed alternatives to the PV were dadelfenni ‘analyse’ (G11), datgymalu ‘disjoint’
(C1), pydru ‘decompose’ (C1, C6) and daddansoddi ‘analyse’ (A3, Y1). This untypical
example may be considered to illustrate lack of confidence of professional speakers with
regard to technical vocabulary.
Table 28. Sentences in the questionnaire containing troi ymlaen ‘turn on’.
Register
R1
R2
R3
R4
Welsh sentence
Bob tro wyt ti’n troi’r teledu ymlaen, maen nhw’n siarad am y
gemau.
Pan wnes i droi y radio lleol ymlaen, roedd pob cân yn Saesneg a’r
unig Gymraeg a glywais mewn
mwy na hanner awr oedd un llinell
mewn newyddion
Os bydd ffermwyr yn troi'r dŵr
ymlaen heb angen gwneud hynny,
gall y tir wynebu syched.
Gall cleifion droi’r larwm ymlaen
pan fo angen.
English translation
Every time I switch on the TV they
are talking about the games.
Acceptability
95%
When I turned the local radio on every song was in English and the only
Welsh I heard in more than half an
hour was a single line in the news.
89%
If farmers unnecessarily turn the water on, the land may face drought.
64%
Patients can turn the alarm on if
need arises.
75%
293
Very similar results were obtained for the PV troi ymlaen. The sentences containing this PV referred to the concept of switching/turning on electric devices: television,
radio, alarm or, in R3, water (Table 28). Also in this case the PV was markedly less accepted in a popular book than in an official leaflet, while being almost universally accepted in speech. One possible reason for the lower acceptability in R3 might be attributed
to the less common context of turning on water, rather than a device. Another explanation
was given by speaker Y2, who said the PV would be appropriate in the official register
because it is very common, in a book however she would avoid it by using the phrase
defnyddio dŵr ‘use water’. Several other speakers proposed this and other verbs to be
used in Situation 3: defnyddio ‘use’ (C1, G9, G11, A5, Y1, Y2), dyfrio ‘water’ (G5),
gwastraffu ‘waste’ (A7). Noticeably, in the spoken informal register, speakers also proposed alternative expressions they would use in everyday life: a dialectal variant rhoi
ymlaen (A4, A5, B3, B5 – all these speakers came from North Wales), troi ar (G6), troi
arnodd (O6) and troi on (C6, O6). All these speakers stated, however, that they would
use the more standard troi ymlaen in the formal situation. Some speakers who rejected
the PV in R4 proposed canu ‘ring’ (L6, Y1), tanio ‘ignite’ (G6) and defnyddio ‘use’ (L8,
B8, Y1) as alternatives.
Table 29. Sentences in the questionnaire containing troi i fyny ‘turn up’.
Register
R1
R2
R3
R4
Welsh sentence
Faint o bobol wnaeth droi i fyny/
droi lan i’r cyfarfod?
Dyw hi ddim yn anarferol i bedair
neu bump ambiwlans droi i fyny/
droi lan â chleifion o fewn hanner
awr.
Mae’n werth cadw rheiddiaduron
yn isel am gyfnodau hirach yn hytrach na’u troi i fyny/ troi lan am rai
oriau.
Rydym yn chwilio am wirfoddolwyr ag egni a syniadau – croeso
mawr i bobl o bob oed ac o bob
rhan o'r ardal droi i fyny/ droi lan.
English translation
How many people turned up to
the meeting?
It is not unusual for four or five
ambulances to turn up with patients within half an hour.
Acceptability
85%
It’s a good idea to keep radiators
low for longer periods rather than
turn them up for some hours.
67%
We are looking for volunteers
with energy and ideas – people of
all ages and from every part of the
region are warmly welcome to
turn up.
31%
65%
In the case of troi i fyny there was a meaningful difference between the acceptability of the PV in a book and a leaflet – 67% and 31%, respectively. In Situation 3 the
verb referred to turning up a radiator, while in R4 it denoted arriving at a place (Table
29). This discrepancy may, as previously, be explained by the fact that in R3 the verb was
294
used in a technical sense difficult to replace. Three speakers admitted they struggled with
finding an equivalent and could not think of anything else. In fact, only one speaker proposed an alternative phrase for this sentence: defnyddio ar wres uchel ‘use at high heat’
(G6). In contrast, in R4 the majority of participants could immediately find an acceptable
word or expression to replace troi i fyny in the sense of arriving: dod ‘come’ (9), dyfod
‘come’ (1), ymuno ‘join’(4), cyfrannu ‘contribute’ (1), mynychu ‘attend’ (3 ), bod yn bresennol ‘be present’ (1), cymryd rhan ‘take part’(1), dod draw ‘come over’ (1), galw heibio
‘call by’ (1) – the latter two being native PVs. Two speakers believed that the phrase
should be deleted from the sentence altogether. Although the PV was much more accepted
in informal speech (85%), 6 speakers said they would rather use the verb dod ‘come’. In
formal speech, where the verb was less preferred (65%) the most common proposed alternative was more formal cyrraedd ‘arrive’ (7); other options mentioned were mynychu
‘attend’ (1) and ymddangos ‘appear’ (2). The dialectal variant troi lan was usually chosen
only in the spoken informal (10) or in both spoken registers (8) by speakers from South
or Mid Wales. Only two speakers chose troi lan in every context.
In general, the verb troi i fyny in its main sense was associated with the informal
register, with speakers being sensitive to differences between spoken and written mode.
However, only one speaker indicated directly that the PV was “incorrect” on paper.
Table 30. Sentences in the questionnaire containing edrych ar ôl ‘look after’.
Register
R1
R2
R3
R4
Welsh sentence
All dy chwaer edrych ar ôl Mabon
benwythnos yma?
Mae henoed sy'n edrych ar ôl plant
yn dioddef yn ariannol
Mae ffermio coco yn golygu edrych ar ôl y coed coco a chynaeafu'r codau, eplesu a sychu'r ffa a'u
pacio mewn sachau yn barod i gael
eu troi yn eich hoff far o siocled.
Os ydych chi’n edrych ar ôl
rhywun ag anghenion gofal a chymorth a fyddai'n methu ymdopi heb
eich help chi, rydych yn ofalwr.
English translation
Can your sister look after Mabon
this weekend?
Elderly people who look after
children suffer financially.
Farming cocoa involves looking
after cocoa trees and harvesting
the pouches, fermenting and drying the beans and packing them in
sacks ready to be turned into your
favourite bar of chocolate.
If you look after someone who
needs care and assistance and
who could not cope without your
help, you are a caretaker.
Acceptability
87%
58%
47%
45%
The case of the prepositional verb edrych ar ôl also demonstrated the difference
between spoken and written language. The verb was used in the same sense in all situations, referring to taking care of children, plants or sick patients (Table 30). The PV was
295
highly accepted only in informal speech (87%), while in formal speech the percentage
dropped to 58% and in written registers it oscillated around 45%. This is explained by the
existence of several Welsh verbs which could replace edrych ar ôl. The number of speakers proposing alternative verbs reflected the proportion of acceptability: only 5 speakers
suggested equivalents for R1, while it was 20 for R2 and 19 for R3 and 4. The most
popular alternatives suggested were gofalu ‘look after’ (R1– 2; R2 – 6; R3 – 11; R4 – 16)
and gwarchod ‘look after’ (R1– 3; R2 - 14; R3 – 5; R4 - 1). Other verbs proposed were
cynnal ‘support’ or meithrin ‘nourish’ (1), trin ‘cultivate’, ‘treat’ (R3 – 2, R4 – 1) and
amdiffyn ‘protect’ (R3 – 1). Although the PV was highly accepted in the informal spoken
register, in more formal contexts half of the speakers expressed preference for one-word
equivalents.
Table 31. Sentences in the questionnaire containing rhedeg allan o ‘run out of’.
Register
R1
R2
R3
R4
Welsh sentence
Mae’r ffôn wedi rhedeg allan o /
rhedeg mas o fatri eto.
Mae’r llywodraeth wedi colli ysbrydoliaeth a rhedeg allan
o/rhedeg mas o stem.
Mae llawer un yn pryderu byddwn
yn rhedeg allan o/ rhedeg mas
o'r olew.
Os ydych wedi rhedeg allan
o/rhedeg mas o‘r feddyginiaeth,
ewch i fferyllfa sydd cymryd rhan
yn y rhaglen.
English translation
The phone has run out of battery
again.
The government has lost inspiration
and run out of steam.
Acceptability
76%
Many people worry that we will
run out of oil.
53%
If you have run out of medicine, go
to the pharmacy which takes part in
the programme.
44%
69%
The phrasal prepositional verb rhedeg allan o ‘run out of’ was the least accepted
of the six PVs examined in informal speech (see Table 26), yet its acceptability was still
rather high, at 76%. Notably, there was no large difference between the acceptability of
this PV in informal and formal speech (69%). This may be explained by the fact that in
R2 the verb was used in an idiomatic expression rhedeg allan o stêm, ‘run out of steam,
energy’ (Table 31). The speakers found it difficult to find an alternative and only few of
them proposed equivalents: colli stêm ‘lose steam’ (4), dim stêm ‘no steam’ (1), diffyg
stêm ‘lack of steam’ (1). In R1 the proposed equivalents for “running out of battery” were:
mynd yn fflat ‘go flat’ (2), wedi mynd ‘gone’ (1), darfod ‘finish’ (2), mynd yn marw ‘go
dead’ (1), gorffen ‘finish’ (1). In R3 speakers proposed dim ar ôl ‘no left’ (1), dod i ben
‘finish’ (2), dirwyn i ben ‘finish’ (1), na fydd digon ar gael ‘there won’t be enough available’ (1), gorffen ‘finish’ (1), while in R4 the suggested options were: gorffen ‘finish’
296
(10), nad oes ar ôl ‘that there is no left’(4), os nad oes ‘if there is no’ (1) and dod i ben
‘finish’ (3). Speakers from South and Mid Wales chose the dialectal variant rhedeg mas
in speech (10) or only in informal speech (3). Only two speakers chose this variant also
in informal writing, and one speaker in all four situations.
Generally speaking, the participants considered this PV to belong to the informal
registers of Welsh. One speaker openly stated that it is too informal for writing (O4).
Speaker A1, a teacher of Welsh, mentioned the PV of his own accord in the opening
interview as an example of a calque from English. Upon seeing it in the questionnaire he
added that the expression did not make sense and that he would try to draw the attention
of his students to the “mistake” by asking “run where?” if they use it. However, curiously,
the speaker accepted the PV in R4.
Table 32. Sentences in the questionnaire containing ffeindio allan ‘find out’.
Register
R1
R2
R3
R4
Welsh sentence
Wnaeth rhywun ffeindio allan/ ffeindio mas bod nhw’n gwerthu tocynnau o hyd.
Rwyf wedi siarad â’r staff Gwasanaethau Ambiwlans, ac wedi ffeindio
allan/ ffeindio mas byddai’n haws
iddyn nhw ddefnyddio’r wybodaeth
pe bai GPS ar gael.
Pe gallwn ffeindio allan/ffeindio
mas mwy am effaith y cemegyn,
byddwn ni’n gwybod sut i ymdopi â’r
broblem.
Gallwch ffeindio allan/ ffeindio mas
rhagor am hyfforddiant a gweithgareddau dan y ddolen isod.
English translation
Someone found out that they are
still selling tickets.
Acceptability
87%
I have spoken with the staff of
Ambulance Services and found
out that it would be easier for
them to use the information if
GPS was available.
If we could find out more about
the effect of the chemical we
would know how to cope with
the problem.
You can find out more about the
training and activities under the
link below.
44%
27%
22%
Finally, the PVs ffeindio allan (Table 32) was the least accepted one in registers
other than informal speech, where it was chosen by a very high proportion of the speakers
(87%). Only two speakers suggested an alternative in this register: canfod ‘find’ (1) and
ffeindio ‘find’ (1). In contrast, in formal registers, the majority of speakers rejected the
PV. The most common alternative proposed was darganfod ‘find’, ‘discover’ (R2 – 10;
R3 – 14; R4 – 17). Other equivalents were: canfod ‘find’ (R2 – 2; R3 – 6; R4 – 1), cael
gwybod ‘get to know’(R2 – 1), cael rhagor am ‘get more about’ (R2 – 1), dysgu ‘learn’
(R2 – 1, R4 – 1), deall ‘understand’ (R2– 1), cael mwy o wybodaeth ‘get more information’ (R4 – 5), dod o hyd i wybodaeth ‘find information’ (R4 – 1), cael hyd i wybodaeth
297
‘find information’ (R4 – 1), darllen rhagor ‘read more’ (R4 – 1). Many speakers from
South and Mid Wales chose the dialectal variant ffeindio mas in speech (13). Only 5
speakers accepted ffeindio allan in all registers and there was an exceptional case of a
speaker who accepted the PV only in R4 (L9).
5.4.3.2. Extralinguistic variables – statistical analysis
The results of Questionnaire B were analysed statistically with respect to three extralinguistic variables which were considered potentially valid regarding the homogenous character of the sample: sex, age and education (see 5.4.1.) Participants’ answers were transposed into scores 0-6 for each PV in each of the four registers (R1, R2, R3, R4). The total
score was also calculated for each participant. The data was analysed in SPSS and R
statistical software.
Prior to verifying each hypothesis the distributions of variables were tested for
their conformance with the normal distribution. If the distribution was normal the hypotheses were verified using parametric student’s t-test; otherwise, the non-parametric MannWhitney U test was applied. The equality of variance was assessed with Levene’s test for
each t-test and each time the null hypothesis of equal variances was not rejected. In the
analysis of correlation regarding the variable age, the distribution of all variables was
non-normal, therefore non-parametric Spearman's rank correlation coefficient was used.
Sex
Firstly, the distributions for all four registers and the total scores of the sample divided
by sex were checked for the normality of distribution. A Shapiro-Wilk test of normality
found significant departures from the normal distribution in scores R1, R2, and R3, and
no significant departure from normality in the R4 score and total score.
Student’s t-test analysis showed that there were no significant differences between
sexes in R4 and the total score (Table 33).
298
Table 33. The results of student’s t-test for differences between sexes in Register 4 and total
score with normal distribution.
Sex of the speaker
R4
total
N
M
SD
male
19
2.79
1.40
female
36
3.08
1.68
male
19
16.05
4.21
female
36
15.67
4.96
t
df
p
-0.65
53
0.52
0.29
53
0.77
The analysis with Mann-Whitney U test showed no statistically significant differences between sexes for R1, R2 and R3 (Table 34). This means that there is no evidence
of male - female differences with respect to the acceptability of phrasal verbs.
Table 34. The results of the Mann-Whitney U test for differences between sexes for Registers 1, 2 and 3 with non-normal distribution
Sex of the speaker
N
MR
male
19
female
R1
R2
R3
Percentile
25
50
75
27.97
5
6
6
36
28.01
5
6
6
male
19
29.76
4
4
6
female
36
27.07
3
4
6
male
19
29.71
1
3
5
female
36
27.10
2
3
5
Z
p
-0.01
0.99
-0.61
0.54
-0.58
0.56
MR- Mean Rank
Age
The second variable tested was age. The participants were assigned to six age groups
which were then transposed into ranks for calculations. A Shapiro-Wilk test of normality
found significant departures from the normal distribution in all questionnaire scores except for the total score. Therefore, non-parametric Spearman's rank correlation coefficient
was used to verify the correlation hypothesis. The analysis did not show any significant
correlations between age of the participants and the score obtained in the questionnaire
(Table 35). Thus the analysis provided no evidence for participants’ age influencing the
acceptability of phrasal verbs.
299
Table 35. The results of Spearman's test for correlation between age and Questionnaire B
scores for Registers 1-4 and the total score.
age
ρ
p
N
R1
-0.03
0.82
55
R2
-0.14
0.30
55
R3
-0.09
0.50
55
R4
-0.24
0.08
55
Total score
-0.20
0.14
55
Education
The third variable examined was the level of education of the participants. The majority
of the studied speakers (47) had received university education, while only 8 participants
had been through secondary education. A Shapiro-Wilk test of normality found significant departures from the normal distribution in all scores except for the total score of the
two groups distinguished according to the level of education.
Student’s t-test found a significant difference in the total score between speakers
with secondary and higher education t (53) = 2.35, p < .05 (Table 36). Speakers with
higher education had significantly lower scores in the total sum of accepted PVs (M =
15.21, SD = 4.43) in comparison with speakers with secondary education (M = 19.25, SD
= 4.92) (Figure 22).
Table 36. The results of student’s t-test for differences between speakers with secondary
and higher education in the total score.
Education
Total score
N
M
SD
secondary
8
19.25
4.92
higher
47
15.21
4.43
t
df
p
2.35
53
<.05
300
25,00
19,25
20,00
Total score
15,21
15,00
10,00
5,00
0,00
Secondary
Higher
Education
Figure 22. Mean results for the total score according to education.
The analysis with the non-parametric Mann-Whitney U test (Table 37) showed a
significant difference between speakers in R3, the semi-formal written register, Z = 2.69,
p <.01. Speakers with higher education had a significantly lower score in R3 (Mdn = 3.00)
than speakers with secondary education (Mdn = 6.00) (Figure 23).
Table 37. The results of the Mann-Whitney U test for differences between speakers with
secondary and higher education in Registers 1-4.
R1
R2
R3
R4
Education
N
MR
secondary
8
higher
Percentile
25
50
75
31.25
5.25
6.00
6.00
47
27.45
5.00
6.00
6.00
secondary
8
31.00
3.25
5.00
6.00
higher
47
27.49
3.00
4.00
6.00
secondary
8
41.88
5.00
6.00
6.00
higher
47
25.64
2.00
3.00
4.00
secondary
8
37.56
2.25
5.00
6.00
higher
47
26.37
2.00
3.00
4.00
Z
p
0.72
0.55
0.59
0.58
2.69
<.01
1.88
0.07
MR- Mean Rank
301
7,00
6,00
6,00
6,00
5,00
R3 score
5,00
4,00
4,00
3,00
3,00
2,00
2,00
1,00
0,00
secondary
high
Education
Percentile: 25
Percentile: 50
Percentile: 75
Figure 23. Mean results for R3 scores according to education.
The statistical analysis showed that within the studied group of professional speakers
there were significant differences with regard to the level of education in the written formal register and in the total score. To scrutinise the matter it is worth presenting acceptability percentage results for the two groups compared with the mean score of all the
participants of the study. The results for the spoken register are presented in Table 38,
while Table 39 shows the results for the written mode. Characteristic differences between
the two groups are marked in bold.
Table 38. Acceptability of PVs in Part B in spoken registers with regard to speakers’ level
of education.
PV
Secondary Higher edu- All speakeducation
cation
ers
R1 - Spoken informal
Secondary
education
Higher eduAll speakcation
ers
R2 - Spoken formal
torri i lawr
88%
98%
95%
88%
94%
89%
troi ymlaen
100%
96%
95%
100%
91%
89%
troi i fyny
100%
85%
85%
75%
68%
65%
edrych ar ôl
100%
87%
87%
75%
57%
58%
rhedeg allan o
88%
77%
76%
100%
68%
69%
ffeindio allan
88%
89%
87%
75%
43%
44%
302
Table 39. Acceptability of PVs in Part B in written registers with regard to speakers’ level
of education.
PV
Secondary
Higher eduAll speakeducation
cation
ers
R3 - Written semi-formal
Secondary
Higher eduAll speakeducation
cation
ers
R4- Written official
torri i lawr
88%
55%
60%
88%
79%
78%
troi ymlaen
88%
60%
64%
75%
77%
75%
troi i fyny
88%
66%
67%
63%
26%
31%
edrych ar ôl
88%
43%
47%
75%
40%
45%
rhedeg allan o
88%
47%
53%
63%
40%
44%
ffeindio allan
63%
21%
27%
50%
17%
22%
The above results point to the following observations:
1. The acceptability of phrasal verbs was generally negatively correlated with the
level of education.
2. In the spoken registers there were no statistically significant differences between
speakers with secondary and higher education regarding the acceptability of PVs.
3. However, it should be noted that in the spoken formal register acceptability scores
for rhedeg allan o and ffeindio allan were markedly higher among speakers who
had received no higher education.
4. In the written semi-formal register there were statistically significant differences
between speakers with secondary and higher education. All the PVs had significantly higher scores of acceptability among speakers with no higher education.
5. There were no statistically significant differences in the written official register;
however, a trend of higher acceptability among less educated speakers was observed in all cases except for troi ymlaen.
6. Among speakers educated at secondary level the acceptability of PVs was lower
in the written official compared with the written semi-formal register.
Comparison of Groups B and L
In view of the above findings, it was considered worthwhile investigating differences
between the two groups which had the highest and the lowest percentage of speakers
having been through higher education (100% Groups B and 60%, in Group L84). Since
84
It can be noted that the participants in Group L with no higher education (4 speakers) comprised half of
the whole group in the sample (8 speakers).
303
this factor proved to be important, the two groups might be expected to have markedly
different results which might point to other extralinguistic factors influencing their judgements. The answers of these participants are presented in Table 40.
Table 40. The results of Part B for Groups B and L compared with the mean.
Spoken informal
Spoken formal
Written semiformal
Written official
all
95%
B
100%
89%
60%
78%
88%
38%
88%
L
90%
90%
100%
80%
all
95%
89%
64%
75%
B
100%
100%
25%
63%
L
100%
100%
100%
80%
all
85%
65%
67%
31%
B
88%
25%
38%
13%
PV
Group
torri i lawr
troi ymlaen
troi i fyny
edrych ar ôl
rhedeg allan o
ffeindio allan
L
100%
80%
100%
60%
all
87%
58%
47%
45%
B
100%
63%
63%
75%
L
100%
50%
80%
60%
all
76%
69%
53%
44%
B
38%
75%
0%
0%
L
100%
80%
90%
70%
all
87%
44%
27%
22%
B
88%
50%
0%
0%
L
90%
60%
50%
30%
The differences between the two groups were further examined with statistical analysis. Comparing Groups B and L, a Shapiro-Wilk test found significant departures from
the normal distribution in scores R1, R3 and R4 and no significant departure from normality in the R2 score and total score.
Student’s t-test showed a significant difference in the total score between speakers
in the analysed groups, t (16) = 3.87, p < .001 (Table 41). Speakers in Group L had a
significantly higher total score (M = 19.30, SD = 3.40) in comparison with Group B (M
= 13.13, SD = 3.31) (Figure 24).
Table 41. The results of student’s t-test for differences between Groups B and L for the
total score and R2.
group
total
L
N
M
SD
t
df
p
10
19.30
3.40
3.87
16
<.001
304
R2.
B
8
13.13
3.31
L
10
4.60
1.43
B
8
4.00
1.41
0.89
16
0.39
25,00
19,30
Total score
20,00
15,00
13,13
10,00
5,00
0,00
L
B
group
Figure 24. Mean total score in Groups L and B.
The Mann-Whitney U test showed a significant difference between speakers in Groups B
and L in R1, the spoken informal register, Z = 2.08, p <.05 (Table 42). Speakers in Group
L had significantly higher acceptability scores in R1 (Mdn = 6.00) compared with Group
B (Mdn = 5.00) (Figure 25). The analysis also showed significant differences in R3, the
written semi-formal, Z = 3.45, p <.001. In this register speakers in Group L had significantly higher scores (Mdn = 5.50) compared with Group B (Mdn = 1.50) (Figure 26).
Table 42. The results of the Mann-Whitney U test for differences between Groups B and L
for R1, R3 and R4 with non-normal distribution.
R1
R3.
R4
group
N
MR
L
10
B
Percentile
25
50
75
11.45
6.00
6.00
6.00
8
7.06
4.25
5.00
6.00
L
10
13.30
4.50
5.50
6.00
B
8
4.75
1.00
1.50
2.75
L
10
11.20
2.00
3.50
6.00
B
8
7.38
1.25
3.00
3.00
Z
p
-2.08
<.05
-3.45
<.001
-1.57
0.12
MR- Mean Rank
305
7,00
6,00
6,00
6,00
6,00
6,00
5,00
R1 score
5,00
4,25
4,00
3,00
2,00
1,00
0,00
L
B
group
Percentile: 25
Percentile: 50
Percentile: 75
Figure 25. Percentile distribution of R1 scores for Groups B and L,
7,00
6,00
5,50
6,00
R3 score
5,00
4,50
4,00
2,75
3,00
1,50
2,00
1,00
1,00
0,00
L
B
group
Percentile: 25
Percentile: 50
Percentile: 75
Figure 26. Percentile distribution of R3 scores for Groups B and L.
Altogether, statistical analysis pointed to significant differences between members of
Groups B and L in R1, R3 and the total score. Looking at the percentages of acceptability
one can see that the most striking differences between the two groups were observed in
the acceptability of directly calqued PVs troi i fyny, rhedeg allan o and ffeindio allan in
the written registers. The PVs were almost universally rejected (0%-13%) by members of
Group B, except for troi i fyny in the written semi-formal register, which was accepted by
a higher proportion of speakers (38%) presumably due to the fact that it was used in the
more technical sense of ‘increasing the heat’. In R3 (writing a popular book) the majority
306
of speakers in Group B rejected also other calqued PVs, torri i lawr and troi ymlaen,
although they accepted them in the written official register. The only PV accepted by the
majority of Group B was edrych ar ôl and in R4 it was selected by as many as 75% of the
speakers. With regard to spoken registers, members of Group B declared high acceptability of all PVs in the spoken informal, except for rhedeg allan o which was rejected by the
majority of speakers. In the spoken formal, however, the same PV was accepted by 75%
of the informants, probably due to its use in a more idiomatic context (“running out of
steam”), while the acceptability of other PVs did not deviate significantly from the mean
score except for troi i fyny, accepted by only 25% of speakers.
In contrast, members of Group L declared high acceptability (70%-100%) of all
the PVs in all registers, except for edrych ar ôl and ffeindio allan. Ffeindio allan was
recognised as unacceptable by half of the speakers in R3 and over two thirds in R4. Both
verbs were also rejected by about half of the speakers in the spoken formal register. The
low acceptability of ffeindio allan can be attributed to its status of a loanblend associated
with borrowing from English.
The most noticeable difference between the two groups can been seen in comparing the levels of acceptability between the two written registers: members of Group B
were more tolerant of PVs in the official document than a popular book, as opposed to
Group L, who were more reluctant to use PVs in the official context. This may be partly
explained with the comments given by some of the employees of Canolfan Bedwyr, who
advocated the Cymraeg Clir approach, namely using simple, understandable language in
official communication with the public. In turn, Group L appeared to associate the official
register with more formal, standard language. The significant difference between the
groups in the acceptability of rhedeg allan o and ffeindio allan may point to a strong
prescriptive norm against such constructions among members of Canolfan Bedwyr staff.
However, what should also be noticed is that both groups almost universally accepted the
PVs in the spoken informal register.
The results of examining Questionnaire B scores against three extralinguistic variables showed that while sex and age did not seem to play a role in influencing the speakers’ judgment, the level of education might have been a significant factor, especially in
the written semi-formal register. In informal speech, however, levels of acceptability were
307
high regardless of the participants’ education. The comparison of Groups B and L suggests that employees of Canolfan Bedwyr much more aware of a norm against phrasal
verbs in writing and more sensitive to calquing phenomena.
5.4.4. Interview data
The following sections present the data obtained from the interviews accompanying the
questionnaire. The data are organised around major topics and themes which emerged in
the course of the conversations.
5.4.4.1. Acceptability and usefulness of borrowings
While the majority of the informants appeared to feel comfortable with the fact that they
used English vocabulary in their speech, some expressed concern about using too many
loanwords or, on the contrary, emphasised that they made a conscious effort to avoid
them.
A quarter of the speakers (14), most of them from groups A and B, manifested
a rather purist approach, saying that they try to avoid borrowings in general. From among
those, A1, A3, B5 and B6, four teachers of Welsh, stated that they were very conscious
about the use of borrowings and tried to avoid them as a rule. For example, A1 made a
decision after graduating from high school to eliminate English words from his vocabulary.
A1: beth nes i o’dd (.) jyst penderfynu yn fy meddwl i raddau reit dw i’n mynd i stopio bod
yn ddiog (...) yn hytrach na defnyddio gair Saesneg lle o’n i’n arfer gwneud yn yr ysgol (.)
iawn be ydy’r gair Cymraeg am (.) o hyn allan dw i’n mynd i ddefnyddio hwnna (.) a bob
tro dw i’n dweud rhywbeth Saesneg o’n i’n y n STOPIO be fydda yn Gymraeg sut i ddweud
yn Gymraeg yn gywir a (.) (...) dros tair blynedd (.) nes i lwyddo (1) i gael gwared o ran
fwya yno.
‘What I did was, just decide in my mind to some degree “right I’m going to stop being
lazy” (…) rather than using an English word as I used to do in school, “right, what is the
Welsh word for.. from now on I’m going to use that” (…) and every time I say something
English I STOPPED, what would it be in Welsh, how to say it in Welsh correctly and (…)
in three years, I managed … to get rid of most of them.’
308
Two speakers said that they would like to use fewer English words and borrowings. Similarly to A1, G11 associated using English vocabulary with a kind of “laziness”;
Y2 said that using too much English causes her to feel embarrassed with some people.
G11: llawer gormod. (...) a dwi ddim yn gwbod pam (...) be sydd yn dylanwadu ar hynny
achos (.) fel fi’n gweud bo fi’n siarad Cymraeg yn rhan fwyaf o’r amser ond (.) weithia mae
meddwl yn mynd yn diog a chi’n gweld y gair cynta sy’n dod i’r pen a mae hwnnw yn
weithia yn Saesneg (.) yn anffodus.
‘Much too many (...) and I don’t know why (...) what influences it because, as I say, I speak
Welsh most of the time but sometimes the mind goes lazy and you see the first word that
comes to your head and sometimes unfortunately it is in English.’
Y2: hoffwn i defnyddio llai o Saesneg a weithiau a: yn enwedig efo rhai pobol wi’n ymwybodol bo fi’’n defnyddio gormod o Saesneg t’mod bron yn embaras (..)
‘I’d like to use less English and sometimes, especially with some people, I am aware I use
too much English you know, rather embarrassing.’
Some speakers noticed that they had become more aware of the English element
in their speech over the course of time. For example, the bilingual school environment
and contact with L2 speakers made A4 realise that her Welsh is heavily influenced by
English. O6 made a similar discovery after a monthly trip to Vietnam, when she realised
that local people could pick up many English words in the Welsh she spoke with her
companions.
All of the interviewees indicated that the amount of English used varied depended
on the situation and the register they used. Although a question about adapting one’s language to the interlocutor was not asked by the researcher, nearly half of the speakers (21)
made this point themselves. “It depends with whom I speak” was by far the most common
reaction to the first question about borrowings.
Examples of adapting one’s language to the speaker were varied. Some informants
simply acknowledged that they imitated the habits of others. Speaker A4 noticed that her
Welsh has changed due to contact with L2 speakers.
A4: achos bod na blant o gefndiroedd Saesneg yn yr ysgol maen nhw yn amlwg yn defnyddio geirfa Saesneg. a wedyn dwi’n mynd adre (.) a ma’ mhartner i’n fath o dweud fath
o mae’n tynnu arna fi fath o bo fi’n dweud petha fel na fath o geiriau Saesneg dw i ddim
yn sylweddoli bo fi’n neud o yn aml iawn.
R: felly maen nhw’n dylanwadu [eich Cymraeg chi]
A4: [yndi o yndi] mae’r plant bod nhw â’r acen mor rhyfedd ond ydy (.) dych chi’n picio
pethau fyny (...)
309
‘A4: Because there are children from English backgrounds in the school, they obviously
use English vocabulary and then I go home and my partner sort of, says, sort of, he teases
me, sort of, that I say things like that, sort of English words sometimes I don’t realise I’m
doing it.
R: So they influence your Welsh.
A4: Yes, oh, yes, the children, that they have such a strange accent, but yes, you pick things
up.’
Adaptation may also mean purposefully simplifying one’s language while talking
with people who are less confident about their Welsh, such as learners (A2, A6, B1, B7,
B8, C7, O1, O2). This was mentioned primarily by teachers of adults:
C7: achos bo fi’n diwtor Cymraeg (…) mae Nghymraeg fi wedi symleiddio (.)
‘Because I’m a Welsh teacher (…) my Welsh has become simpler.’
O1: ond wedyn pan dw i ella yn siarad efo rhywun sydd ddim CWEIT mor hyderus (.) wrth
siarad Cymraeg dw i’n gwybod mod i’n defnyddio mwy o eiriau Saesneg (...) yng nghanol
brawddegau Cymraeg.
R: mhm er mwyn swnio’n fwy naturiol falle? neu =
O1: = naturiol [dw i yn]
R: [neu jyst] i roi hyder i bobol?
O1: hyder i bobol yn benna dw i’n meddwl ym ma’ pobol yn dweud yn aml iawn mod i’n
siarad fel athrawes Cymraeg (...) <imitate> o’n i’n meddwl mai <en> teacher </en> oeddech chi. <imitate>(...) felly mae defnyddio ychydig bach mwy o eiriau Saesneg dw i’n
meddwl yn gwneud pobol sy ddim yn hyderus i deimlo yn llai ymwybodol (.) o unrhyw
camgymeriadau sgynnon nhw yn y Gymraeg.
‘O1: And then when I maybe speak with someone who is not QUITE so confident while
speaking Welsh, I know I use more English words in the middle of Welsh sentences.
R: Mhm, in order to sound natural maybe, or...?
O1: Natural, I...
R: Or maybe make people confident?
O1: Confidence, mostly, I think, people very often say I speak like a teacher of Welsh. “I
thought you were a teacher”. So using a bit more English words I think makes people that
aren’t confident feel less aware of any mistakes they make in their Welsh.’
O2 believed that using some “Wenglish” might help learners to gain confidence and fluency.
O2: dw i’n defnyddio tipyn bach efo dysgwyr (1) o Saesneg neu Wenglish (...) os ydw i’n
canolbwyntio ar batrwm (.) a dw i ddim isio i nhw colli rhedeg neu rhediad y peth (.) dw
i’m taflu geiriau Saesneg i mewn (.) jyst i helpu efo’r dealltwriaeth
‘While speaking to learners I use a bit more… of English or Wenglish (...) If I focus on a
pattern I don’t want them to lose the flow of things, I throw in English words just to help
with understanding.’
310
B1 described the situation at Bangor University where, according to the speaker, many
native speakers are intimidated by the academic environment, being afraid that the Welsh
they use is not good enough. The speaker said he tried to simplify his language to give
them confidence.
B1: (.) be sy’n rhyfedd yn yr ardal yma ydy cymaint o siaradwyr Cymraeg mamiaith sydd
yn swil o defnyddio eu Cymraeg. (...) yn enwedig (.) gan bod nhw’n gweithio mewn prifysgol (.) maen nhw’n poeni (...) felly pan mae rhywun yn siarad efo pobol fel hynna mae
rhywun yn trio (.) addasu cywair yr iaith
R: mhm felly defnyddio mwy o eiriau Saesneg ?
B1: ie (.) wel efalle peidio defnyddio holl ystod yr eirfa Gymraeg sydd gen i felly.
‘B1: What is strange in this area is that there are so many native Welsh speakers who are
shy in using their Welsh (...) especially since they work at the university, they worry (...)
so if you speak with people like that you try to adapt the register of the language
R: Mhm, so using more English words?
B1: Yes, well, maybe not using the whole range of the Welsh vocabulary I have.’
A reverse situation was mentioned by other participants, who make their Welsh more
“pure” and sophisticated while talking to people who they believe have better command
of the language. These would be, for example, friends who received good education in
Welsh (A8), scholars (Y2), or in the case of a journalist, interviewees of the older generation whom she respects (G11).
Another factor influencing the use of accommodation techniques would be the
interlocutor’s education. Institutional rules may also play a role. B2, for example, describes a “hierarchy” of his friends and colleagues with respect to the kind of Welsh he
would speak.
B2: ella ‘swn i’n siarad efo (.) cydweithwyr yng Nghanolfan Bedwyr ella ‘swn i’n (...) fwy
ie parod i ddefnyddio geiriau Cymraeg lle os ydw i’n siarad efo (.) ffrindia coleg lle ‘swn
i’n defnyddio mymryn bach mwy o eiria Saesneg ond ffrindiau ysgol lle byddwn i’n defnyddio LOT mwy o eiriau Saesneg
‘Maybe if I talked with colleagues in Canolfan Bedwyr maybe I would be yeah ready maybe
to use Welsh words if I talk with friends from college I would maybe use a little bit more
of English words but with friends from school I would maybe use a LOT more of English
words’
Speaker B7 noticed that adapting to the perceived needs of other might be quite misleading as one may label people according to their social status. She quoted a situation of
using English vocabulary purposefully while talking with workers who, surprisingly for
the speaker, answered back with Welsh words.
311
B7: os dw i’n siarad efo rhywun tu allan i’r gwaith (.) bydda i’n ofalus i beidio defnyddio
(.) geiriau (.) sy’n ym (.) ddiarth oherwydd mae mae’n gallu (.) neud rhywun yn teimlo’n
chwithig (...) dw i di cael yn nal allan (.) wedi siarad efo (.) er enghraifft efo trydanwr neu
plymar adra (.) a defnyddio rhyw air Saesneg am feddwl na fydden nhw’n deall y gair
Cymraeg ac wedyn maen nhw’n defnyddio y gair Cymraeg cywir yn ôl [arna fi] (...) ie
ch’mod dyn ni’n weithia gallu bod yn nawddoglyd efo pobol
‘If I speak with someone outside work I’ll be careful not to use words which are strange,
because they can make someone feel awkward (....) I have been caught out having spoken
with for example an electrician or a plumber at home and using some English word thinking
they wouldn’t understand the Welsh word, and then they use the correct Welsh word back
on me (...) yes, you know, we can sometimes be patronising with people’
Finally, five speakers mentioned that they try to speak more correctly to their own
children. They indicated that having children made them more sensitive to their own vocabulary and that avoiding English vocabulary in their Welsh for the sake of children may
develop into a habit. Some of these speakers referred to PVs in particular:
O4: ers cael plant falle mod i’n defnyddio mwy o eiriau Cymraeg fel (.) e: peiriant golchi
peiriant sychu pan o’n i’n blentyn ‘swn i’n dweud <en> washing machine tumble dryer
hairdryer </en> ond dwi’n cofio ers cael plant dw i’n cofio TRIO defnyddio y geiriau Cymraeg. (...) a dw i wedi cario ymlaen i ddefnyddio geiriau Cymraeg.
‘Since having children maybe I’m using more Welsh words like… peiriant golchi [washing
machine], peiriant sychu [tumble dryer]. When I was a child I would say washing machine,
tumble dryer, hairdryer, but I remember since having children, I remember TRYING to
use the Welsh words (...) and I have carried on using Welsh words.’
G2:
wel (.) diffodd faswn i’n dweud (.) ond (1) rhywbeth mwy diweddar ydy hynny
(...) lle (.) rhai blynyddoedd yn ôl dw i’n siŵr (.) fyswn i’n dweud troi i ffwrdd neu troi
ymlaen (...) yn enwedig t’mod gyda plant (.) dw i’n meddwl (...) ‘chos bod isio i’r iaith bod
yn (1) fwy (.) fwy cywir (.) o bosib
‘Well, I would say diffodd [turn off] but... this is something more recent (...) while some
years ago I’m sure I would say troi i ffwrdd [turn off] or troi ymlaen [turn on] (...) especially
you know with children, I think (...) because you want the language to be… more, more
correct possibly’
Five other speakers paid attention to the fact that their own parents used to correct them
and that is why they are more sensitive to certain words.
O6: dw i wedi cael ffrae (.) gan mam dros y blynyddoedd am ddweud droi off.
‘I would have a row with mum over the years for saying troi off [turn off]’.
312
During the initial and follow-up interviews the majority of informants shared their
opinions about the place of English borrowings in Welsh. The prevailing attitude towards
them was positive. The most popular stance was that English loanwords are a natural
element of the language (15) and that some of them are well-established, sometimes very
old (8). Speakers viewed borrowings as something “inevitable, a part of the way everyone
speaks” (B2), something that comes along with bilingual upbringing and bilingual society
(B5). Fourteen interviewees emphasised that specific borrowings are an inherent part of
their idiolect. As stated by C5: “If you just talk from your heart... if you just talk naturally... English words just come out”. C6, a young speaker living in Cardiff, mentioned
playing a kind of game with herself once, when she tried to speak Welsh without using
any English words. The “pure” language she produced sounded awkward and it made her
feel somewhat guilty:
yn y car unwaith (.) hanner fel jôc (.) nath fi benderfynu baswn i ddim yn deud unrhyw
eiriau Saesneg (.) ac o’dd o’n rîli anodd (.) a pan o’n i’n siarad o’n i’n teimlo mod i’n
swnio’n annaturiol a cweit gwirion a ddylwn i ddim teimlo fel ‘na (.) ond fel ‘na o’dd o.
‘Once in a car, half as a joke, I decided I wouldn’t say any English word and it was really
difficult, and when I was speaking I felt that I sound unnatural and quite silly and I shouldn’t
feel like that...but so it was.’
The speaker seemed to be at loss why she used so much English vocabulary despite knowing Welsh vocabulary:
dw i’n trio defnyddio llai. a dw i yn gwbod (.) dw i yn gwbod y geiriau Cymraeg (.) jyst am
ryw reswm ma’ y geiriau benthyciada yn teimlo mwy naturiol.
‘I try to use fewer and I do know, I do know the Welsh words, it’s just that for some reason
the borrowed words feel more natural.’
Many informants suggested that speaking “purer” Welsh involves conscious effort
and can be quite difficult:
B4: ie mae dylanwad Saesneg mae’n anodd (2) (dwyn i gof) petha mwy Cymreig weithia
er bod nhw’n swnio’n lot gwell ‘lle. (...) a dim yn rhy anodd chwaith.
‘Yes, there is English influence, sometimes it’s difficult… to remember more Welsh things
although they sound much better maybe (...) and not too difficult either.’
313
The awareness of the foreign element in one’s language and an effort to speak “properly”
to the researcher could be observed in the way the participants self-monitored themselves
during the interview. It was not uncommon for them to notice borrowings in their speech
and correct themselves by supplying a Welsh word.
A8: dw i’n trio peidio ond dw i newydd dweud dw i newydd dweud trio [peidio] (...) mae’r
hwnna’n air Saesneg.
‘I try (trio) not to [use borrowings – ML] but I’ve just said I’ve just said trio not to... that’s
an English word.’
C5: [mae’n jyst digwydd]. so mae’n rhaid ichi meddwl am y peth (.) rili (.) y:m. (.) o ie (.)
dw i’n trio (.) o dw i jyst wedi gweud dw i’n trio (.) a t’mod gair ceisio (.) [ceisio yw’r gair
ffurfiol
‘It just happens so you have to thing about it, really, um, o yes, I try [trio]... oh, I’ve just
said trio, and you know ceisio, ceisio is the formal word.’
B4: ti ddim isio rhoi pobol (.) off fel dwi’n dweud @ <@>ddim eisio off<@>
‘You don’t want to put people off, as I say (laughter) don’t want off.’
Some speakers explained that, in certain contexts, English borrowings sound more natural
than Welsh coinages, especially with regard to modern terminology (A4, G5). Speakers
A5 and L3 referred to calquing English proper names such as Facebook and Power Point
(Gweplyfr, Pwerbwynt) and considered such translations rather unnecessary. L8 believed
that Welsh lacks certain specialist terms due to the fact it is “older than English” and
a “spoken” language. These deficiencies in vocabulary, in the speaker’s opinion, make it
more difficult to avoid English words
Another point in favour of using borrowings was made by journalists G2 and G9,
who believed that, since some Welsh words are unknown to general public or sound unnatural, using borrowings may facilitate communication with a wider audience. Thus the
speakers stated they would use loanwords on purpose:
G2: os ydw i’n sgwennu erthygl neu rywbeth (.) bydda i’n tueddu i beidio ysgrifennu
sgwennu’n rhy ffurfiol (.) ac yn or-ffurfiol (...) er enghraifft os dw i’n y:m (1)sgwennu am
gerddoriaeth y:m (.) a sôn am gig neu (.) y lluosog gig felly (.) yn well imi ddefnyddio
GIGS (.) yn hytrach na GIGIAU (...) mae na rei mae na rei geiriau (.) os na (.) dw i’n
meddwl (.) fydda pobol sy’n darllen yn deall y gair y: (.) y gair ffurfiol Gymraeg (.) felly
(.) bydda i’n tueddu defnyddio y:m gair Saesneg neu yn sicr gair sy’n sy’n deillio o o gair
Saesneg mewn ffordd.
314
‘If I write an article or something I’d rather not write too formal and “overformal” (...) for
example if I write about music and mention a gig or... so the plural of gig… I’d prefer to
use gigs rather than gigiau (...) there are some words, if I think people who read would not
understand the word, the formal Welsh word, I would tend to use an English word or surely
a word that comes from an English word in a way.’
G9: dw i’n trio osgoi y:m cystrawennau Saesneg (.) pan dw i’n sgwennu yn Gymraeg (.)
weithia fydda i’n defnyddio ambell air (.) Saesneg WEDI GYMREIGIO (...) yn fwriadol
yn enwedig os ydw i’n y:m anelu at gyrraedd cynulleidfa (.) hh hynny yw dw i’n trio defnyddio iaith faswn i’n tybio mae’r gynulleidfa yn gyfforddus efo hi
‘I try to avoid English constructions when I write in Welsh, sometimes I would use some
Wallicized English word (...) on purpose especially if I aim to reach an audience... that is I
try to use a language that I suppose the audience is familiar with.’
Another recurring theme was associating borrowings with one’s dialect. This was mentioned by 10 speakers (B2, B6, B8, G9 – from Caernarfon; G7 – Gwynedd; C5, G1, G4
– Carmarthenshire; L3, Y3 – border of Pembrokeshire and Ceredigion). Speakers from
Caernarfon in particular mentioned the large amount of English vocabulary in their dialect. Two speakers pointed out, however, that the Welsh used in Caernarfon is “naturally
Welsh” as it retains Welsh structures even though much of the vocabulary is borrowed:
B2: yndi (.) em yndi. se sech chi’n t’od i’n cerdded (i lawr) (.) ar y stryd yng Nghaernarfon
(...) felly bod na cymaint o eiriau benthyg (.) Saesneg yn y frawddeg (1) na t’od falle tasech
chi’n deud nid Cymraeg ydy o ond Cymraeg ydy o achos mae na gystrawen yn Gymraeg
‘There are, yes, there are if you, you know, walk down the street in Caernarfon (…) so there
are so many English borrowings in the sentence... you know maybe you’d say it isn’t Welsh
but it is Welsh because the structure is Welsh.’
B6: ces i fy magu yn rhannol yng Nghaernarfon (.) ac yn fan ‘ma mae pobol yn gollwng
geiriau Saesneg bron pob yr ail air yn aml iawn. ond mae patrymau @ <@>Cymraeg</@>
‘I was partly raised in Caernarfon and people there drop in English words almost every
second word very often. But the patterns are Welsh.’
Using borrowings might form part of one’s class identity as well, as mentioned by speaker
O3, who said she preferred to use borrowed vocabulary while talking to members of her
family who belong to the working class. She considered using loanwords natural in this
environment.
Altogether, information obtained in the interviews suggests that speakers tend to
prefer borrowings to native Welsh words which are not well-established in everyday language and/or which belong to the formal register. Consequently, speakers have difficulties using them or might be afraid of social rejection.
315
On the whole, participants made considerably few explicitly negative comments
on borrowings from English. Four speakers – all of them teaching Welsh to adults – gave
examples of loanwords which sound “ugly” (C7), “hurt the ear” (O1), “get under the skin”
(O6), they “don’t like them” (O3). These comments concerned only specific examples of
loanwords, however. Most speakers believed that borrowings from English have their
place in the Welsh language. Three participants noticed that borrowings of vocabulary,
and even calquing idioms, are not as problematic for the language as changes in structure.
This is consistent with B2’s previously mentioned comment about the Caernarfon dialect
– as long as the structure remains Welsh, the language is Welsh. The same view was
expressed by G8:
G8: tra bod ti’n wedi ymlacio a siarad bod dydd wyt ti YN iwso geirie Saesneg ond ar
ddiwedd y dydd ma’ (.) ma’ gwraidd yn Gymraeg
‘When you are relaxed and speak everyday you DO use English words but at the end of the
day the root is Welsh.’
Only a relatively small group expressed concern about overusing borrowings in
their speech. Sometimes borrowings were viewed as a valuable, distinguishing element
of one’s dialect, sometimes they were seen as useful with regard to specialist terminology.
Borrowings were also said to help in bridging the gap between native speakers and learners, making the latter more confident. On the other hand, some speakers stated that they
thought it important to keep their Welsh devoid of a strongly English element out of a
sense of responsibility for younger people, be it students or their own children. Another
factor that reinforced monitoring the purity of one’s vocabulary was pressure from the
professional environment, such as university. Above all, however, speakers drew attention to the necessity to adapt to needs and abilities of other people to ensure communication in a bilingual society.
5.4.4.2. Ideologies of purism
Since the participants generally considered English influence as natural in the bilingual
context, none of them outwardly supported ideologies of purism. Nevertheless, six in-
316
formants expressed the opinion that although changes in the Welsh language are inevitable, speakers of Welsh should do their best to keep the language “pure and idiomatic”.
Such a stance was often associated with the policy of the institution where the speakers
worked. For example, A6 said that unless the purity of Welsh is retained, the language
might be lost and it is the mission of the school to help young people learn to speak Welsh
as a natural, community language. Similarly, B1 and B6 stated that it is the mission of
Canolfan Bedwyr to help people speak better Welsh, namely a language in which “English patterns” are avoided. On the other hand, the same speakers admitted that it is extremely challenging to try and stop the growing influence of English, particularly among
young people, who develop the “bad habit” of speaking English with their peers. B6 noticed that the changes are profound, affecting the structure of the language and expressed
hope that the structure will not “change too much”.
A more common opinion expressed by speakers was that the most important thing
for the future of Welsh is that it is spoken, while its “correctness” and “purity” is less
crucial (10). Two participants illustrated this with a popular saying gwell Cymraeg slac
na Saesneg slic ‘slack Welsh is better than slick English’. This view was particularly
common among teachers from group A, who function in a bilingual school environment
where Welsh is the second language for many students. The teachers pointed out that
students should speak a language which they find natural (A1) and be confident in speaking Welsh even if it is imperfect (A6). A7 emphasised the need of young people to be
able to communicate and “have fun” in Welsh rather than worry about correctness. A8
referred to her experience of teaching French and Spanish to claim that confidence is the
crucial aspect in gaining the ability to communicate. A4 noticed that being able to communicate in Welsh is very important for maintaining it as a community language, and
stated that it is essential to have Welsh speakers working in areas such as health service.
The speaker acknowledged the need to write correctly, but believed it is not necessary for
communication in everyday life of the community:
(.) maen nhw’n meddwl bod o’n (.) bod angen bod yn IAWN bod yn GYWIR. dw i’n
meddwl os dyn nhw’n gallu siarad yn gymdeithasol (.) a jyst fath o ar lefel ar lefel cymdeithasol dw i’n meddwl dyna peth sy’n bwysig wrth mynd i byd gwaith yn yr ardal yn
enwedig achos (.) oce dw i’n meddwl mae na le lle mae nhw angen gallu ysgrifennu’n
gywir efalle ond (.) hynna sen nhw’n mynd i weithio lle bynnag maen nhw’n mynd i weithio
(...) felly os maen nhw’n gallu siarad yn Gymraeg a hynny’n doctoriaid nyrsys (.) mae’r
henoed yn enwedig yn licio clywed nhw’n siarad Cymraeg
317
‘They think there is a need to be RIGHT, to be CORRECT. I think that if they can speak
socially and just sort of on the social level I think this is what is important while entering
the professional sphere in the area because, ok, I think there is space where they need an
ability to write correctly maybe, but if they are going to work wherever they’re going to
work (…) so if they can speak Welsh as doctors, nurses… elderly people especially like to
hear them speak Welsh.’
Several members of Cymdeithas yr Iaith also saw purism as dangerous for maintaining
the language. For example, speaker C1 believed that puristic tendencies may discourage
Welsh speakers to share their language due to lack of confidence and that revitalisation
efforts should focus on other aspects of language maintenance.
‘swn i ddim yn dueddol i godi y fath yma o bwnc y dyddiau ‘ma (...) mae cwestiynau llawer
iawn mwy pwysig ynglŷn a FAINT o Gymraeg y:m (.) dyn ni’n defnyddio (.) pa mor HYDERUS dyn ni wrth (.) a (‘na fel) defnydd o Gymraeg mewn gwahanol feysydd (.) gwahanol
sefyllfaoedd (.) hwn ydy’r peth pwysig i mi swn i’n dweud.
‘I wouldn’t raise this type of subject these days (…) there are much more important questions regarding HOW MUCH Welsh we use, how CONFIDENT we are while… and then
like the use of Welsh in different fields, different situations, this is what is important to me
I would say.‘
Nine participants expressed the opinion that people should not be criticised for
using English words. Similarly, speakers C8 and L3 remarked that everyone is entitled to
use the language they like. L3 noticed that some people might consciously choose to
speak “Wenglish”. G8 believed that correcting children is the best way to make them
switch to English:
G8: cer ti i gywiro plant ifenc yma (...) maen nhw’n gweud stwffo (.) byddech chi’n cael
Saesneg (.) na (.) ti’n mynd i ladd e (.) yndwyt ti?
‘You go correcting these young children (...) they will say stuff you, you’ll get English, no,
you’re going to kill it [Welsh – ML], right?’
Upon mentioning the issue of language devoid of an English element, several
speakers talked about the so-called “language police” (plismon iaith, heddlu iaith),
a name commonly attributed in Wales to language purists. A vivid picture of such people
was drawn by C5:
C5: ni’n cael rhyw fath o heddlu’r iaith (...) achos maen nhw’n y: (.) gobeithio (.) mae’n
swnio’n yn ofnadw (.) ond gobeithio maen nhw’n marw mas nawr
R: mhm
318
C5: y:m mae na <en> sort of </en> rhai hen bobol sy’n meddwl galle pawb jyst yn siarad
Cymraeg pur heb dim gair o Saesneg o gwbl (.) jyst Cymraeg pur a mae ‘na pobol fel fi
sy’n sy’n siarad Cymraeg naturiol (.) e: defnyddio geiriau Saesneg (.) ni yw’r rhan o broblem achos ni’n newid yr iaith?
R: ie
C5: <en> so </en> (.) maen nhw’n gweud (1) pan o’n i’n gweud pethau fel o’n i’n trio
achub yr iaith Cymraeg (.) maen nhw’n gweud (.) wel pwy iaith dych chi’n achub? Dych
chi’n achub y iaith Cymraeg go iawn neu (.) dych chi’n achub (.) jyst unrhyw beth sy’n
swnio fel Cymraeg? <en> so </en> (.) ie (.) imi mae’n agwedd ofnadw achos (1) mae lot o
bobol gyda diffyg hyder (...) dylwn ni annog y bobol na i siarad Cymrâg (...)
‘C5: We have a kind of language police (…) because they, I hope...it sounds terrible but I
hope they are dying out by now…
R: Mhm.
C5: There are sort of some old people who think that everyone can just speak pure Welsh
without a word of English, just pure Welsh, and there are people like me who speak natural
Welsh, use English words. We are part of the problem, because we are changing the language?
R: Yeah.
C5: So, they say… when I was saying things like “I’m trying to save the Welsh language”,
they say, well, whose language are you saving? Are you saving the real Welsh or, are you
saving just anything that sounds like Welsh? So, for me this is a terrible attitude because…
there are many people who lack confidence (…) we should encourage these people to speak
Welsh (...)’
C5 comes from a bilingual family and speaks Welsh with a pronounced accent. He describes himself as “someone who speaks natural Welsh”, which in his view is a language
which contains English words. The purists whom he encountered, on the other hand, did
not consider his language to be “real” Welsh and criticised him for changing the language,
associating the English element with deterioration. The speaker feels the injustice of such
claims as someone actively engaged in language revitalisation efforts as a member of
Cymdeithas yr Iaith. He expresses his bitterness in the wish that purists, whom he identifies with the older generation, “die out” (incidentally, he uses a calqued PV marw mas).
Similarly, speaker B5 remarked that purism can be detrimental for learners, in that
it might discourage them from studying. The speaker referred to her experience as a
teacher in beginner courses of Welsh and told the anecdote of learners being criticised for
using colloquial forms of the verb such as r’on i ‘I was’ instead of the full literary form
of the verb yr oeddwn i.
Speaker Y3 recollected that purist attitudes were common among some students
of Welsh in her university days. A5 called herself a “language policeman”, but only in
correcting her own daughter rather than students at school. The speaker felt that this was
a natural thing to do, as it had been done by her own parents.
319
Five speakers, all of them teachers, mentioned that speaking Welsh devoid of borrowings and using less familiar Welsh words may cause people to view one as “posh”
(A1, O1, O3), “weird” (C5, O6), “speaking like a teacher”, “a snob” (O1). A1, who had
previously described his efforts to de-anglicise his Welsh, admitted that he tended to simplify his language and use some English vocabulary to avoid being seen as “posh”.
A1: dw i ddim isio dod draws yn yn BUR ac yn ac yn GYWIR (...) felly dwi’n fwy debygol
i symleiddio’r iaith a i ychwanegu ambell air Saesneg achos (.) mi fedri di gael rhai pobol
sy’n jyst yn mynd o wyt ti’n siarad Cymraeg <en> posh <en>
‘I don’t want to come across as PURE and CORRECT (...) so I’m more likely to simplify
the language and add an occasional English word because you can have some people who
just go “oh, you speak posh Welsh.”
C5 emphasised that it is important to teach people colloquial, dialectal vocabulary
in order to accommodate them in the local community.
C5: os dw i’n dysgu dosbarth yn Sir Gâr (.) ni’n trio dysgu y: iddyn nhw y tafodiaith? achos
os maen nhw’n mynd mas o ddosbarth a siarad i bobol ar y stryd (.) os maen maen nhw’n
siarad <en> like </en> rili gogleddol a rili ffurfiol bydd pobol yn jyst troi mas a gweud be?
(...) a edrych arnyn nhw fel maen nhw’n math o (.) <en> like weirdos so </en>(...)
‘If I’m teaching a class in Carmarthenshire, we try to teach them the dialect. Because if
they go out of the class and talk to people on the street if they talk like really North and
really formal people will just turn out and say “what”? (...) and look at them as if they were
sort of, like weirdos, so (...)
The same opinion was given by three teachers from group O. O1 pointed to the dangers
of teaching people non-integrated words, telling a story about a group of learners who
attended a Welsh course to re-learn the language because the variety they had been initially taught in language courses was too formal. As a result, native speakers were reluctant to speak with these learners because they sounded "perfect and posh".
O1: dw i’n swnio’n POSH. meddai un ddynes. a dw i ddim yn posh yn Saesneg felly dw i
isio siarad yr iaith sy’n adlewyrchu (.) pwy ydw i (.) y:m <whisper> roedd hi’n defnyddio
geiriau od </whisper> @
‘’I sound POSH’, said one woman. ‘And I am not posh in English so I want to speak a
language which reflects who I am’. (whisper) She used weird words (laughs).
On the other hand, some informants, mostly teachers, pointed out that it is the new speakers who can be a bit “purist” when it comes to the choice of vocabulary (5) in that they
320
feel disappointed to hear Welsh words similar to English. For example, O1 said that in
the courses she conducts students often dislike being taught the word licio ‘like’ which is
a centuries old borrowing, preferring hoffi, which they consider “real Welsh”. Speaker
B2 believed that lack of English vocabulary can be a distinguishing feature of a learner’s
language.
B2: ti’n gallu sbotio dysgwyr yn aml iawn oherwydd bod nhw’n siarad yn fwy pur a hefyd
‘chos (.) maen nhw’n defnyddio geiriau Cymraeg go iawn
‘You can spot a learner very often because they speak more purely and also because they
use real Welsh words.’
Following this line of thinking, L1 expressed the belief that the researcher’s language is
more “correct” than the language of native speakers.
L1: ar hyd y blynydde dw i wedi pigo’r hynny i fyny a dw i’n eu defnyddio nhw heb feddwl
lle mae na rhywun fel ti wedi dysgu Cymraeg a wedi dysgu y Gymraeg yn GYWIR, yn
hytrach na fenthygau ac dw i’n meddwl bod dy Gymraeg di yn fwy cywir
‘In the course of years I have picked them up [borrowings – ML] and I use them without
thinking while there is someone like you who has learned Welsh and learned Welsh CORRECTLY rather than borrowings and I think your Welsh is more correct’.
An important point was made by B6, who noticed that the purist tendencies among learners concern only loanwords from English and not calquing, which is often done unconsciously.
Overall, the attitude of the participants towards ideologies of purism was negative
as they were associated with criticism, discouragement and stigmatisation. Most of the
speakers who expressed their opinion on the subject believed that maintaining Welsh requires accommodating all Welsh speakers, including learners and those who choose to
speak “Wenglish”. This may mean adapting one’s language to the needs of the interlocutor and using more English vocabulary. Several speakers indicated that it is better to
correct other people in a more indirect way, for instance “by giving a good example”
(C1). Referring to school practice, many teachers said they usually corrected students by
repeating what they say using a Welsh word (A1, A2, A3, A5, A6, A7, A8, B6, C7). They
commented that such indirect correction should be done in a delicate, constructive or
joking manner.
321
5.4.4.3. The influence of English on Welsh vocabulary and structures
Conversations on the purity of Welsh and the presence of borrowings encouraged many
speakers to make general observations regarding the influence of English on Welsh and
the change that the language is currently undergoing. Eleven participants emphasised that
the influence of English on Welsh changes the language of ordinary speakers. They noticed that not only English vocabulary but also English patterns are overwhelming due to
the ubiquity of English-language media, in particular the Internet and social media.
B3: oherwydd mae grym y Saesneg (.) mor hollbresennol mae o ym mhob man. (...) a mae
o’n cynyddu wrth gwrs mae wedi cynyddu yn ofnadwy efo cyfryngau cymdeithasol (...) ac
efo’r we a datblygiada diweddar ynde.
‘Because the strength of English is so omnipresent, it is everywhere (…) and of course it is
increasing, it has increased terribly with social media (…) and with the Internet and recent
developments.‘
B1: er bod gynnon ni sianel teledu a’n sianel radio (...) maen nhw’n gwrando’n llawer mwy
ar batryma Saesneg ac yn clywed llawer iawn mwy o Saesneg felly jyst hynny ydy o. dim
unrhyw (.) YMDRECH i osgoi patrymau Cymraeg (.) ond jyst (.) yn cael ei (.) bentyrru
arnon ni o bob cyfeiriad.
‘Although we have a television and radio channel (...) they [Welsh speakers – ML] listen
much more to English patterns and hear much more English so it’s just that, no EFFORT
to avoid Welsh patterns but just, it is piled upon us from each direction.’
Education in English was mentioned as another factor contributing to the Anglicisation
of Welsh. For example, speaker O6 said that she has trouble communicating with a family
member who used many “annoying” calques from English after a period of studying at
a university in England.
Many speakers of group B, linguists and teachers, were sensitive to the widespread
phenomenon of calquing in Welsh. B1, who had taken part in advanced language skills
courses for many years, noticed that currently what the participants struggle with most is
the use of English patterns in their Welsh, while in the past the most problematic areas
were strictly grammatical, such as the use of mutations. Speakers B3 and B5 gave examples of preposition stranding in Welsh, an evident example of calquing English structures.
Speakers B6 and B7 believed that Welsh is changing syntactically, and that translating
English idioms is particularly prevalent. B8 noticed that calquing was very common in
written work of students due to the fact that the majority of sources they read is in English.
322
The speaker was able to quote many examples of structural calquing, such as preposition
stranding or double definite article in noun phrases. B2 noticed that calquing is widespread among inexperienced translators, as observed during examinations:
ti ddim yn gwbod be sy’n mynd trw’ feddwl yr @ creaduriaid @ bach yn yr arholiad (1)
dan bwysau ac ati ac yn aml iawn ella bod na gystrawen Saesneg neu ella jyst diffyg cystrawen @ yn gyfan gwbl weithiau (...)
‘You don’t know what is going through the minds of these poor creatures during the
exam… stressed and so on and very often there can be English structure or sometimes just
(laughter) total lack of structure (…)‘
Two experienced journalists, G9 and G11, noticed that written Welsh has undergone deep changes in the last few decades. G9 stated that the written language has become
much more colloquial and some native Welsh words had become strange to ordinary
speakers. G11 believed that many idioms were being lost as people think more and more
in English. While writing, the speaker often tries to think what her grandparents would
say and is worried about the next generations, considering that people of her age seem to
struggle with natural Welsh language. She also believed that losing natural idiomaticity
is damaging for the richness of cultural heritage:
G11: yr un peth â cholli alawon gwerin (...) mae rhywun yn poeni am yr AMRYWIAETH
hynod hynod arbennig sy di (1) bwydo’r iaith.
‘The same as losing folk melodies (...) you worry about the remarkably, remarkably special
VARIETY which has… fed the language.’
On the other hand, a number of speakers saw the English influence as a natural
part of language development (8). G1 and O1 pointed out that every language evolves
and is influenced by other languages, while A5 noticed that the influence is more overwhelming in the case of a minority language. Speaker O1 seemed to be in two minds,
accepting borrowing English vocabulary, but worrying about calquing and losing native
idioms:
O1: mae pob iaith yn benthyg (.) o ieithoedd eraill mae’n datblygu mae na eiriau newydd
(.) ma be sy’n iawn i mi o ran (.) licio e:m a geiriau tebyg felly (.) maen nhw’n dod yn eiriau
gwahanol i’m merch i (.) pan fydd hi yn tyfu (.) <@>i fyny</@>.
R: @@
O1: a’r hi’n hŷn mi fydd gynni hi ei ei chorpws bach ei hun byth [o eiriau} Saesneg
R: [ie}]
O1: sy’n iawn i ddefnyddio yn y Gymraeg.
323
R: mhm
O1: <whisper> felly dw i ddim yn gwybod. </whisper> mae’n bechod (.) ie. cyfieithu’n
slafaidd o’r Saesneg. a cholli ymadroddion a idiomau Cymraeg. mae hynny’n biti.
‘O1: Every language borrows from other languages, it develops, there are new words, what
is ok for me when it comes to licio and words like that, they will become different words
to my daughter when she grows (laughter) up.
R: (laughter)
O1: And when she is older she’ll have her own little corpus of English words.
R: Yeah.
O1: That are ok to use in Welsh.
R: Mhm.
O1: (whisper) So I don’t know. (loud) It’s a shame, yes, slavish translation from English,
and losing Welsh phrases and idioms. That’s a pity.’
Even though the participants were professional speakers, some of them admitted that they
struggle with the omnipresent influence of English in everyday life:
A4: er bo fi’n iaith gynta (.) fath o dw i wedi dysgu Cymraeg a adre o rieni bob dim. ond
dw i dal i feddwl o be di gair yn Gymraeg achos ma’ rhywun weithia. mae Saesneg yn dod
mor hawdd mae o gwmpas ni.
‘Even though I am first language sort of I learned Welsh at home from my parents and all,
but I still think, oh, what some word is in Welsh, because sometimes you... sometimes
English comes so easily, it surrounds us.’
A theme which emerged in several interviews was that the influence of English is so
strong it becomes the language in which Welsh speakers think. Some participants claimed
this about themselves: for example L5, a middle-aged native speaker, admitted that she
always thinks in English while speaking Welsh. In the interview, the speaker used a large
amount of English vocabulary and, outside the interview with the researcher, codeswitched to a great extent. Another native speaker, L8, said he found it difficult to establish which language is the “first” one he used in his mind as he thinks in both. He attributed his trouble with finding Welsh vocabulary to lack of readership.
L8: un o gwestiynau mae pawb yn gofyn ydy be ‘di’r iaith cynta sy’n dod i’ch pen (.) [pryd
dach chi’n meddwl?] (...) a dw i’n <en> probably </en> dweud mai Cymraeg ydy o (.) ond
bo’ fi yn defnyddio Saesneg lawer so weithiau dw i yn (.) mae o’n gallu mynd y ffordd arall
(.) ie (.) dw i’n (.) <en> probably </en> yn fwy aml ‘wrach (.) meddwl am y gair Cymraeg?
achos dw i’m yn (...) <en> probably </en> ddim yn darllen digon o Gymraeg
‘One of the questions everybody asks me is “which language is the first that comes into
your head when you think” (...) And I probably say it is Welsh but I am using English a lot
so sometimes I... it can go the other way, yes, I probably more often maybe think about the
Welsh word? Because I don’t (...) probably I don’t read enough Welsh.’
324
Two of the speakers who work as interpreters said that this work had made them
notice that some people evidently think in English when they speak Welsh due to the fact
it is exceedingly easy to translate their Welsh into English.
B2: y yn aml iawn dyn ni’n mynd i gyfarfodydd (2) a ti’n ffindio rhai pobol yn eitha
HAWDD i gyfieithu (.) pan maen nhw’n siarad Cymraeg ti’n gallu cyfieithu fo i’r Saesneg
(.) ar y pryd yn eitha hawdd. a ti’n sylweddoli pam (.) achos maen nhw yn eu pennau (.) yn
cyfieithu o’r Saesneg (.) i’r Gymraeg (.) ac yn cadw at gystrawen y frawddeg Saesneg (.)
ac felly mae’n hawdd iawn newid hwnna’n ôl (.) wrth gyfieithu
‘Very often I go to meetings... and I find some people quite EASY to translate, when they
speak Welsh you can interpret it into English quite easily. And you notice why – because
they translate from English into Welsh in their heads and hold to the English sentence structure and so it’s very easy to turn it back while interpreting’
G11: a dw i’n weithia (.) be ydw i i dechra neud (.) gwrando rhai adrodd y: gohebwyr ar
Radio Cymru ac fi’n cyfieithu ar y pryd (.) ti’n gwbod mewn cyfarfodydd (...) ar fy liwt fy
hun (1) a dw i wedi dechra cyfieithu nhw’n ôl i Saesneg (.) a maen nhw’n siarad yn UNION
(.) mae’r strwythur y frawddeg cystrawen yn swnio y geiriau I GYD (1) dim yn Saesneg
mae’n swnio’n fel Cymraeg DA (.) safonol (...) ond (.) mae pobol wedi dechra cystrawennu
(.) yn yn Saesneg.
R: mae pobol yn meddwl yn Saesneg te.
G11: ie.
‘And sometimes what I start doing is to listen to some reports on Radio Cymru and I interpret, you know during meetings as a freelancer... and I’ve started translating them back to
English, and they speak EXACTLY, the structure of the sentence, syntax, it sounds, ALL
the words... not in English, it sounds like GOOD, standard Welsh but people have started
structuring in English.
R: So people think in English.
G11: Yes.’
On the other hand, a group of speakers emphasised that they always thought in Welsh (8).
Speaker L6, who learned Welsh as a teenager, draws a sharp distinction between two
languages and is unable to switch easily from one to another.
All in all, a number of speakers expressed concern over the way Welsh is changing
under the influence of English. However, it should be noticed that none of the speakers
expressed negative views on English itself, but worried about losing the “native”, “natural”, “idiomatic” elements in Welsh. The next section will examine whether PVs were
viewed as part of this threat.
5.4.4.4. Speakers’ views on phrasal verbs in Welsh
325
As seen above, the phenomenon of structural calquing was rarely mentioned by the informants in the preliminary interviews due to the fact that they associated borrowing primarily with loanwords. While some speakers appeared to monitor themselves to avoid
English loanwords in the interview, they freely used calqued PVs. Only in the short extracts cited in this chapter one may find a number of idiomatic PVs, such as: picio/pigo
i fyny ‘pick up’, (A4, L1), dal allan ‘catch out’ (B7), cario ymlaen (O4), sbïo i fyny ‘look
up’ (C6), rhoi off ‘put off’ (B4), marw mas ‘die out’ (C5), dod draws ‘come across’ (A1).
Other examples of PVs used include: ysgrifennu i lawr ‘write down’ (B2), pigo lan ‘pick
up’ (G8, G10), edrych i fyny (L2), cadw i ffwrdd ‘keep away’ (L8), bod ymlaen ‘be on’
(G8, L10), dod allan ‘come out, be published’ (O1), taflu i mewn ‘throw in’ (O2), gweithio allan/mas ‘work out’ (O2, Y2), dod i fyny ‘come up’, ‘appear’ (O4), tynnu i lawr
‘pull down’, ‘make simpler’ (O6), mynd ymlaen ‘go on’ (Y1), ffonio lan ‘phone up’ (C2),
troi off ‘turn off’ (B4). This is unsurprising bearing in mind the results of Part B of the
Questionnaire where the acceptability of PVs in spoken informal register was very high
regardless of the PV type.
Accordingly, a large group of speakers (14) said that all PVs in the questionnaire
appear natural to them while other 8 participants believed that only some PVs are natural.
One speaker considered the constructions very common and found it surprising that they
can be classified as borrowings. Three other participants specified that PVs are natural in
speech (G6, Y2, O3) and three speakers associated PVs with dialect and slang (G4, G6,
G8).
Some speakers drew attention to the gap between the spoken and the written language. Y2 emphasised the difference between the two and claimed that it does not depend
so much on external standards but the fact that writing gives one more time to think.
Y2: dw i’n credu bodd hi’n ffordd eitha da o destio fe oherwydd mae’n neud ichi sylweddoli
ma’ da chi safonau gwahanol i’r ysgrifennu a’r siarad. dyw e dim cweit SAFONAU mae e
jyst mater o amser (...)
‘I believe that this [the questionnaire – ML] is quite a good way to test it because it makes
you realise you have different standards for writing and speaking. It’s not quite STANDARDS, it’s a matter of time (…)’
Similarly, for speaker G4, PVs were tokens of “lazy” Welsh spoken on everyday basis.
326
G4: (rhywbeth) DIOG yw e (.) ond mae’n jyst swno’n (.) y:m dyw e ddim yn swno (1) mae
fel (.) dw i’n trio meddwl am enghraifft o beth dw i wedi gweud (1) digon posib dw i wedi
gweud llwyth ohonyn nhw ond dw i ddim yn cofio (.) mae jyst mae bron fel rhan o (1) o’ch
sgwrs pob dydd. erbyn hyn (.)
‘It’s something LAZY but it just sounds, it doesn’t sound… it’s like.. I’m trying to think of
an example of what I’ve said… It’s quite possible I’ve said a load of them but I don’t
remember, it’s just almost like part of your everyday conversation by now.’
C6 said that although she would avoid PVs in formal writing, she would use them in
official communication, such as composing a leaflet about the health service, where there
is a need to be “clear” and use natural, widely-used expressions:
C6: (…) achos mae yn cael ei ddefnyddio lot mwy (.) a dw i’n meddwl i bobol sy ddim yn
darllen lot o (.) Gymraeg ella bod ymadroddion ‘ynna yn haws i’w deall.
‘(…) because they are used a lot more and I think to the people who don’t read much Welsh,
maybe such phrases will be easier to understand.’
While these participants emphasised the naturalness of PVs in Welsh, another
group of speakers highlighted their non-standardness and the influence of English on
them. Ten speakers considered PVs to be calques. Eight speakers stated that they tried to
avoid calqued PVs, while six other participants specified that PVs should be avoided in
standard formal writing. However, only two speakers described such constructions negatively as “incorrect” and four declared that some PVs sound wrong. O1 stressed, however,
that the examples in the questionnaire are well-established and it is mostly new calques
from English that she found unacceptable.
Some members of Group G commented on the presence of PVs in the press. Two
experienced journalists (G7, G9) said that the use of PVs in the press had increased over
the years as they appear in quotations from interviews. G11 showed a letter to the press,
in which a reader complained about the use of PVs as “horrible” calques from English.
The letter was kept on the board in the magazine’s office to remind the journalists to avoid
word-for-word translation.
Other six speakers expressed a view that some PVs are unnecessary. B3 paid attention to pleonasms such as torri i lawr, where the particle does not add any meaning,
but admitted that the PV is “well-rooted”. Speakers B1 and C7 believed that there is no
327
need to use calqued English idioms if there are simple and effective Welsh words to express the same concept. B1 referred to the concept of Cymraeg Clir, which teaches people
to use simple Welsh words rather than calqued idioms.
Three speakers reacted to seeing phrasal verbs by referring to the prescriptive rules
introduced in advanced language courses. B1 noticed that these are precisely the constructions he would teach people to avoid in Gloywi Iaith courses, while L8 recollected
being taught not to use PVs in such a course he had himself attended. B1 explained that
once people uncover the pattern of calquing they might decide to eliminate this type of
constructions from their language. Another teacher from Canolfan Bedwyr, B8, was more
sceptical in this respect, suggesting that the norm against some PVs introduced in the
courses does not have much effect on the way people speak. However, the speaker did
not find the phenomenon as something particularly threatening for the language.
B8: dw i’n meddwl bod hi’n bwysig bod ni’n cadw (.) y patryma a cystrawena naturiol yn
y Gymraeg (.) ond efo petha fel na (.) dydy hi ddim yn mhoeni fi cymaint (.) defnyddio
idioma fel na (...)
‘I think it is important that we keep the patterns and natural constructions in Welsh but with
things like that [phrasal verbs – ML] it doesn’t worry me that much, using such idioms (...)’
Two speakers, G9 and C1, talked about phrasal verbs of their own accord before
seeing them in the questionnaire. G9 mentioned the PV syrthio allan ‘fall out’, ‘argue’ as
an example of a structure he would correct while proofreading other people’s texts.
C1 mentioned calqued PVs with up as an example of constructions he was “proud of”
avoiding in writing:
C1: mae na lot o idiomau Saesneg yn defnyddio’r gair <en> up </en> (...) t’mod (.) y:m
<en> tidying up (.) washing up (...) growing up (.) </en> (.) y:m a wedyn m- m- mae rhai
o hynny ar lafar yn tueddol o gyfieithu eu hunan’n llythrennol (.) mae mae patrymau Saesneg yn effeithio ar y Gymraeg. ond yn y gwaith ysgrifenedig dw i’n cymryd rhywfaint o
falchder yn osgoi y math yna o (.) droada lle mae natur amlwg Saesneg ar frawddeg.
‘There are many English idioms using the word up, you know, tidying up, washing up,
growing up and then there’s a tendency to translate those literally, English patterns influence Welsh, but in writing I take some pride of the fact I avoid this kind of translations
where the nature of the sentence is obviously English.’
This example illustrates that some of the informants were highly sensitive to word-forword translation from English. A3, a teacher of Welsh, said in the follow-up interview
328
that she had often thought about the constructions in the questionnaire as obvious tokens
of the widespread phenomenon of calquing that may result in losing Welsh vocabulary:
A3: o mae pethau ydw i’n meddwl amdanyn drwy’r amser @@ (...)
R: pam? @
A3: oherwydd (.) y:m dw i yn teimlo dan ni’n cyfieithu’n llythrennol o’r Saesneg DRWY’R
AMSER (...) a dan ni’n neud o ar lafar heb inni sylwi bod ni yn ei wneud o. y:m ond ‘chod
dan ni’n mynd i golli’r geiria (.) dw i’n meddwl (.)
‘A3: Oh, these are things I’m thinking about all the time (laughter) (...)
R: Why?
A3: Because I feel that we translate from English ALL THE TIME (...) and we do it in
speech without realising but, you know, we are going to lose words, I think.’
Noticeably, A3 says “we”, admitting that she used the calqued constructions herself. Similarly, other speakers who viewed PVs as calques also indicated that despite the existence
of a prescriptive norm, they do not adhere to it every time, particularly in speech. For
instance, B8 that she would use a PV torri i lawr herself even though she would encourage
her students not to do it.
Moreover, six participants (A6, A8, B4, B6, C8, G9) pointed out that some PVs
are difficult to replace. They referred especially to the case of troi ymlaen ‘switch on’
whose one-word equivalents such as tanio ‘ignite’ or galluogi ‘enable’ sound awkward.
Finally, four speakers (B3, C1, G2, G9) declared that the questionnaire made them
realise they were inconsistent in what is acceptable and what is not:
B3: a dydi’r linyn ddim yn gyson bob tro chwaith dw i’n teimlo mae rhywun yn hwyrach
yn defnyddio rhyw ymadrodd a ryw air (1) mewn un sefyllfa neu ar un adag (.) a (.) ddim
yn ei ddefnyddio yr adag arall mae’r (.) dw i’n teimlo bod fi’n beth bynnag yn reit (.) yn
reit ANWADAL i ddeud a (mympwyol) weithiau e e yn fy nefnydd.
‘The borderline is not consistent every time either, I feel you sometimes maybe use some
expression and some word… in one situation or at one time… and not use it in another,
it’s… I feel that at least I’m pretty, pretty FICKLE to say, and sometimes capricious in my
usage.’
C1: dw i’n licio meddwl mod i’n trio osgoi rhai ymadroddion Saesneg (.) ond ond neud imi
gwestiynu (.) pa idiomau yn union ydy hynny (.) mae’n bosib bod’na rai (.) llai amlwg imi
efallai y sy’n rhan o ddull ysgrifenedig (.) dw i ddim wedi meddwl amdano cymaint (.)
mae’n diddorol.
‘I like to think that I try to avoid some English phrases but this made me wonder which
idioms are those exactly, it’s possible that there are some less obvious for me maybe that
are a part of the written style, I haven’t thought about this that much, it’s interesting.’
329
To summarise, the informants who commented on Welsh generally agreed that they were
a natural part of the spoken language. Little more than half of the speakers felt that at
least some PVs are constructions calqued from English, but very few viewed it as a negative thing. Several informants referred to the norm prescribing the use of PVs, which
they encountered in proficiency courses, but some of them indicated such prescriptions
are somewhat ineffective.
5.4.4.5. Views on standard Welsh
In the follow-up interviews, a number of speakers were willing to engage in a conversation about standards in Welsh and attempted to identify sources of the linguistic norm. In
discussing standards, many informants stressed the difference between spoken and written language. The notion of standard was generally associated with writing, with 13
speakers saying that standards should be maintained in written language. Two speakers
believed that there is no need for a spoken standard because Welsh people understand
each other (L5) and one should maintain the variation in dialects (G1). C8 emphasised
that the need to maintain standard Welsh is a responsibility of the intellectual elite, who
function in institutionalised environment when norms can be retained. In the case of “ordinary” speakers, however, it is impossible to exercise any pressure to speak more formal
Welsh. Similarly, C2 stated that the majority of Welsh speakers are entirely unfamiliar
with the concept of standard Welsh because they do not use it in their professional lives.
Many speakers identified the standard with the language of the media. Eight informants expressed the opinion that broadcast Welsh is important for developing standards and that it is the mission of the Welsh media to give good example of language use.
Speakers also evaluated the broadcast language they heard. Five speakers, most of them
teachers from Group A, said that standards are generally high and the media manage to
keep the right balance between formal and informal language. However the informants
also shared some critical remarks. For example, A8 believed that the language of some
radio speakers is not natural enough due to the fact they make too much effort to use only
Welsh vocabulary.
330
A8: dw i’n mwynhau gwrando arnyn nhw (.) ond weithiau dw i’n teimlo bod un neu ddau
ohonyn nhw (.) mae nhw’n (.) maen nhw’n trio mor galed yn yn defnyddio Cymraeg geiriau
Cymraeg a dydy o ddim yn swnio’n naturiol iawn weithiau (…)
‘I enjoy listening to them but sometimes I feel that one or two of them, they try too hard to
use Welsh, Welsh words and they sometimes don’t sound very natural (…)’
Similar points were raised by B2, B6 and O6. B2 criticised radio journalists who read preprepared text, which sounded too formal and unnatural. O6 said that the media should use
more natural spoken language, except for news items and programmes about poetry,
which are formal in nature. The speaker added, however, that in her opinion the media
were already making attempts to make the language more colloquial. According to B6
the language of television news was too difficult and for that reason discouraging for
many people. The speaker believed that the situation could improve if there were more
than one channel, claiming that when two Welsh-language news programmes existed in
the past, there was more variation in the style they used.
On the other hand, several speakers believed that media standards were too lax as
regards formality and correctness (6). For A1 and O3 this concerned especially online
news portals. B7 complained about excessive colloquialisms and careless translations
from English in Welsh press. Having a member of the family working in a Welsh magazine, the speaker explained that the deteriorating standards are caused by a hurry in preparing articles and lack of supervision over young people who are strongly influenced by
English.
maen nhw’n gweithio dan bwysa (.) bwysa amser wedyn (.) ychydig iawn o bobol lot o
waith (.) ychydig iawn o amser. mae pethau fel na yn llithro i fewn (.) a mae lot o bobol
sy’n gweithio yn y cyfryngau yn bobol iau (.) a wedyn mae gynnon nhw dw i’n meddwl (.)
mae mwy o ddylanwad Saesneg ar eu hiaith nhw.
‘They work under pressure, the pressure of time, and then very few people, a lot of work,
very little time. Things like that [calqued idioms – ML] slip in, and lots of people working
in the media are young people and then they have, I think, their language is more influenced
by English.’
Similarly, B8 claimed that the younger generation of journalists working in Radio
Cymru seem to have problems with using natural Welsh, including structures and correct
mutations. The station was also mentioned by O4, who believed that for some journalists
the wish to appeal to popular taste takes precedence over correctness – as an example she
331
quoted a popular programme which had been using an incorrect phrase as a running joke,
which in the speakers opinion might spread the mistake. The speaker said that the problem
of calquing and incorrectness was to be seen in other Welsh media as well. She believed
that the media should strive to find a balance between making the audience confident and
maintaining high standards
Five speakers pointed to the impact of social media on changing the standards due
to the fact that people become accustomed to writing down spoken language and mixing
Welsh and English. On the other hand, one of the young participants, G5, said that she
used only Welsh in social media with her close circle of friends. This makes her feel that
she lives her life entirely in Welsh and it gives her more confidence in using it.
Another domain that the speakers associated with standard Welsh was the language found in official documents. The comments on that subject illustrated some burning issues in that respect. Four speakers said that many Welsh people have trouble understanding official Welsh. Speaker A5 admitted having problems with understanding some
documents herself, despite having a degree in Welsh; she also quoted examples of official
forms issued by the local council of Gwynedd, full of errors and “slavish” translations
from English. O5 also mentioned difficulties in using Welsh forms due to lack of familiarly with the “jargon”. Speakers O2, A7 and C3 admitted that they never used official
forms in Welsh as they were too difficult. Other speakers indicated that official Welsh
was too formal (Y3, O4, O6), “posh” (L10) and “stiff” (A5).
Over one fifth of the speakers highlighted that official Welsh should be simple
and clear (14). In some cases, this statement referred to the questionnaire as the speakers
justified why they would accept some PVs in an official leaflet but not in a book. In many
cases they referred to the Cymraeg Clir approach.
Several speakers mentioned the role of literature in maintaining standard Welsh.
All the three interviewed writers (Group Y) said they believed the language of novels
should reflect the reality and the vocabulary people use. At the same time, Y2 and Y3
believed that the narrative parts of the novels should be written in standard language with
fewer English elements. Speaker B7 commented on fiction from a reader’s perspective,
claiming that there was a growing tendency in novels to use colloquial idioms even in the
narrative parts. In the speakers’ opinion such a technique was unnecessary and made
books difficult to understand for speakers from different regions of Wales; she also believed that since written language never reflects the way people speak in reality, writers
332
should attempt to find the balance between sounding “real” and maintaining literary
standards. Characteristically, only one speaker (A1) referred to the most traditional literary norm, mentioning the Welsh Bible as the source of “the best” Welsh.
For two speakers the best standard Welsh was the natural idiomatic language used
by public persons (A8, G11). Journalist G11 listed several names of her linguistic “role
models” and described the language they used as “rich”, “robust” and difficult to translate
in that it does not copy English constructions.
Education was another domain the participants associated with maintaining standards. For speakers B7 and C8 the educational system was the most important channel,
alongside the media, of introducing standard language. B7 expressed hope that there
would be more Welsh-medium schools which might improve the language of young people. Speaker C9 complained about the level of education and the training of teachers in
his area, expressing worry about the future of Welsh. The participants who were teachers
themselves, particularly from Group A stressed their responsibility for introducing standard Welsh to the younger generation. Teachers of Welsh in particular (A1 and A3) said
that they saw themselves as models of good language and described their attempts to
sensitise students to differences between spoken and written idiom. This may be a difficult task, however, as students tend to be entirely unfamiliar with higher registers of the
language. A3 noticed that young people are often afraid to speak standard Welsh and
sound “like a book”. She described a situation in which she tried to encourage students to
use a literary form es i ‘I went’ by suggesting a colloquial pronunciation “esh i”, rather
than using dialectal aru fi fynd ‘I went’.
A3: maen nhw’n dweud <imitate.> ie mae’n iawn i chi miss <imitate> dych chi’n siarad
am tecstio a bethe a nhw’n dweud <imitate> ie swn ni’n ddim dweud hynny pan dyn ni’n
siarad efo’n gilydd <imitate> a nes i ddeud reit trïwch pan ydych chi’n siarad efo’i gilydd
yn lle mynd <imitate> o aru fi fynd <imitate> trïwch ddweud o ES I I. <imitate> ond dyn
ni ddim isio swnio fel llyfr <imitate> ES I I. dych chi ddim gorfod dweud na fel llyfr trïwch
i ddweud o <pvc> ESH I I </pvc> siopa
‘They say “yeah, it’s fine for you, miss”, you talk about texting and so on, and they say
“yeah, we wouldn’t say that when we talk to each other” and I said right, try, when you talk
with each other, instead of going aru fi fynd, try saying oh ES I I [I went (to)]. “But we
don’t want to sound like a book”. ES I I, you don’t have to say it like a book, try to say
“ESH I I” shopping.
333
A6 stated that despite all the problems with persuading the younger generation to use
“good” Welsh (or any Welsh at all), the education system should continue trying to maintain Welsh as young people are really proud of it as part of their Welsh identity.
A6: mae’n RHAID inni gario ymlaen i i gwffio (.) mae rhaid iddy- (.) MAEN nhw’n falch
o’u hiaith (...)dyna beth sydd yn od maen nhw’n falch bod nhw’n Gymry maen nhw’n falch
o’u hiaith (1) ond wedyn ma’ nhw mewn (1) <en>habit</en> ydy o <en>habit</en> drwg
(.) sydd yn gair Saesneg (...) A6: <en>habit</en> drwg o (.) defnyddio Saesneg efo ffrindia.
(...)ond allwn ni’n allwn ni’n (.) stopio. (...)mae rhaid inni gario ymlaen.
‘We HAVE TO carry on fighting, they have to… they ARE proud of their language (...)
this is what is strange, they are proud to be Welsh, they’re proud of their language, but then
they have a… habit, it is a bad habit, which is an English word, of using English with
friends (…) but we can’t, we can’t stop (…) we have to carry on.’
On a more optimistic note, several teachers defended students, stating that they are able
to speak “good” Welsh if only they make an effort (A2, A3, A5). B7 and A6 also noticed
that careless calquing and code-switching is a natural thing for teenagers and it had been
done in their own time alike. B5 claimed that the problem of “bad” language did not
concern only Welsh, because teenagers may also struggle with speaking correct English.
Deterioration of standards regarding language correctness was another theme
which emerged in several conversations. Participants gave different reasons for what they
saw as falling standards, such as lack of readership (B6, C9):
B6: mae na lot o anwybodaeth am beth sydd yn gywir (...) OHERWYDD (...) dydyn nhw
ddim yn darllen (.) dim yn t’mod achos dyna’r ffordd (.) naturiol ora o weld o ddysgu’r
Cymraeg o safbwynt cywirdeb.
‘There is a lot of ignorance about what is correct (,,,,) BECAUSE (...) they don’t read, you
know, because this is the best natural way of seeing, of learning Welsh as far as correctness
goes.’
C9: hefyd mae o’n diffyg (1) darllen (...) ma’ ‘na wahaniaeth rhwng darllen (.) ar gyfer
arholiad (.) a darllen cymdeithasol (.) darllen adref. yn bendant mae’na ddiffyg o hynny
hefyd.
‘There is also lack of reading, there is a difference between reading for exams and social
reading, reading at home. Definitely, there is a lack of that too.’
334
Other speakers said that standards in Welsh deteriorate due to the fact that people do not
have enough contact with natural language (6). For B2, one reason for this was the decline
in chapel attendance, where people used to be exposed to standard Welsh:
B2: mae’r oes lle oedd (2) pobol yn gallu dibynnu ar (1) cystrawenna mae nhw wedi dysgu
i co (.) mewn capeli ac yn y blaen wedi mynd (...) pan o’n i’n (.) blentyn deg mlynedd ar
hugain yn ôl (.) dw i’n siŵr bod na llawer iawn mwy o bobol yn dod ar draws i bobol hŷn
wedi mynd i gapel ac roedd eu Cymraeg yn fwy graenus ac yn [fwy cadarn] na (...) mae
hyn yn cael ei golli a does dim byd yn dod yn lle yr agwedd (.) yna oedd yn rhoi addysg
neu’n dysgu pobol ynglŷn â chystrawenna (.) Cymreig t’od (.) y patryma Cymreig da t’mod
(...) dw i’m yn gwybod be dy’r ateb (.) ond hyn ydy mae’r (.) ma’r patryma DA (.) yn fwy
prin. (.)
‘The age when... people could depend on… the constructions they had learnt by heart in
chapels and so on is gone (…) when I was a child thirty years ago, I’m sure there were a
lot more people who came across older people who had gone to the chapel and their Welsh
was more polished and stronger than… (…) this is lost and nothing comes to replace that
attitude which would give education or teach people about Welsh construction, you know,
patterns of good Welsh (…) I don’t know what the answer is but the thing is the GOOD
patterns are rarer.’
Speaker G11 stated that it is difficult to write good Welsh because there is so little
of it around, especially on the Internet. The speaker also believed that receiving education
in Welsh is not enough to provide people with the ability to speak good Welsh if they do
not have enough contact with proficient speakers, particularly from the older generation.
Another problem which the informant identified was the excessive translation from English and insufficient use of Welsh in the government bodies:
G11: [y problem yw] bod fi’n meddwl (...) bod na lawer gormod o gyfieithu yn cael ei
wneud yng Nghymru. (...)yn hytrach na annog (1) e:m (.) pobol i gwella a defnyddio eu
Cymraeg yn fwy. mewn awdurdoda lleol mewn (...) cynghorau lleol (...)mae’r system
gyfieithu mor gryf mae cymaint o gyfieithu mae p pethe yn (.) cael eu sgwennu yn Saesneg
a s’dim angen iddyn nhw.
‘The problem is I think that there is too much translation done in Wales (...) rather than
urging people to improve and use their Welsh more, in the local authorities, in the local
council, the translation system is so strong, things are written in English and they don’t
need to be.’
Lack of prestige associated with political power was also mentioned by C1, who
believed that using more Welsh in the National Assembly would help to create a new
standard.
335
C1: dw i’n meddwl bod YN drueni mawr t’mod fod na ddim o ddefnydd o Gymraeg yn y
Cynulliad er enghraifft. (...) achos mae hwnna’n (.) y:m (.) (mwy na’r) math o sefydliad y
bysa YN creu ac ac yn safoni defnydd o’r Gymraeg ella mewn ffordd sydd yn esblygu (.)
ella mewn ffordd sydd ddim yn draddodiadol (.) mi fasa fo yn creu y:m y:m ac yn gwneud
pobol yn fwy gyfarwydd efo efo efo’r Gymraeg mewn cyd-destun deddfwriaeth basa
hwnna (.) t’m- helpu i bobol ddod yn fwy cyfarwydd efo cywair mwy (.) t’mod safonol (.)
o’r iaith
‘I think this IS a great pity, you know, that Welsh is not used for example in the Assembly
(…) because that is a type of institution that WOULD create and standardise the use of
Welsh maybe in a way that evolves, maybe in a way that is not traditional, that would
create, um, and make people more familiar with Welsh in the context of legislation, it would
help people to become more familiar with the more, you know, formal register of the language.’
5.5. Summary and discussion of the results
The study aimed to investigate the homogeneity of the linguistic norm in Welsh among
professional speakers by examining the level of integration of the selected PVs and the
presence of monolingual-oriented ideologies in the speakers’ judgements.
The first aim of the study was to measure the acceptability of selected PVs in
semi-formal registers of Welsh against alternative expressions. The analysis of 12 onomasiological profiles in Part A of the questionnaire confirmed Hypothesis 1 stating that,
given a choice between PVs and their native equivalents, the participants preferred nonPV constructions, with one exception of bwrw ymlaen ‘carry on’. The findings of the
questionnaire have demonstrated that, at least in the investigated cases of frequently occurring PVs, they are not likely to replace alternative native items in written language. It
can be seen that, in the investigated cases, the vast majority of items coexisted with native
equivalents in the minds of the speakers. Insertion pertained mostly only to the case of
troi ymlaen, which is difficult to express by other phrases. The only non-PV item that has
been seen to fall into disuse was cael crap ar ‘get a grasp on’. However, the low acceptability of this item did not stem from the fact that it competed with the PV pigo i fyny
‘pick up’, but rather from the homogeneity of the impolite English word crap.
The second aim concerned the integration of idiomatic PVs in the standard written
language in comparison with the spoken mode. The results of Part B of the questionnaire
and data from interviews have confirmed Hypothesis 2 that the selected PVs were con-
336
siderably well-integrated in informal spoken Welsh and seen by the informants as an entirely natural part of their spoken idiom. A fairly large number of idiomatic PVs occurred
naturally in the participants’ speech, despite the fact that most of them used a rather formal register of Welsh in the situation of a recorded interview, avoided code-switching
and monitored themselves with regard to using English loanwords. This observation
agreed with the results of the questionnaire, where a high proportion of the speakers declared they accepted the use of PVs also in formal spoken register. This indicates congruity between speakers’ declared and real lexical choices.
The study results have also confirmed the hypothesis that differences in the degree
of acceptability would be associated with the status of a PV with regard to contact with
English. The high acceptability of bwrw ymlaen which is, to all probability, a native
Welsh PV, as well as dod ar draws ‘come across’ and edrych ymlaen at ‘look forward
to’, which also might have emerged naturally in Welsh rather than be directly borrowed
from English (3.3.2.3, 4.5.5), leads to the conclusion that the small number of native PVs
in Welsh are well-established in the written language. In contrast, as was expected,
loanblends such as cario ymlaen, ffeindio allan, troi off, whose form evokes the English
influence, were rejected in the semi-formal written register by the vast majority of speakers. The degree of acceptability of pleonastic and directly translatable calques varied considerably and appeared to depend on a number of linguistic and extralinguistic factors.
The final aim of the study was to identify some possible motivations for accepting
or rejecting phrasal verbs and tie them to standardisation-related ideologies. As had been
stated in Hypothesis 3, the results demonstrated a lack of universally accepted norms and
a variety of approaches to the issues of language contact.
The analysis has identified several possible motivations for rejecting a PV which
combined to form the speakers’ judgement:
-
perceived effectiveness of non-PV equivalents;
-
perceived degree of formality of a PV;
-
perceived English origin of a PV;
-
knowledge about a rule prescribing the use of a PV.
The first and probably most important motivation for rejecting PVs was the exist-
ence of another Welsh phrase which could express the same concept and its perceived
effectiveness related to qualities such as simplicity, naturalness and intelligibility. For
instance, simple verbs dysgu ‘learn’, and digwydd ‘happen’, corresponding to ffeindio
337
allan ‘find out’ and mynd ymlaen ‘go on’, were chosen by a higher proportion of speakers
than the more formal and sophisticated phrase disgwyl yn eiddgar ‘ardently expect’ corresponding to edrych ymlaen at ‘look forward to’. The most frequently rejected non-PV
item was the rather obsolete idiom cael crap ar ‘get a grasp on’, presumably falling out
of use due to its connotation with the English word crap. Similarly, the participants preferred to dispose of the particle in the case of pleonastic eistedd i lawr ‘sit down’ and tyfu
i fyny ‘grow up’, for the sake of linguistic economy and simplicity. There is also an interesting comparison to be made between two PVs expressing the concept of switching mechanical devices “on” and “off”: troi ymlaen and troi i ffwrdd (off/bant). While the speakers tended to acknowledge the need to accept troi ymlaen ‘turn on’ for lack of a better
expression, troi i ffwrdd ‘turn off’ was largely rejected due to the fact that there exists a
simple and well-established equivalent diffodd ‘turn off’, ‘extinguish’. A similar phenomenon was observed with troi i fyny, used in two different senses, where the more technical
and transparent sense of “turning up the heat” was more accepted in the written mode
than the idiomatic sense of “coming, appearing”, which can be expressed in Welsh by a
number of non-PVs. The observation that certain PVs may be considered unnecessary in
view of the existence of an effective alternative item corresponds to the findings of Chapter 4 concerning the choices of lexicographers in excluding some verb-particle constructions when another equivalent can be proposed.
Secondly, both the interviews and questionnaires showed that the informants were
highly aware of differences between the formal and informal registers of Welsh. For this
reason some PVs, such as rhedeg allan o ‘run out of’, gweithio allan ‘work out’, troi
i fyny ‘turn up’, ‘arrive’ were considered inappropriate in semi-formal contexts. This is
unsurprising, given the similar stylistic properties of their English counterparts. On the
other hand, some of the proposed Welsh equivalents, such as dychwelyd ‘return’, were
occasionally rejected due to their excessive formality. What is more, the speakers universally rejected dialectal forms of PVs (gweithio mas, pigo lan) and the PV troi off with the
stylistically marked particle directly borrowed from English.
The third motivation expressed directly by some informants was the fact that a PV
was a calque from English. Around half of the speakers expressed awareness regarding
the influence of English on Welsh PVs. However, very few speakers perceived the calquing of PVs as something to be avoided as a general rule, but rejected mostly the less well-
338
established items in more formal registers. Several speakers acknowledged their inconsistency as to which borrowings they found appropriate.
Finally, participants occasionally declared that the rejection of PVs was motivated
by prescriptive rules, which the speakers encountered at language courses or that had been
taught to them by members of older generation e.g. parents.
With regard to extralinguistic factors, the results of Part B were analysed against
three variables: age, sex and level of education. No significant differences between men
and women or between different age groups were confirmed in the statistical analysis.
The latter finding does not confirm Hirata’s (2012) suggestion that age might play a role
in the acceptability of calques.
However, the statistical analysis showed that education is a factor that might significantly influence speakers’ lexical choices. It seems logical to assume that highly educated speakers are more likely to have been taught about prescriptive rules or sensitised
to the phenomenon of calquing during their education process. On the other hand, no
significant differences in acceptability between speakers with secondary or higher education were found in the spoken formal or written official registers. A possible reason for
that emerging from interviews with the participants is the belief that these registers should
be closer to the informal spoken language given the need to communicate with the public
in a simple and effective way.
Other than education, there are certainly numerous other factors which empower
prescriptive ideologies among speakers. One of them might be family background. Data
from interviews has indicated that linguistic rules are transmitted across generations: several speakers recollected their own parents correcting them, while other speakers mentioned monitoring the way their children speak. Moreover, a lack of contact with older
generations was mentioned as one of the factors behind the perceived deterioration of
standards observed by some of the participants.
One more aspect that needs to be considered are institutional norms in a given
workplace. Following Eckert and McConnell-Ginet (1992), workplaces can be described
as communities of practice that is aggregates of people “who come together around mutual engagement in some common endeavour” (Eckert and McConnell-Ginet 1992: 95)85.
It could be argued that at least one of the investigated workplaces, Canolfan Bedwyr,
85
The notion of community of practice is derived from an older concept of a speech community, understood
as a group of people who share the same norms in regard to language (Labov 1972: 158).
339
could be viewed as a community of practice sharing some norms regarding the avoidance
of English influence on Welsh. As this institution is responsible for a variety of codification and standardisation projects in fields such as lexicography, translation and teaching
it is unsurprising that the centre’s staff were the only group who manifested stronger support for prescriptive norms in the interviews. This was also the best educated group in the
sample. A comparison of the scores of Group B with Group L showed significant differences in acceptability in informal spoken and written semi-formal register as well as in
the total score. On the other side of the spectrum, teachers in Group A and O showed an
approach of tolerance towards language variation due to their experience with new speakers of Welsh.
Finally, regarding the standardisation- related ideologies (Aim 3), the interviews
have provided some valuable information on proficient speakers’ views on phrasal verbs
in the context of language contact. The calquing of PVs was rarely recognised as an important issue in view of the linguistic norm. Bearing in mind Thomas’s (1991: 74) claim
that tolerance towards calquing is negatively correlated with the degree of purism, the
speakers’ opinions indicate that ideologies of purism were not very strong among the
interviewed groups. The calquing of PVs was placed in a wider context of the overwhelming influence of English on Welsh chiefly by those speakers who exhibited high linguistic
awareness due to their profession of teachers, researchers and translators. These informants were able to supply plenty of examples of calquing phenomena, stemming from their
contact with teenage Welsh speakers and new speakers. A few participants expressed
concern that Welsh might lose its native idiomaticity in the future and become “translated
English”. However, PVs were generally not seen as part of this trend. The reasons for that
were twofold: firstly, the speakers perceived the phenomenon of PVs as purely lexical,
not affecting the structure of the language. Secondly, phrasal verbs were not seen as substitutes for native words and idioms
The findings of the study are in accordance with Robert’s (2013) claims about the
general tolerance of Welsh speakers towards borrowings and language variation in vernacular practice. Data from the interviews indicate that prescriptive ideologies are in constant competition with other priorities, such as accommodating non-native speakers of
Welsh in school or the classroom environment or the need to communicate with wide
audience. In the present study, the majority of participants saw language practice and
effective communication as crucial for maintaining the Welsh language, even at the cost
340
of changing the traditional monolingual norm. The most common opinion on borrowings
from English given by the participants was that they are a natural element of their language. Speaking Welsh devoid of any English element was seen by some speakers as
awkward, associated with middle-class elitism and negative ideologies of purism.
Indeed, attitudes of the speakers towards purism were invariably negative as the
ideology was associated with criticism and stigmatisation. Anecdotal evidence provided
by the speakers regarding purists who criticise the Welsh of new speakers deemed such
ideology as detrimental for maintaining and revitalising Welsh, dividing society and discouraging people from using the language. Instead, participants of the study advocated
other methods of preventing the deterioration of Welsh, primarily the practice of “giving
good examples” themselves. Teachers in particular saw it as their mission to encourage
students to use “good” Welsh, which would at the same time be “natural” and that students would feel comfortable to use. In doing that, teachers need to regard not only abstract norms of “correctness” but also the conditions of the environment in which they
function and the abilities of students. On the other hand, it was also mentioned that new
speakers may desire to hear Welsh which is more “pure” at the lexical level as they expect
it to be “different” from English and are more conscious in separating the two languages.
Thus the emerging picture is that of a constant negotiation of standards between more and
less proficient speakers of Welsh, where lack of confidence in speaking is seen as a much
greater problem than lack of “correctness” or using Anglicised vocabulary and structures.
All in all, the interviewed speakers manifested support for diglossia, but at the
same time were aware of what Musk called the “dichotomy between the everyday practice
of bilingualism and the ideal” (Musk 2006: 405) embedded in the monolingual norm. In
his study of a bilingual school community, the author mentioned that members of
a Welsh-dominant group did not contest ideologies based on the monolingual norm. Similarly, the professional speakers in the present study, most of whom were native speakers
of Welsh, were supportive of the existence of the monolingual norm at least in writing as
an instrument of maintenance and transmission of the language (cf. Schiffman 1998: 211).
However, the majority of speakers indicated that implementing such norms in real life
was a challenging task.
Nearly all the speakers admitted having difficulties in recalling Welsh vocabulary
on an everyday basis and the majority declared that they regularly used dictionaries, primarily in writing, but sometimes also during conversation. This shows the importance of
341
dictionaries in maintaining the linguistic norm among professional speakers, in particular
Geiriadur yr Academy, which was seen as the main authority on Welsh vocabulary, confirming Prys et al.’s findings (Prys et al. 2015, see 1.2.6.2). However, only about a quarter
of the informants stated that they strove to make their Welsh better by supressing the
ubiquitous English influence, avoiding borrowings and code-switching. This approach
was particularly strong among teachers, who perceived themselves as linguistic models
for students. In most cases, the research had shown a tendency towards increased informality of the semi-formal registers and accommodating English vocabulary in them. The
idea of rigid separation between the two languages was not popular among the speakers
(cf. Selleck 2013 and the idea of “flexible bilingualism” 1.2.1). Overall, the informants
expressed acceptance for English borrowings motivating it with various factors: for instance, they saw borrowings as tokens of their dialect, and a universal means of communication with learners or people with perceived weaker command of Welsh. Adaptation
to the speaker or audience was seen as a crucial criterion for the level of adherence to
monolingual norms. Prescriptive ideologies were most prominent in writing, where the
vast majority of the informants declared they would avoid using borrowings from English.
The final issue addressed in the study was the notion of standard Welsh. The study
has confirmed observations presented in Chapter 1 according to which modern standard
Welsh is a rather vague concept even in the minds of proficient Welsh speakers, owing
to multiple channels of its transmission in the space left after the decline of the traditional
norm connected with the chapel and literary language. The opinions gathered in the interviews show that the interviewed professional speakers associated the standard primarily with writing and its different nature from the spoken idiom. With regard to spoken
standard, the speakers were evidently less confident in talking about spoken formal variety than written one due to the fact that in general they declared having fewer opportunities to use it on everyday basis. Moreover, using the spoken standard was sometimes seen
as more demanding since speaking gives little time to monitor one’s language.
However, despite the fact that standards were mostly associated with written language, it was the broadcast media rather than literature or press that was mentioned more
often by the speakers while specifying the domain of the standard language. This demonstrates something of a clash between the traditional monolingual norms rooted in written
language and new standards introduced through the spoken language of the media where
342
the bilingual element is evidently stronger. For this reason, a number of speakers believed
the media language to be too informal and advocating “bad” Welsh influenced by English.
For others, however, the Welsh they heard on the radio or on television were still too rigid
and artificial.
Several speakers pointed to the increasing informality of the written language.
Writers of the Y group did not believe in maintaining traditional forms in fiction as it
should reflect the real language. Similarly, journalists of group G have given evidence of
how the norm in press has adapted to the changing needs of the bilingual audience over
the years. Another factor mentioned by several speakers was the influence of social media
where code-switching is widespread. Having been asked in the questionnaire to indicate
what they would expect in standard language used in the press, some informants drew
attention to the fact that their expectations were in contrast to what they saw in actual
usage.
The official variety of Welsh was most often discussed in a negative context during the interviews as one that fails to serve as the source of linguistic norm. Several of the
professional speakers interviewed admitted that they did not use the official language or
had problems in understanding this register. Many speakers recognised the need to advocate the Cymraeg Clir approach, especially in Group B, whose members worked in the
institution promoting the initiative. Consequently, members of this group were more tolerant of calqued PVs in official writing in comparison with semi-formal written register
(writing a book) as they acknowledged the need for effective communication with the
public. Several members of Group C who were sensitive to the presence of Welsh in
official bodies also commented on the need to increase the accessibility of Welsh documents and the presence of Welsh in politics.
5.6. Limitations of the study
The study presented in this chapter has a number of limitations which will be now addressed:
As discussed in 1.5., the obtained data contain a number of self-reported claims, which
should be treated with caution as they do not provide information about the actual
343
behaviour of participants. Nevertheless, such information is valuable regarding speakers’ perception of their language and their ideologies, which were the main subject of
the analysis. Moreover, it has been observed that the results of Questionnaire B regarding the informal register were generally in congruence with actual use of PVs in
the participants’ natural speech.
Regarding the acceptability of PVs it should be noticed that the format of Questionnaire A where participants had a choice between PV and non-PV items might have
compelled some speakers to choose the one “right” option despite having been told
about the possibility to mark as many phrases as they saw appropriate. This might have
resulted in a bias towards non-PV items.
Statistical results of the analysis of extralinguistic factors in relation to Questionnaire B scores should not be generalised as they are representative only for the selected
small sample of speakers.
Choosing the method of semi-structured interviews allowed for freedom of expression
but made it impossible to touch upon the same themes with all the participants. Moreover, due to time constraints (many of the participants were interviewed during their
working hours and some of them needed more time than usual to fill the questionnaire)
it was not always possible to cover all the intended topics in the interview. As a result,
the data collected in the interviews are unsystematic and cannot be measured quantitatively but rather show trends in speakers’ judgements.
As the first field study devoted to the subject of phrasal verbs, the investigation has outlined major themes related to the research questions, which may be the basis for future
studies of Welsh PVs using wider samples of Welsh speakers and comprising not only
self-reported use but also investigating actual usage and language attitudes.
5.7. Conclusions
The main aim of this chapter was to investigate the level of integration of phrasal verbs
in Welsh, using the judgements of language-aware professional speakers as a litmus test
for the strength of the linguistic norm. The study revealed the speakers’ strong attachment
to diglossic character of Welsh: whilst in the informal spoken PVs of all categories were
344
found to be extremely well-integrated, in the written language a prescriptive norm against
them was manifested in the informants’ judgement, although levels of acceptability varied considerably across the investigated cases. There was a negative correlation between
the level of education and acceptability of PVs in the written language, but only in a
situation involving writing a book, not in public communication. The acceptability of
idiomatic PVs in writing depended on their stylistic formality, semantic functionality and
their status with regard to contact with English. The naturalness of non-PV equivalents in
a given context also played a major role in the speakers’ judgement. In nearly all the
investigated cases, the speakers declared their preference for native Welsh equivalents to
PVs with the exception of the selected native verb-particle construction bwrw ymlaen.
The acceptability of loanblends was markedly lower than PVs of other language contact
categories. The results showed a high linguistic awareness of the speakers and their sensitivity to differences in register and style.
Within the wider context of language ideologies, it can be concluded that prescriptive ideas were manifested primarily in the domain of writing, but even there the norm
against using PVs was not particularly strong. Although a group of speakers expressed
concern over the phenomenon of calquing from English, they did not perceive PVs as
a major threat to the Welsh language, viewing them as constructions belonging mainly to
colloquial speech that are not likely to replace native items entirely. The prevailing ideology among the interviewed speakers was that of tolerance towards PVs specifically and
borrowings in general. The speakers did not expect themselves or others to be “double
monolinguals” according to the traditional normative definition of bilingualism (Clyne
1998: 311) and considered the calqued verb-particle constructions as a natural and sometimes useful element of their vocabulary. Ideologies of purism were seen as detrimental
by a number of speakers, who believed that prescriptive rules should be introduced in
a positive, encouraging way. Effective communication and accommodating new speakers
of Welsh were seen as crucial for maintaining the language.
Moreover, the data gathered in the interviews have shown that the professional
speakers perceived tension between the “ideal” monolingual norm and the bilingual reality which requires a constant negotiation of standards. Many of the informants saw it as
their duty to keep the norm in the written language, given their position as teachers, translators, journalists etc., but believed it neither possible nor necessary to adhere to such
norms in speech. Thus the influence of English on spoken Welsh was seen as a natural
345
and inevitable thing. However, a number of speakers have indicated that the written language is subject to change as well, becoming increasingly informal and anglicised. The
discrepancy between the expectations of the professional speakers regarding the written
norm and actual practice is shown in the fact that while the speakers generally rejected
verb-particle constructions given a choice between PV and non-PV items, the questionnaire contained authentic sentences which originally did contain phrasal verbs.
In view of the complex picture emerging from the results presented in this chapter,
the conclusions drawn from the empirical study are by necessity generalising. The multitude of opinion on the standardisation of Welsh and issues of language contact expressed
by the participants points to the phenomenon acknowledged in recent studies on minority
languages standardisation, that is the “acceptance of pluralism and/or ambiguity by actors
participating in minority standardisation projects which embrace the diversity of speakers” (Costa et al. 2017: 13–14). Future linguistic norms in Welsh and new semi-formal
registers are bound to arise out of an amalgam of monolingual-oriented normative practices and everyday communication in a bilingual environment, which currently are in the
process of emerging. The degree of uncertainty and variation in the acceptability of
phrasal verbs in standard Welsh is part of this phenomenon.
346
Conclusion
Phrasal verbs have been a largely neglected subject in Welsh linguistics, except for the
pioneering article by Rottet (2005), which has provided a solid basis for the present work.
In my dissertation I have strived to expand on Rottet’s findings by conducting a larger
and more in-depth study using a variety of methodologies. First, by conducting a thorough
literature review and analysing my own corpus of the press and fiction I identified and
investigated major features of phrasal verbs in Welsh. This builds on Rottet’s (2005) findings in a more systematic way and amends a simplified and erroneous description of these
constructions found in Thomas (1996). Secondly, as idiomatic phrasal verbs in Welsh are
primarily, though not solely, a phenomenon of language contact, I have focused my study
on the level of their integration in the standard written language, which, as explained in
the Introduction, is not only of theoretical but also practical importance. Special attention
has been paid to idiomatic constructions and their status in relation to the linguistic norm
in order to put the phenomenon within the broader context of language standardisation
processes and changes which take place in Welsh bilingual society of today.
The changing situation of the Welsh language and the processes influencing the
linguistic norm were described in Chapter 1, providing a background for discussing issues
of language contact and standardisation. Chapter 2 presented the definition and classification of Welsh phrasal verbs on the basis of literature on the subject in English and
Welsh. The subsequent analytic chapters investigated the constructions from several complementing perspectives: Chapter 3 presented a corpus study of phrasal verbs occurring
in the press and fiction, accompanied with a cognitive analysis of particle semantics,
which identified marked particle senses; Chapter 4 investigated the representation of
phrasal verbs in Welsh pedagogical resources and dictionaries; finally, Chapter 5 presented the results of a field study among professional speakers of Welsh, assessing the
347
acceptability of idiomatic phrasal verbs in various registers of the language, with focus
on standard written varieties. The studies presented in Chapters 3-5 allowed for an indepth analysis of numerous issues outlined by Rottet (2005). I have broadened the scope
of his research by redefining the concept of phrasal verb, including idiomatic phrasal
prepositional verbs and prepositional verbs. Based on systematic corpus data, I have analysed different categories of phrasal verbs, providing new data on their frequency,
productivity, transitivity and stylistic properties. I have also conducted a systematic analysis of a representative sample of lexicographic and pedagogical materials containing the
most up-to-date materials and conducted the first field study on the acceptability of
phrasal verbs in various registers of Welsh. The gathered data and findings provide a firm
basis for answering the main research questions posed in the Introduction:
1. What are the major characteristics of phrasal verbs in Welsh?
2. Are Welsh phrasal verbs integrated in the standard written language of today?
3. How are Welsh phrasal verbs represented in contemporary grammars, teaching materials and dictionaries?
4. What ideologies towards phrasal verbs are manifested by proficient speakers of
Welsh?
The results of the study have provided answers to these questions which, taken together,
form a comprehensive description of phrasal verbs in Welsh in relation to the Welsh linguistic norm. Rather than answering the questions one by one, the collected observations
are summarised below in one section, offering a foray into further studies and a potential
source of data for linguists, lexicographers and authors of teaching materials.
Welsh phrasal verbs – general characteristics
Definition and categorisation
Phrasal verbs (berfau ymadroddol) in Welsh can be defined as highly lexical syntactic
combinations of a verb followed by a particle: preposition, prepositional adverb or spatial
adverb (2.4.1). Syntactically, phrasal verbs can be categorised into adverbial phrasal
348
verbs consisting of a verb and an adverbial particle (e.g. dod ymlaen), prepositional verbs
consisting of a verb and a preposition (e.g. dod ar draws) and phrasal prepositional verbs
consisting of a verb and two particles, the second of which is a preposition (e.g. edrych
ymlaen at) (2.4.3). Phrasal verbs can be also divided along the continuum of idiomaticity
into transparent, semi-idiomatic and idiomatic constructions (2.4.2).
Transparent vs. idiomatic
In transparent phrasal verbs the verb retains its literal meaning and expresses verbal action, while the particle expresses directionality of the action, e.g. dod i mewn ‘come in’,
taflu i ffwrdd ‘throw away’. Transparent phrasal verbs are natural to Welsh, meaning that
they have been attested in the earliest written accounts of the language and there is rarely
an alternative way to express the same concept (2.2.2, 2.4.2, 4.3). As both the verb and
the particle are exchangeable, transparent constructions are relatively free and only a limited number of them can be replaced with a single word e.g. dod yn ôl ‘come back’ with
dychwelyd ‘return’, mynd i ffwrdd ‘go away’ with gadael ‘leave’. Thus, many transparent
phrasal verbs may be perceived by speakers as free combinations of verb and particle
rather than single units. For this reason, it is the idiomatic constructions which are most
often associated with the term phrasal verb.
In idiomatic phrasal verbs the meaning of the phrase is not fully predictable from
the meaning of its constituents and there is no possibility of contrastive substitution of the
verb or particle, e.g. dal i fyny ‘catch up’, torri ar draws ‘interrupt’. Semi-idiomatic
phrasal verbs are those in which the verb retains its literal meaning, while the particle
adds an aspectual interpretation of the verb e.g. tyfu i fyny ‘grow up’.
Native vs. borrowed idiomatic phrasal verbs
Idiomatic and semi-idiomatic phrasal verbs in Welsh have emerged as an outcome of two
independent yet intertwining linguistic phenomena: 1) historical extension of particle
meaning from transparent to idiomatic and 2) transfer from English. It has been observed
that Welsh is capable of forming its native idiomatic phrasal verbs: prepositional verbs,
such as torri ar draws ‘interrupt’, lit. break across’, colli ar ‘lose your senses’, lit. ‘lose
on’, as well as adverbial phrasal verbs, such as bwrw ymlaen ‘continue’, lit.
349
‘throw/hit/spend on’, codi allan. ‘stir out of doors’, lit. ‘rise out’ (3.2.3.4). The corpus
data suggest that native prepositional phrasal verbs outnumber native adverbial phrasal
verbs. The very rare native phrasal prepositional verbs are usually variants of adverbial
constructions e.g. bwrw ymlaen â ‘continue with’. There is evidence to claim that extension of particle meaning from transparent to idiomatic is a natural process in the Welsh
language. The study based on the cognitive model of particle semantics (3.3) has identified particle senses which do not occur in English e.g. ar draws conveying an idea of
interruption, as in torri ar draws ‘interrupt’, siarad ar draws ‘interrupt’, and one Welsh
particle which is not directly translatable into English, draw, used in constructions such
as galw draw ‘call by’ to express the concept of visiting for a short time (3.3.2.10,
3.3.2.12). The naturalness of idiomatic phrasal verbs in Welsh is also proven by the presence of ad hoc metaphorical constructions found for instance in creative writing, such as
diferu heibio ‘flow by’ (about time), lit. ‘drop by’ (3.2.3.4).
Notwithstanding the potential of the Welsh language to form idiomatic verb-particle constructions, it cannot be doubted that this tendency has been immensely reinforced
by contact with English. A cross-examination of corpus findings with the historical dictionary of Welsh Geiriadur Prifysgol Cymru has demonstrated that this process is not
recent given that a great number of constructions which are directly translatable into English entered the language centuries ago (3.3). However, further research on historical corpora is needed to provide more detailed observations. As far as the contemporary language is concerned, idiomatic phrasal verbs are a widespread phenomenon: the study on
a small corpus of less than 1 million words has identified 271 types of idiomatic phrasal
verbs with 398 different senses, which sample is by no means exhaustive. A rough frequency count performed for the fiction subcorpus showed that the amount of idiomatic
phrasal verbs per 1000 words varied from 0.9 to 2, depending on the author (3.4).
Classification of transferred phrasal verbs
Around 90% of the items in the corpus had their exact English counterparts. Although in
some cases the similarity of idiomatic phrasal verbs in Welsh and English may be coincidental (3.3), the fact that the vast majority are directly translatable into English must be
attributed to language contact. Consequently, most idiomatic phrasal verbs in Welsh can
350
be classified into the following categories according to the mechanisms of transfer from
English (3.2.3.4):
1.
calques, that is word-for-word translations of English verb-particle constructions
consisting of native Welsh morphemes, for instance mynd ymlaen ‘go on’, ‘happen’,
troi i fyny ‘turn up’, ‘arrive’. They constituted around two thirds of idiomatic phrasal
verbs in the corpus.
2.
loanblends, that is calques which combine borrowed and native morphemes. They
are most frequently formed with a borrowed English verb with the ending -(i)o, e.g.
ffeindio allan ‘find out’, cario ymlaen ‘carry on’; in some cases the particle or both
elements may be borrowed, as in ‘troi off’ ‘turn off’, switcho off ‘switch off’.
Loanblends constituted about one tenth of phrasal verbs in the corpus.
3.
pleonastic calques, that is semi-idiomatic calques in which the particle is, from a
logical standpoint, redundant. These are usually formed with prepositions i lawr, i
fyny, o gwmpas and allan, e.g. eistedd i lawr ‘sit down’, tyfu i fyny ‘grow up’,
chwarae o gwmpas ‘play around’, helpu allan ‘help out’. Pleonastic calques constituted less than a tenth of phrasal verbs in the corpus.
4.
loan renditions, that is constructions in which the translation diverges from the English model rather than reproduces it element by element. A common type of loan
renditions are those rendering English phrasal verbs containing get with the verb dod
‘come’, e.g. dod ymlaen ‘get on’ (with someone). This type of transfer is relatively
rare.
5.
wholesale borrowings, that is constructions in which the whole linguistic material is
borrowed from English. They are a marginal phenomenon occurring in a couple of
stock phrases such as cym on ‘come on’, howld on ‘hold on’. While in written texts
these items can be distinguished by a Welsh spelling, in speech they might be difficult to differentiate from code-switching.
Productivity
Welsh phrasal verbs are typically formed with common lexical verbs associated with
physical movement. The corpus study has identified over 120 verbs which enter idiomatic
verb-particle combinations. The most productive verbs were mynd ‘to go’ and dod ‘to
351
come’. Noticeably, over one third of phrasal verbs in the corpus were formed with verbs
borrowed from English (3.2.3.1). As for the particles, the study has identified 28 particles
– 15 adverbial and 13 prepositional – which can be used to form idiomatic phrasal verbs,
excluding numerous dialectal and colloquial variants. The most productive particles were
allan, i fyny and ymlaen (3.2.3.1, 3.4). Allan was the most frequent particle in the corpus,
forming a quarter of all phrasal verb types. It occurred in items recorded centuries ago as
well as recent calques from English. Generally speaking, it has been observed that phrasal
verbs are highly productive constructions as one can identify a number of new phrases
expressing modern concepts. Among these are mostly direct calques from English, e.g.
optio allan ‘opt out’, logio i mewn ‘log in’ but also cases of loan renditions such as llwytho
i fyny ‘upload’. It has been shown that a considerable number of calqued phrasal verbs
blend naturally into Welsh due to the fact that the particles convey figurative senses that
are relatively transparent and well-established in the language (3.3).
With regard to syntactic classification, the most frequent type of Welsh phrasal
verbs are those formed with an adverbial particle. They constituted around two thirds of
phrasal verbs in the sample. Phrasal prepositional verbs constituted less than 20% of the
types and half of them were variants of adverbial phrasal verbs. Prepositional verbs constituted around 15% of the items. The fact that the proportion of native constructions
and/or loan renditions was the highest in this syntactic category suggests that idiomatic
prepositional verbs are better established in Welsh than adverbial constructions.
Transitivity
Depending on the syntactic category, Welsh phrasal verbs have different properties regarding transitivity. Prepositional phrasal verbs always take a direct object. Adverbial
phrasal verbs may be intransitive or transitive, while phrasal prepositional verbs with two
particles are mostly intransitive; only a very limited number may take a direct object
(2.4.3).
There has been some inconsistency in the literature regarding the position of the
direct object in transitive adverbial phrasal verbs (2.4.3.1). The corpus study has shown
that although both Verb-Object-Particle and Verb-Particle-Object orders are possible, the
Verb-Object-Particle is the prevalent one. The rare Verb-Particle-Object order appears to
be archaic and therefore stylistically marked in contemporary texts. However, it is also
352
occasionally encountered in informal spoken language, and also with multi-word objects.
The re-emergence of this word order can be attributed to the influence of English and
exemplify a case of lexico-syntactic rather than purely lexical calquing.
Stylistic properties and acceptability
The corpus study and field study data have shown that idiomatic phrasal verbs are widely
used in both spoken and written language. Although the corpus contained only samples
of written texts, the analysis acknowledged a division between narrative and conversation
modes, with the latter reproducing spoken Welsh in dialogues or interviews. The distribution of items depending on the mode of narration was almost equal and there were very
few marked differences between them with regard to language contact categories: native
phrasal verbs and loan renditions were slightly more prevalent in narrative, whilst wholesale borrowings occurred almost exclusively in conversation (3.2.5.4). Only a small number of items in the corpus were stylistically marked and non-standard in that they were
associated with slang used by teenagers, members of the working class and learners. The
majority of phrasal verbs in the corpus were unmarked and used in semi-formal as well
as informal contexts (3.2.6).
Questionnaire results and data gathered in interviews with professional speakers
(for this definition see 5.3.3.1) have confirmed that phrasal verbs are highly integrated
into informal spoken language. In a judgement test, all of the investigated phrasal verbs
were highly accepted by the speakers in the context of informal conversation. In the formal spoken register the acceptability of some highly idiomatic phrasal verbs was markedly lower. In the written registers there was a considerable decrease of acceptability of
loanblends whose form evokes the English influence. In contrast, native phrasal verbs
were accepted in every register investigated. Overall, the participants’ stated choices revealed a tendency to avoid calqued phrasal verbs in more formal context, in particular in
the written language (5.4.3).
A similar dichotomy has been noted in the representation of phrasal verbs in lexicography and pedagogical materials (4.2, 4.5). On the one hand, Welsh dictionaries contain a meaningful number of accepted, well-established PVs, including both native and
borrowed constructions, while teaching materials for less advanced users, focusing on
everyday communication introduce a small number of most common phrasal verbs. At
353
the same time, it has been found that there exists a rather consistent linguistic norm prescribing the use of phrasal verbs. Most Welsh dictionaries have a tendency to exclude
calqued phrasal verbs and occasionally warn against them (4.3, 4.5), while a considerable
number of teaching materials contain sections on translation from English with particular
focus on phrasal verbs (4.2.1). These materials usually concentrate on the standard written
language and are targeted at proficient users who are bound to use Welsh in formal situations and in writing. Within this norm, phrasal verbs are generally associated with informal speech and/or seen as tokens of careless calquing from English.
All in all, the degree of acceptability and markedness of idiomatic phrasal verbs
varies considerably depending on a number of factors. The study has identified several
major motivations for accepting or rejecting a phrasal verb, which combine to form the
linguistic norm:
The existence of an alternative Welsh item. Many of the calqued phrasal verbs are core
borrowings, duplicating native items. In the questionnaire, the participants showed
strong preference for single-word items over calqued phrasal verbs (5.4.2.). In contrast,
the speakers acknowledged the need to embrace a small number of phrasal verbs which
fill a lexical gap. Thus, for example, troi ymlaen ‘turn on’ was generally accepted due
to the lack of a well-established Welsh expression for the same concept, while troi i
ffwrdd ‘turn off’ was often rejected due to the fact that there exists a simple equivalent
diffodd ‘turn off’, ‘extinguish’. The same sense of commitment to native Welsh vocabulary was found in lexicography and teaching materials which occasionally underlined the superiority of native items over calqued ones (4.2, 4.3, 4.5).
Perceived degree of formality of a phrasal verb and alternative Welsh items. As the
majority of idiomatic phrasal verbs in Welsh are calqued, there is a case for suggesting
that the perception of informality of phrasal verbs in English influences their status in
Welsh. In consequence, a decision to use a phrasal verb or an alternative item may
depend on the formality of a situation. In the questionnaire, participants occasionally
rejected alternative items due to their excessive formality in comparison with a wellintegrated phrasal verb (e.g. dychwelyd ‘return’ vs. dod yn ôl ‘come back’, disgwyl yn
eiddgar ‘ardently expect’ vs. edrych ymlaen at ‘look forward to’) (5.4.2). This might
indicate increasing informalisation of the written language. On the other hand, many
354
phrasal verbs are strongly associated with informal spoken idiom and therefore rejected in writing (e.g. mynd ymlaen ‘go on’, ‘happen’). The research highlighted that
the boundaries between spoken and written idiom were not clear-cut and speakers were
aware of a wide range of registers with different degrees of formality. For example,
some phrasal verbs were seen as inappropriate in a non-fiction book but effective for
official communication in a leaflet (5.4.3).
Perceived English origin of a phrasal verb. This criterion is connected with the feeling
of ‘naturalness’ of a metaphorical meaning of the particle in the Welsh language. A
comparison of the cognitive analysis of particle semantics in 3.3 with the study in 4.5
has revealed which particles are less integrated in Welsh and therefore rejected by
some speakers. The most important of these are:
o
allan indicating maximum boundaries, completeness of an action (3.3.2.1);
o
i fyny and i lawr in most of their figurative senses (3.3.2.2 and 3.3.2.4);
o
i mewn (i) in its figurative senses (3.3.2.5);
o
off (see 3.3.2.7);
o
o gwmpas (see 3.3.2.9) in its figurative senses.
Knowledge about a prescriptive rule. Prescriptive rules against the use of some
calqued phrasal verbs have been observed to be passed from generation to generation:
authors of dictionaries and teaching materials occasionally refer to authorities from the
past (4.2.1, 4.3); similarly, some of the interviewed speakers declared they had learnt
about forbidding rules from their parents or teachers (5.4.1.5, 5.4.4.2). Higher educational institutions certainly play a role in maintaining linguistic norms; in the study on
professional speakers, university education was found to negatively correlate with the
acceptability of phrasal verbs in the semi-formal written registers (5.4.3.2.). Moreover,
prescriptive rules might be introduced at other institutions and workplaces, such as
Canolfan Bedwyr, via language courses and trainings advocating sensitivity to the use
of calques constructions.
Phrasal verbs and linguistic norms and ideologies
The latter motivation behind the acceptability of Welsh phrasal verbs leads to the question
of underlying linguistic ideologies. The study presented in Chapter 5 has revealed a number of dynamic processes shaping the current linguistic norm in Welsh that is constantly
355
negotiated between different groups of speakers. At the risk of making some generalisations, it is worth outlining the prevailing stances towards the phenomenon of phrasal verbs
which emerged in the course of the research.
The speakers recruited for the field study acting in the selected groups were mostly
language influentials (see 1.2.1), that is members of society who contribute to the emergence of the linguistic norm (teachers, translators, journalists, writers, language activists,
etc.). It has been suggested by Robert (2013), who investigated the use of loanwords
among similar groups of Welsh speakers, that the prevailing ideology was tolerance towards borrowings and language variation in the spoken language, coexisting with what
the author calls “purism” at the institutional level. My study, focusing on calquing, has
confirmed this twofold approach to the English element in Welsh, expressed in the vast
difference of acceptability of phrasal verbs across various registers.
However, I would like to problematize the term purism here. For those of my
participants who discussed the subject, the term purism had invariably very negative connotations. It was associated with condemnation and stigmatisation and seen as detrimental
for language revitalisation. In discussing the future of Welsh, speakers perceived a lack
of ‘correctness’ or ‘purity’ as a less serious problem than a lack of confidence in using
Welsh, as it is the latter that might result in abandoning the language. Language practice
and effective communication were seen as crucial for maintaining the Welsh language
even at the cost of changing the traditional norms. The participants often emphasised that
prescriptive rules should be introduced in a positive, encouraging way, bearing in mind
the abilities of Welsh speakers at various levels of proficiency (5.4.4).
For this reason, in describing the ideologies manifested by the speakers, I would
avoid the term purism. Purism, especially in the context of minority languages, involves
negative discourse of rejecting language change and creating boundaries between speakers who use the ‘correct’ or ‘corrupted’ varieties of language (cf. Dorian 1994). Instead,
I would employ the term protectionism, which I understand as an ideology that involves
a sense of responsibility toward the Welsh language and in the context of calquing is
manifested by the wish to protect Welsh from the inconspicuous influence of English.
Since calquing is associated with low awareness and lack of control, protectionism implies monitoring one’s language and raising the linguistic awareness of others. At the
same time, advocates of protectionism do not want to force others to monitor themselves
356
in their use of calqued structures. The interviewed speakers generally believed that education should not take the form of criticism and imposing prescriptive rules but rather
provide people with the ability to switch from a bilingual mode involving transfer phenomena, such as calquing, to a monolingual mode whenever they see fit. The speakers’
notion of proficiency in the language was thus rooted in the traditional, structuralist approach in that it was identified with the ability to control and monitor one’s language,
separating the Welsh and English elements. This agrees with the findings of Musk (2006)
and Selleck (2013) who investigated linguistic ideologies in a bilingual school environment.
As the interviewed speakers were generally inclusive and tolerant towards linguistic variation, there were considerably few voices expressing concern over the Anglicisation of Welsh lexicon or calquing (5.4.4.). The presence of English vocabulary in spoken
Welsh was generally accepted as an inevitable outcome of bilingualism. The calquing of
phrasal verbs in everyday conversation was seen as natural and largely unnoticed by the
vast majority of participants. However, with regard to writing, a much higher proportion
of speakers were cautious about transfer from English. In general, the speakers were supportive of the existence of a monolingual norm in writing and saw it as their duty to
maintain standards in more formal contexts. Yet, calquing of phrasal verbs was not seen
as a major threat due to the fact that it was less noticeable or seen as a lexical, not structural phenomenon. Many of the participants believed that as long as the structure of the
language remained Welsh, lexical transfer was not a serious issue (5.4.4.3).
On the other hand, many speakers indicated that maintaining monolingual norms
in real life was a challenging task, given the strong influence of English. The participants
also manifested a degree of uncertainty and insecurity with regard to linguistic norms
which are in the process of change. It has been demonstrated that standard Welsh was a
rather vague concept in the participants’ minds, associated mostly with two domains: formal writing and broadcast language (5.4.4.5). In consequence, the new emerging standard
appears to amalgamate traditional monolingual-oriented written language with spoken
idiom used in the media, where the bilingual element is more prominent. The emergence
of new media, such as social networks and problems with normalising Welsh as the language of law and administration also contribute to the uncertainty regarding future direction of standardisation. Many of the speakers acknowledged a discrepancy between their
357
expectations regarding written language and actual practice, indicating that writing is becoming increasingly informal and anglicised. Stories of personal experience reported by
the speakers showed that prescriptive ideologies are in constant competition with other
priorities, such as accommodating speakers with different linguistic backgrounds. This
was visible in the study of the acceptability of calqued phrasal verbs which were seen as
more effective than native vocabulary items in situations such as writing for a general
audience (5.4.3).
In conclusion, it appears that in view of the current processes of standardisation,
phrasal verbs are bound to become increasingly integrated in the semi-formal varieties of
Welsh. This is also confirmed by their prevalence and unmarkedness in the corpus. This
can be seen as a token of the growing Anglicisation of Welsh lexicon. On the other hand,
it can be argued that the presence of phrasal verbs doubling native items creates stylistic
variation which indicates the healthy status of the language.
Limitations of the study
The multi-perspective approach adopted in this study aimed at providing a comprehensive
description of phrasal verbs in Welsh. I have used a mixed methodology, focusing on
qualitative research, which allowed for an in-depth study of issues related to the complex
phenomenon investigated. However, the present study had a number of limitations, in
particular with regard to its quantitative aspect:
1. Firstly, one should acknowledge a multitude of possible definitions of phrasal verbs.
It should, for example, be noted that the definition adopted in this study has included
prepositional verbs, which are often treated as a separate category, especially in studies
focusing on syntactic properties of verb-particle constructions (2.1.1). Another debatable issue are the criteria for deciding about the level of idiomaticity of a phrasal verb
as in some borderline cases the ultimate decision rests on the researcher’s intuition
(2.4.2.1). In consequence, the present findings may be inappropriate to use in comparative studies with other languages or incompatible with other research on the subject.
This is a common and unresolved issue in studies of phrasal verbs in English linguistics.
358
2. The analysis in Chapter 3 presented a corpus created specifically for the purpose of
the dissertation. Several limitations of this study have been presented in 1.5. and 3.2.
The most important ones are:
1) the small size of the sample,
2) underrepresentation of some registers and dialects, in particular the spoken and administrative language,
3) possible thematic bias,
4) possible errors due to manual annotation.
For this reason, although the study has brought a number of valuable observations, the
quantitative values presented in Chapter 3 should be treated with caution.
3. The field study presented in Chapter 5 consisted of semi-structured interviews and
judgement test questionnaires. A number of limitations of this study was addressed in
5.6. Semi-structured interviews, which allowed for a degree of flexibility, made the
data unsystematic, outlining general trends and a variety of opinions rather than
providing consistent statistical data. It should also be borne in mind that, while the
chosen method of judgement test investigated the perception of phrasal verbs among
professional speakers of Welsh and the stated acceptability, it did not examine the
participants’ actual linguistic behaviour. The choice of a judgement test might have
also resulted in a bias towards non-PV items.
4. Although it can be argued that the sample of speakers participating in the study was
representative of professional speakers of Welsh, it should be noted that the geographical distribution of speakers was uneven, with the majority coming from Gwynedd,
Ceredigion and Carmarthenshire. The research would have certainly benefited from
including other groups located in South Wales. It would be particularly interesting to
investigate groups working in central authority institutions in Cardiff, which is a rapidly growing Welsh-speaking community and one with considerable intra-dialect contact. Unfortunately, speakers from this area were not consulted due to time constraints.
359
Recommendations for future research
In my thesis I have attempted to study the phenomenon of phrasal verbs from several
perspectives relevant to the research questions. Yet, the description of phrasal verbs produced is by no means exhaustive. Possible directions for future research include:
diachronic studies on the development of verb-particle constructions. The present dissertation touched very briefly on the subject drawing mainly from lexicographic data.
Studies on historical corpora would certainly yield more in-depth results;
the presence of phrasal verbs in the Patagonian variety of Welsh which has, for a relatively long time, developed independently from English influence;
studies on the use of phrasal verbs in the Welsh broadcast media, which have been
found to significantly influence the current linguistic norm;
corpus-based studies on larger samples of texts encompassing a wider variety of registers with a focus on quantitative data, measuring the degree of ‘success’ (cf. Zenner
et al. 2012) of calqued phrasal verbs in comparison with native items. Such studies
could be of particular use for lexicographers;
comparative studies on the use of phrasal verbs in natural occurring and monitored
speech, using the technique of participant observation for generating the most effective
naturalistic data;
as the study has identified some possible motivations for accepting or rejecting phrasal
verbs, the data could be used in future studies of attitudes towards phrasal verbs in
order to prepare rankings of motivations and investigate covert and overt attitudes of
speakers and their linguistic behaviour;
bearing in mind that the groups investigated in Chapter 5 of the present study were
selected on the basis of specific criteria of their language awareness and for that reason
are certainly not representative of the whole population of Welsh speakers, studies on
larger samples of speakers regarding the acceptability and/or attitudes of speakers
would prove an interesting direction of study;
finally, a comparative study analysing differences between Welsh and English would
be a valuable contribution to Welsh translation studies. Incidentally, as I have been
informed, a book on comparative stylistics of Welsh and English by Rottet and Morris
360
(in preparation) to be published a month after this thesis is submitted (September
2018), is going to contain a chapter on phrasal verbs which does exactly that.
Practical implementation of the findings
As mentioned in the Introduction, this thesis arose from my personal experience as a student of Welsh and the tensions I perceived between the norm and usage. Therefore, to
conclude this study, I would like to put forward some recommendations for a practical
implementation of my findings.
I. It has been shown that phrasal verbs are generally omitted in grammars of Welsh.
This does not seem to be justified, given that these constructions are not only extremely widespread in the spoken colloquial language but also that a number of
them also belong to formal registers of Welsh. What is more, they do not represent
a simple case of calquing, but a complex phenomenon which occurs natively in
the Welsh language. This calls for the inclusion of sections on phrasal verbs in
future grammar books.
II. It has also been demonstrated that some of the well-established, frequent and accepted phrasal verbs are missing from major dictionaries of Welsh. For instance,
the major historical dictionary Geiriadur Prifysgol Cymru does not include 6 out
of the 10 most frequent idiomatic phrasal verbs in the corpus, such as edrych
ymlaen at or mynd ymlaen. Therefore, a revision of the dictionary’s wordlist
should be considered and relevant items added. It would be worthwhile defining
a set of clear criteria for including phrasal verbs in Welsh dictionaries, pursuant
to a chosen normative or descriptive approach; such criteria could be stated in the
dictionary’s preface.
III. Incidentally, the questionnaire has revealed that Geiriadur yr Academi was the
most widely used dictionary among the interviewed professional speakers of
Welsh and it was perceived as the main authority on the language. The same has
been noticed by other researchers (Robert 2013:179, Prys 2015:358). In view of
361
the importance of this work, it is highly recommended that the digitalisation process of its online version be completed. Moreover, as the dictionary was published
over twenty years ago, a revision should be considered as well.
IV. Insights from linguistic research on phrasal verbs may be used in preparing Welsh
courses for young and adult speakers. It would be beneficial for teaching materials
to acknowledge the differences between various categories of phrasal verbs in relation to contact with English. Authors should be careful not to oversimplify the
issue by introducing general prescriptive norms, but rather draw students’ attention to the stylistic, semantic and dialectal variation in these constructions.
I hope that this thesis has laid the foundation for an accurate and systematic description of verb-particle constructions in modern Welsh and contributed to the discussion
about future linguistic norms. It has been shown that phrasal verbs are an essential element of the Welsh language and may be analysed from a variety of perspectives, with
many unanswered research questions left to pursue. Since this was the first more extensive investigation of the subject, the research may not be devoid of shortcomings characteristic of a little investigated field. Critical appraisal and further research are to be welcomed in order to provide more insights, spark debate and help transform scholarly work
into practical measures and output.
362
ABSTRACT
The dissertation is a comprehensive study of verb-particle constructions known as phrasal
verbs (PVs) in the Welsh language in relation to linguistic norm and contact with English.
Despite their widespread use, PVs in Welsh have not been studied in detail thus far, except for pioneering articles by Rottet (2000, 2005). The few other sources that mention
these constructions treat them as borrowings from English situated outside the linguistic
norm. The aim of my thesis was to provide a comprehensive albeit still preliminary description of PVs in Welsh and revise some oversimplifying statements regarding these
constructions by investigating the degree of their standardisation.
In the first chapter of the thesis, I present the situation of Welsh as a minority
language and attempt to define the sources of the current linguistic norm and standards. I
describe the decline of the so-called literary standard and the emergence of a new norm,
the so-called official or semi-formal register. The second part of the chapter discussed
relevant issues connected with English influence on Welsh with focus on lexical transfer.
Chapter 2 begins with a review of literature on PVs in English and Welsh. By
analysing contemporary grammars of Welsh, I demonstrate that the constructions are
most often excluded or erroneously described. The number of sources on Welsh PVs is
very limited as they are usually considered a priori to be calques from English. On the
basis of studies by Rottet (2000, 2005) and a comparative analysis of papers on PVs in
Irish (Stenson 1997, Veselinović 2000) it is shown that the issue of origins of these constructions in Celtic languages is much more complex. The key element is the division
between transparent and idiomatic PVs: while verbs from the first group occur naturally
in Celtic languages, the others are generally borrowed, although native constructions exist
as well. This implies that idiomatic PVs in Welsh are not only the outcome of direct
translation from English and in some cases may originate in natural processes of language
evolution, i.e. the semantic extension of transparent constructions towards idiomaticity.
In the second part of the chapter, I present a definition and classification of idiomatic PVs
used in the subsequent chapters.
363
In the third chapter I present a corpus of written Welsh created for the purpose of
the dissertation, comprised of works of fiction and periodicals which are aimed to be a
representative sample of semi-formal registers of Welsh. The results of the corpus analysis show the frequencies of PVs according to text type, syntactic categories and classification related to language contact. The corpus data are then used to compare the grammatical description of the constructions found in literature with my own observations.
The next part of the chapter is a qualitative analysis of corpus texts focusing on the examples of stylistic markedness of PVs, which indicates their acceptability in the linguistic
norm. In the final sections I analyse the meaning of phrasal particles found in the corpus
using Rudzka-Ostyn’s (2003) cognitive model of the semantics of particles. By comparing the obtained data with lexicographic materials I identify semantic extensions which
are most likely to have emerged due to contact with English.
Chapter 4 discusses the description of PVs in Welsh teaching materials and dictionaries. The analysis of numerous sources demonstrates the existence of a norm which
prescribes or discourages the use of some PVs as borrowing from English. However, the
norm can be observed primarily in materials aimed at native speakers or advanced learners, while resources used at the early stages of learning the language generally avoid
statements against borrowings. In the second subsection of the chapter I analyse dictionary entries for 25 most frequent PVs in the corpus, showing lack of consistency in accepting PVs as part of the linguistic norm.
Chapter 5 presents the results of a field study conducted in 2016 and 2017 on 55
professional speakers of Welsh in seven groups. The study consisted of an interview and
a questionnaire. The questionnaire checked the standardness and acceptability of PVs,
while the interviews touched on more general issues related to borrowing and standardisation. The results show variation in acceptability of PVs depending on the type of verb
and the used register. They also illustrate the complexity of attitudes related to the changing linguistic norm and point to potential factors which shape speakers beliefs on that
matter.
In the conclusions I propose a general description of the phenomenon of PVs in
Welsh and offer recommendations for including the constructions in grammars, dictionaries and pedagogical materials.
364
STRESZCZENIE
Niniejsza praca doktorska jest szerokim opracowaniem zagadnienia obecności konstrukcji czasownikowo-partykułowych zwanych czasownikami frazowymi w języku walijskim w kontekście normy językowej i kontaktu z językiem angielskim. Pomimo ich powszechności w uzusie czasowniki frazowe w języku walijskim nie doczekały się
dotychczas obszerniejszego opisu za wyjątkiem prekursorskich artykułów Rotteta (2000,
2005). Pozostałe rzadkie wzmianki w literaturze prezentują omawiane konstrukcje jako
zapożyczenia z języka angielskiego, a tym samym zjawisko spoza normy językowej. Niniejsza praca ma na celu dokonanie szeroko zakrojonej, aczkolwiek wciąż wstępnej, charakterystyki czasowników frazowych w walijskim oraz zrewidowanie nieraz upraszczających stwierdzeń na temat tych konstrukcji poprzez zbadanie stopnia ich
ustandaryzowania.
Rozdział pierwszy pracy przedstawia sytuację walijskiego jako języka mniejszości oraz stanowi próbę zdefiniowania źródeł normy językowej i standardów funkcjonujących w tym języku. Opisuję zanik tradycyjnej tzw. normy oraz wytworzenie nowej
normy, tzw. odmiany oficjalnej lub półformalnej. Druga część rozdziału porusza istotne
dla rozprawy kwestie związane z wpływem angielskiego na język walijski ze szczególnym uwzględnieniem transferu na poziomie leksykalnym.
Rozdział drugi rozpoczynam przeglądem literatury na temat czasowników frazowych w języku angielskim oraz walijskim. Analizując współczesne gramatyki walijskiego, wykazuję, że czasowniki te są w nich najczęściej pomijane lub też opis ich zawiera błędy. Liczba źródeł na temat omawianych konstrukcji w języku walijskim jest
stosunkowo uboga, gdyż najczęściej uznawane są a priori za kalki z języka angielskiego.
Na podstawie artykułów Rotteta (2000, 2005) oraz porównawczej analizy prac dotyczących języka irlandzkiego (Stenson 1997, Veselinović 2000) wykazuję, że zagadnienie
genezy czasowników frazowych w językach celtyckich jest znacznie bardziej złożone.
Kluczowy jest tutaj podział na czasowniki transparentne oraz idiomatyczne, gdyż pierwsze występują naturalnie w językach celtyckich, drugie zaś są w znacznej mierze zapożyczane, choć istnieją również konstrukcje rodzime. Wynika z tego, iż istnienie idiomatycznych czasowników frazowych w walijskim nie polega jedynie na bezpośrednim
365
zapożyczaniu angielskich konstrukcji i w niektórych przypadkach może stanowić element naturalnego rozwoju języka, polegającego na semantycznej ewolucji konstrukcji
transparentnych w kierunku idiomatyczności. W drugiej części rozdziału prezentuję definicję oraz klasyfikacje idiomatycznych czasowników frazowych wykorzystane w kolejnych rozdziałach pracy.
W rozdziale trzecim przedstawiam stworzony na użytek pracy korpus języka pisanego, złożony z dzieł literackich oraz materiałów prasowych, które w założeniu stanowią reprezentatywną próbkę tzw. półformalnych rejestrów języka walijskiego. Wyniki
analizy korpusu ukazują częstotliwości występowania czasowników frazowych w poszczególnych rodzajach tekstów, jak również według występowania poszczególnych kategorii syntaktycznych oraz klasyfikacji odnoszącej się do kontaktu językowego. Ponadto
na podstawie korpusu zestawiam gramatyczny opis tych konstrukcji w dostępnej literaturze z własnymi obserwacjami. W drugiej części rozdziału przeprowadzam analizę jakościową tekstów korpusowych, koncentrując się na przykładach stylistycznego nacechowania czasowników frazowych, badając tym samym ich akceptowalność w normie
językowej. W trzecim podrozdziale przy pomocy kognitywnego modelu semantycznego
autorstwa Rudzkiej-Ostyn (2003) analizuję znaczenia poszczególnych partykuł w korpusie. Porównując uzyskane dane z materiałem leksykograficznym, identyfikuję rozszerzenia semantyczne, które powstały najprawdopodobniej na skutek kontaktu z językiem angielskim.
Rozdział czwarty omawia obecność czasowników frazowych w walijskich materiałach dydaktycznych oraz słownikach. Omawiając szereg źródeł, wskazuję na obecność
normy zakazującej lub zniechęcającej do używania niektórych z omawianych konstrukcji
ze względu na ich status zapożyczeń z języka angielskiego. Norma ta obecna jest jednak
przede wszystkim w materiałach skierowanych do użytkowników rodzimych bądź o zaawansowanej znajomości języka, podczas gdy materiały dla uczniów na początkowych
etapach nauki z reguły unikają stwierdzeń skierowanych przeciwko zapożyczeniom. Następnie na podstawie analizy haseł słownikowych dla 25 najczęściej występujących w
korpusie czasowników ukazuję brak jednorodności w akceptowaniu czasowników frazowych w ramach normy językowej.
Rozdział piąty prezentuje wyniki badania terenowego przeprowadzonego w 2016
oraz 2017 r. na 55 profesjonalnych użytkownikach języka walijskiego zebranych w siedmiu grupach. Badanie składało się z wywiadu oraz ankiety. Ankieta sprawdzała stopień
366
ustandaryzowania wybranych czasowników walijskich oraz to, w jakim stopniu są one
akceptowane przez użytkowników, natomiast w wywiadach poruszane były bardziej
ogólne kwestie związane z zapożyczeniami i standaryzacją. Wyniki ukazują zróżnicowany stopień akceptacji czasowników frazowych w zależności od ich typu oraz używanego rejestru języka, jak również złożoność postaw wobec zmieniającej się normy językowej. Wskazują też na potencjalne czynniki wpływające na przekonania użytkowników
w tej kwestii.
W konkluzjach pracy przedstawiam propozycję ogólnej charakterystyki zjawiska
czasowników frazowych w języku walijskim oraz rekomendacje dotyczące uwzględnienia opisywanych konstrukcji w gramatykach, słownikach i materiałach dydaktycznych.
367
References
Primary sources:
Corpora
Corpws Cyfochrog Cofnod y Cynulliad [Parallell Corpus of the Assembly Proceedings]
[CCCC]. 1999-2010. (http://corpws.cymru/ycofnod/ (date of access: 4 Apr.
2017).
Ellis, N. C., C. O’Dochartaigh, W. Hicks, M. Morgan, N. Laporte. 2001. Cronfa Electroneg o Gymraeg: A 1 million word lexical database and frequency count for
Welsh [CEG]. (http://corpws.cymru/ceg/?lang=en) (date of access: 4 Apr. 2017).
Luft, Diana, Peter Wynn Thomas and D. Mark Smith (eds.). 2013. Rhyddiaith Gymraeg
1300-1425 [Welsh prose 1300-1425]. http://www.rhyddiaithganoloesol.caerdydd.ac.uk. (date of access: 22 Sep. 2017).
Willis, David. 2004. Historical Corpus of the Welsh Language 1500-1850.
LINDAT/CLARIN digital library at the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics, Charles University,
http://hdl.handle.net/11372/LRT-883. (date of access: 22 Sep. 2017).
Course books, reference books
Bangor University. 2008. Llawlyfr gloywi iaith [Textbook for polishing language]. Bangor: Bangor University.
Breese, Nona and Bethan Clement. 2011. Seren iaith [Language star]. Aberystwyth: Atebol Cyfyngedig.
Coleg Cymraeg Cenedlaethol. 2013. Gweiddgareddau. Sgiliau iaith [Activities. Language skills]. (http://sgiliauiaith.colegcymraeg.ac.uk/tasgau/) (date of access: 5
Oct. 2017).
368
Cule, C.P. 1971. Cymraeg idiomatig; Priod-ddulliau byw [Idiomatic Welsh; Living idioms]. Y Bontfaen: D.Brown a’i Feibion.
Davies, Cennard. 1994. Y geiriau bach [Little words]. Llandysul: Gomer.
Davies, Cennard. 2016. Course: Gweithgareddau ymarfer [Course. Practice activities].
(http://www.ybont.org/course/view.php?id=81) (date of access: 29 Sep. 2016).
Davies, Emyr and Eirian Conlon. 2007. Cwrs canolradd: Llyfr cwrs [Intermediate
Course: Textbook]. Cardiff: CBAC.
Department of Welsh Language and Literature, University of Wales. 1988. Gloywi iaith 3
[Polishing language 3]. Caerdydd: Gwasg Prifysgol Cymru.
Gruffudd, Heini. 2000. Cymraeg da [Good Welsh]. Talybont: Lolfa.
Gruffudd, Heini. 2006. Welcome to Welsh - A complete Welsh course for beginners. Talybont: Y Lolfa.
Jones, Dafydd Glyn. 2016. Iawn bob tro. Help bach i’r ysgrifenwr [Correct every time.
A little help for the writer]. Y Bala: Dalen Newydd.
Jones, Morgan D. 1965. Cywiriadur Cymraeg [A guide to correct Welsh]. Llandysul:
Gwasg Gomer.
Jones, Morgan D. 1976. A guide to correct Welsh. Llandysul: Gomer Press.
Meek, Elin. 2005. Cwrs mynediad: Llyfr cwrs [Entry Course: Textbook]. Cardiff: CBAC.
Roberts, Fflur Rees. 2011. Gramadeg [Grammar]. (https://hwb.wales.gov.uk/) (date of
access: 24 Jul. 2017).
Stonelake, Marc and Emyr Davies. 2006. Cwrs sylfaen: Llyfr cwrs [Basic course: Textbook]. Cardiff: CBAC.
Thomas, Gwyn. 2012. Ymarfer ysgrifennu Cymraeg [Practicing written Welsh]. Talybont: Lolfa.
Williams, Cen. 1999. Cymraeg clir: canllawiau iaith [Clear Welsh: language guidelines].
Bangor: Cyngor Gwynedd, Bwrdd yr Iaith Gymraeg a Chanolfan Bedwyr.
Y Ganolfan Dysgu Cymraeg Genedlaethol. 2018a. Mynediad —Entry (A1) Dysgu Cymraeg [Entry (A1) Learning Welsh], part 1. [Unpublished ms.]
Y Ganolfan Dysgu Cymraeg Genedlaethol. 2018b. Uwch 1 (B2) Dysgu Cymraeg [Upper
(B2) Learning Welsh], part 1. [Unpublished ms.]
Dictionaries and collections of idioms
Amiot-Cadey, Gaëlle and Maggie Seaton (eds.). 2009. Collins Spurrell Pocket Welsh
Dictionary (4th ed.). Glasgow: Harper Collins Publisher.
369
Andrews, Tegau and Delyth Prys. 2015. Geiriadur Termau’r Coleg Cymraeg Cenedlaethol [Welsh National College Dictionary of Terms]. Cardiff: Coleg Cymraeg
Cenedlaethol. (http://www.colegcymraeg.ac.uk/termau/) (date of access: 17 Jun.
2016).
Bangor University. 2016. Geiriadur Bangor, an online Welsh-English, English-Welsh
dictionary. (http://geiriadur.bangor.ac.uk) (date of access: 18 Jan. 2018).
Beattie, Susie (ed.). 2017. Collins Spurrell Welsh Dictionary Pocket Edition. Glasgow:
Harper Collins Publisher.
Cownie, Alun Rhys. 2001. A Dictionary of Welsh and English Idiomatic Phrases. Cardiff:
University of Wales Press.
Evans, H. Meurig and Thomas (eds.). 1958. Y Geiriadur Mawr: The Complete WelshEnglish English-Welsh Dictionary. Abertawe: Christopher Davies.
Evans, H. Meurig and Thomas (eds.). 1959. Y Geiriadur Bach; The Welsh Pocket Dictionary. Aberystwyth: Llyfrau’r Dryw.
Griffiths, Bruce and Dafydd Glyn Jones. 1995. Geiriadur yr Academi: the Welsh Academy English-Welsh Dictionary. Cardiff: University of Wales Press. Online version at (http://geiriaduracademi.org/) (date of access: 17 Jun. 2016)
Gruffudd, Heini. 1998. The Welsh Learner’s Dictionary. Talybont: Y Lolfa.
Jones, Ceri. 2013. Dweud eich dweud: a guide to colloquial and idiomatic Welsh. Llandysul: Gwasg Gomer.
Jones, D.B., Delyth Prys and Gruffydd Prys. 2011. Welsh National Terminology Portal.
Bangor: Bangor University. (http://termau.cymru/?lang=en) (date of access: 17
Jun. 2016).
Jones, R.E. 1975. Llyfr o idiomau Cymraeg [A book of Welsh idioms]. Abertawe: Gwasg
John Penry.
Jones, R.E. 1987. Ail lyfr o idiomau Cymraeg [The second book of Welsh idioms]. Abertawe: Tŷ John Penry.
King, Gareth. 2000. The Pocket Modern Welsh Dictionary. Oxford: Oxford University
Press.
Lewis, D. Geraint. 1994. Geiriadur Gomer i’r Ifanc. Llandysul: Gomer.
Lewis, D. Geraint. 2012. Ar flaen fy nhafod [On the tip of my tongue]. Llandysul: Gomer.
Lewis, D. Geraint and Nudd Lewis (eds.). 2016. Geiriadur Cymraeg Gomer [Gomer
Welsh Dictionary]. Llandysul: Gwasg Gomer.
Lewis, D. Geraint. 2018. Gweiadur Pawb, an online Welsh-English, English-Welsh dictionary. (http://www.gweiadur.com) (date of access: 18 Jan. 2018).
370
Macmillan Dictionary Online. 2017. Macmillan Publishers Limited (http://www.macmillandictionary.com/) (date of access: 20 Sep. 2017).
Oxford English Dictionary online. 2017. Oxford: Oxford University Press.
(http://www.oed.com/). (date of access: 12 Sep. 2017).
Prys, Delyth and Gruffydd Prys. 2011. Y Termiadur Addysg. Cardiff: Welsh Government.
(http://www.termiaduraddysg.org/) (date of access: 17 Jun. 2016).
Thomas, R. J., A. Bevan and P. J. Donovan (eds.). 1967-2002. Geiriadur Prifysgol
Cymru: a dictionary of the Welsh language. 4 vols. Cardiff: University of Wales
Press.
University of Wales Trinity Saint David. 2018. Trinity Saint David Dictionary, an online
Welsh-English, English-Welsh dictionary. (www. geiriadur. net) (date of access:
18 May 2016).
University of Wales. 2014. Geiriadur Prifysgol Cymru online. (http://www.welsh-dictionary.ac.uk) (date of access 19 Sept. 2017).
Films
Evans, Marc. 2013. Y Gwyll [The dusk], Series 1, episode 1. Arrow Films.
Novels
Dafydd, Fflur. 2009. Y Llyfrgell [The library]. Talybont: Y Lolfa.
Davies, Daniel. 2011. Tair rheol anhrefn [Three rules of chaos]. Talybont: Y Lolfa.
Griffith, W. J. 1975 [1938]. Storïau'r Henllys Fawr [Stories of Henllys Fawr], Llandysul:
Gomer.
Gruffudd, Robat. 2012. Afallon [Avalon]. Talybont: Y Lolfa.
Jones, Bet. 2013. Craciau [Cracks]. Talybont: Y Lolfa.
Owen, Daniel. 1885. Rhys Lewis, gweinidog Bethel [Rhys Lewis, minister of Bethel].
Wrexham: Hughes and Son.
Prysor, Dewi. 2006. Brithyll [Trout]. Talybont: Y Lolfa.
Prysor, Dewi. 2010. Madarch [Mushrooms]. Talybont: Y Lolfa.
Roberts, Grace. 2010. Adenydd glöyn byw [Butterfly wings]. Llandysul: Gomer.
371
Software
Bangor
University.
Cysill
Ar-lein.
2018.
Bangor:
Bangor
University.
(http://www.cysgliad.com/cysill/arlein/) (date of access: 8 Jan. 2018).
Welsh Bible translations
Beibl Cymraeg Newydd Diwygiedig [Revised new Welsh Bible]. 2004. Cymdeithas y
Beibl, British & Foreign Bible Society. (https://www.bible.com/pl/versions/394bcn-beibl-cymraeg-newydd-diwygiedig-2004) (date of access: 15 May 2017).
beibl.net. 2015. Gobaith i Gymru. (http://www.beibl.net) (date of access: 15 May 2017).
Beibl William Morgan [William Morgan Bible]. 1588, 1612. Cymdeithas y Beibl, British
& Foreign Bible Society. (https://www.bible.com/pl/versions/287-bwm-beiblwilliam-morgan-1588-1620) (date of access: 15 May 2017).
372
Secondary sources:
Aarts, Bas. 1989. “Verb-preposition constructions and small clauses in English,” Journal
of Linguistics 25, 2: 277–290.
Aitchinson, John and Harold Carter. 2013. “The Welsh language – a false dawn?” Cambria Magazine online. (http://www.cambriamagazine.co.uk/the-welsh-languagea-false-dawn/) (date of access: 18 Apr. 2016).
Aitchison, J. W. and Harold Carter. 2000. Language, economy, and society: the changing
fortunes of the Welsh language in the twentieth century. Cardiff: University of
Wales Press.
Aitchison, John and Harold Carter. 2004. Spreading the word: the Welsh language 2001.
Talybont: Y Lolfa.
Altenberg, Evelyn P. and Robert M. Vago. 2004. “The role of grammaticality judgments
in investigating first language attrition. A cross-disciplinary perspective,” in:
Monika S. Schmid (ed.), First language attrition: interdisciplinary perspectives
on methodological issues. Amsterdam, Philadelphia: John Benjamins Publishing,
105–129.
Armstrong, Timothy Currie. 2012. “Establishing new norms of language use: The circulation of linguistic ideology in three new Irish-language communities,” Language
Policy 11, 2: 145–168.
Asmus, Sabine and Cormac Anderson. 2015. “Prolegomena to a study of Welsh vocalims,” in: Aleksander Bednarski, Paweł Czerniak and Maciej Czerniakowski
(eds.), New Perspectives in Celtic Studies. Newcastle upon Tyne: Cambridge
Scholars Publishing, 3–13.
Asmus, Sabine and Siôn Williams. 2014. “Welsh between stability and fragility: consolidated status but increasing linguistic insecurity a threat to diversity leading to
dangerous unity,” in: Sabine Asmus and Barbara Braid (eds.), Unity in Diversity,
Volume 2: Cultural and Linguistic Markers of the Concept. Newcastle upon Tyne:
Cambridge Scholars Publishing, 39–59.
Backus, Ad. 2004. “Convergence as a mechanism of language change,” Bilingualism:
Language and Cognition 7, 2: 179–181.
Baker, Colin. 1988. Key issues in Bilingualism and Bilingual Education. Clevedon: Multilingual Matters.
Baker, Colin. 1992. Attitudes and language. Clevedon: Multilingual Matters.
Baker, Colin. 2003. “Language planning: a grounded approach,” in: Jean-Marc Dewaele,
Hugo Baetens Beardsmore, Alex Housen and Li Wei (eds.), Bilingualism: Beyond
basic principles: Festschrift in honour of Hugo Baetens Beardsmore. Clevedon:
Multilingual Matters, 88–111.
373
Baker, Colin. 2010. “Increasing Bilingualism in Bilingual Education,” in: Delyth Morris
(ed.), Welsh in the twenty-first century. Cardiff: University of Wales Press, 61–
79.
Baker, Colin and Sylvia Prys Jones. 1998. Encyclopedia of bilingualism and bilingual
education. Clevedon: Multilingual Matters.
Ball, Martin J. and Nicole Müller. 1992. Mutation in Welsh. London and New York:
Routledge.
Ball, Martin J. and Enllli Môn Thomas. 2012. “LLARSP: A grammatical profile for
Welsh,” in: Martin J. Ball, David Crystal and Paul Fletcher (eds.), Assessing
grammar: The languages of LARSP. Clevedon: Multilingual Matters, 110–138.
Ball, Martin J., Tweli Griffiths and Glynn E. Jones. 1988. “Broadcast Welsh,” in: Martin
J. Ball (ed.), The use of Welsh: a contribution to sociolinguistics. Clevedon: Multilingual Matters, 182–199.
Bard, Ellen Gurman, Dan Robertson and Antonella Sorace. 1996. “Magnitude estimation
of linguistic acceptability,” Language 72, 1: 32–68.
Basic Skills Agency. 2003. “National standards for adult literacy and numeracy”.
(https://www.excellencegateway.org.uk/content/etf1352) (date of access: 31 May
2016).
Bazoli, Giulia. 2015. “The Welsh experience on GOV.UK: a qualitative research study”.
User Research Blog. (https://userresearch.blog.gov.uk/2015/09/15/the-welsh-experience-on-gov-uk-a-qualitative-research-study/) (date of access: 26 Apr. 2016).
BBC Cymru Fyw. 2015. “Trin geiriau” [Treating words]. (http://www.bbc.com/cymrufyw/35005693) (date of access: 13 Jun. 2016).
Beaufort Research. 2013. “Ymchwilio i ddefnydd iaith siaradwyr Cymraeg yn eu bywyd
bob dydd” [Exploring Welsh speakers' language use in their daily lives].
(http://gov.wales/docs/dcells/research/130808-wels-lang-research-cy.pdf) (date
of access: 14 Jan. 2016).
Bernstein, J. S. 1974. “On the semantics of certain English phrasal verbs and their rendering into Spanish,” Bilingual Review / La Revista Bilingüe 1, 1: 59–66.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad and Edward Finegan
(eds.). 1999. Longman grammar of spoken and written English. Harlow: Longman.
Blackledge, Adrian and Angela Creese. 2010. Multilingualism. A critical perspective.
London: Continuum.
Boers, Frank. 1996. Spatial prepositions and metaphor: a cognitive semantic journey
along the up-down and the front-back dimensions. Tübingen: Gunter Narr Verlag.
Bolinger, Dwight. 1971. The Phrasal Verb in English. Cambridge, Massachusetts: Harvard University Press.
374
Boon, Erin. 2014. Heritage Welsh: a study of heritage language as an outcome of minority
language acquisition and bilingualism.[Unpublished Ph.D. dissertation, Harvard
University.].
Bourdieu, Pierre. 1991. Language and symbolic power. (Translated by Gino Raymond
and Matthew Adamson). Cambridge: Polity Press.
Bright, William. 1998. “Social Factors in language change,” in: Florian Coulmas (ed.),
The Handbook of Sociolinguistics. Oxford, UK: Blackwell Publishing Ltd, 186–
204.
British Broadcasting Corporation. 2015. BBC Wales management review 2014/15.
(http://downloads.bbc.co.uk/annualreport/pdf/2014-15/BBCWalesNationsandRegions201415.pdf) (date of access: 29 May 2016).
Brooks, Simon. 2014. “The growth of English nationalism – Friend or foe? Institute of
Welsh Affairs Lecture, Llanelli National Eisteddfod, 2014”. Institute of Welsh Affairs website. (http://www.clickonwales.org/2014/09/the-growth-of-english-nationalism-friend-or-foe/) (date of access: 10 May 2016).
Brown, Callum G. 2014. Religion and society in twentieth-century Britain. London and
New York: Routledge.
Bryer, Huw. 2016. “Evaluation of workplace Welsh: the Welsh language skills management
diagnostic
tool”.
Cardiff:
Welsh
Government.
(http://dera.ioe.ac.uk/25882/1/160316-evaluation-workplace-welsh-en.pdf) (date
of access: 31 May 2016).
Buczek-Zawiła, Anita. 2014. “Everything must have its place: Accent accommodation in
Modern Welsh borrowings from English,” Studia Linguistica Universitatis Iagellonicae Cracoviensis, 131: 335–351.
Campbell, Harry. 2005. Gwybodiadur: a Welsh informationary: Welsh dictionaries
(gwybodiadur.tripod.com/dicts2.htm) (date of access: 1 May 2016).
Campoy Cubillo, Maria Carmen. 2002. “Phrasal and prepositional verbs in specialised
texts: a creative device,” Ibérica, 4: 95–111.
Cappelle, Bert, Yury Shtyrov and Friedemann Pulvermüller. 2010. “‘Heating up’ or
‘Cooling up’ the brain? MEG evidence that phrasal verbs are lexical units,” Brain
and Language 115, 3: 189–201.
Chambers, J. K. and Peter Trudgill. 1980. Dialectology. Cambridge: Cambridge University Press.
Chen, Junyu. 2007. “On how to solve the problem of the avoidance of phrasal verbs in
the Chinese context,” International Education Journal 8, 2: 348–353.
Chomsky, Noam. 1965. Aspects of the theory of syntax. Cambridge: M.I.T. Press.
Claridge, Claudia. 2000. Multi-word verbs in Early Modern English: A corpus-based
study. Amsterdam-Atlanta: Rodopi.
375
Clark, Stewart and Graham Pointon. 2016. The Routledge student guide to English usage:
A guide to academic writing for students. London, New York: Routledge.
Clyne, Michael. 1998. “Multilingualism,” in: Florian Coulmas (ed.), The handbook of
sociolinguistics. Oxford, UK: Blackwell Publishing Ltd, 301–315.
van Coetsem, Frans. 2000. A general and unified theory of the transmission process in
language contact. Heidelberg: Winter.
Cohen, Gaynor. 2007. “A sense of people and place: The chapel and language in sustaining Welsh identity,” in: Deborah Fahy Bryceson, Judith Okely and Jonathan Weber (eds.), Identity and networks: Fashioning gender and ethnicity across cultures. New York/Oxford: Berghahn Books, 91–102.
“Colloquial Welsh morphology”. 2018 In: Wikipedia: The free encyclopedia.
(https://en.wikipedia.org/wiki/Colloquial_Welsh_morphology) (date of access:
25 Jan. 2018).
Cooper, Robert L. 1989. Language planning and social change. Cambridge: Cambridge
University Press.
Cory, Hugh. 2009. Advanced writing with English in use. Oxford: Oxford University
Press.
Costa, James. 2017. “On the pros and cons of standardizing Scots: Notes from the North
of a small island,” in: Pia Lane, James Costa and Haley De Korne (eds.), Standardizing minority languages: competing ideologies of authority and authenticity
in the global periphery. New York: Routledge, 47–65.
Costa, James, Haley De Korne and Pia Lane. 2017. “Standardising minority languages:
reinventing peripheral languages in the 21st century,” in: Pia Lane, James Costa
and Haley De Korne (eds.), Standardizing minority languages: competing ideologies of authority and authenticity in the global periphery. New York: Routledge,
1–23.
Coupland, Nikolas and Michelle Aldridge. 2009. “Introduction: a critical approach to the
revitalisation of Welsh,” International Journal of the Sociology of Language
2009, 195: 5-13.
Courtney, Rosemary (ed.). 2000. Longman Dictionary of Phrasal Verbs. Harlow: Longman.
Crystal, David. 1999. “Is Welsh safe?” (http://www.davidcrystal.com/?id=4216&fromsearch=true#iosfirsthighlight) (date of access: 16 Mar. 2016).
Crystal, David. 2008. A dictionary of linguistics and phonetics. (6th ed.). Oxford: Blackwell Pub.
“Cymraeg ysgrifenedig” [Written Welsh]. 2015. In: Wicipedia. (https://cy.wikipedia.org/w/index.php?title=Cymraeg_ysgrifenedig&oldid=1666660) (date of access: 18 Jun. 2016).
376
Darwin, Clayton M. and Loretta S. Grey. 1999. “Going after the phrasal verb: An alternative approach to classification,” TESOL Quarterly 33, 1: 65–84.
Davies, Cennard. 1988. “Cymraeg Byw,” in: Martin J. Ball (ed.), The use of Welsh: A
contribution to sociolinguistics. Clevedon: Multilingual Matters, 200–212.
Davies, Howard, Graham Day and Angela Drakakis-Smith. 2010. “Attitudes to language
and bilingualism among English in-migrants to North Wales,” in: Delyth Morris
(ed.), Welsh in the twenty-first century. Cardiff: University of Wales Press, 148–
168.
Davies, Peredur Glyn Cwyfan. 2010. Identifying word-order convergence in the speech
of Welsh-English bilinguals. [Unpublished Ph.D. dissertation, Prifysgol Bangor
University].
Dehé, Nicole. 2002. Particle verbs in English: syntax, information structure, and intonation. Amsterdam, Philadelphia: John Benjamins Publishing.
Dempsey, Kyle B., Philip M. McCarthy and Danielle S. McNamara. 2007. “Using phrasal
verbs as an index to distinguish text genres,” in: David Wilson and Geoff Sutcliffe
(eds.), Proceedings of the Twentieth International Florida Artificial Intelligence
Research Society Conference, FLAIRS 2007. Menlo Park, California: The AAAI
Press, 217–222.
Deuchar, Margaret. 2005. “Minority language survival: code mixing in Wales,” in: James
Cohen, Kara T. McAlister, Kellie Rolstad, and Jeff MacSwan (eds.), ISB4: proceedings of the 4th International Symposium on Bilingualism. Somerville, MA:
Cascadilla Press, 608-620.
Deuchar, Margaret. 2006. “Welsh-English code-switching and the Matrix Language
Frame model,” Lingua 116, 11: 1986–2011.
Deuchar, Margaret and Peredur Davies. 2009. “Code switching and the future of the
Welsh language,” International Journal of the Sociology of Language 2009, 195:
15-38.
Deuchar, Margaret, Kevin Donnelly and Caroline Piercy. 2016. “‘Mae pobl monolingual
yn minority’: Factors Favouring the Production of Code Switching by Welsh–
English Bilingual Speakers,” in: Mercedes Durham and Jonathan Morris (eds.),
Sociolinguistics in Wales. London: Palgrave Macmillan, 209–239.
Deuchar, Margaret, Peredur Davies, and Kevin Donnelly. (2018). Building and Using the
Siarad Corpus: Bilingual Conversations in Welsh and English. Amsterdam: John
Benjamins.
Dorian, Nancy C. 1994. “Purism vs. Compromise in Language Revitalization and Language Revival,” Language in Society 23, 4: 479–494.
Doyle, Aidan. 2001. “Verb particle combinations in Irish and English,” in: John M. Kirk
and Dónall P. Ó Baoill (eds.), Language Links: The Languages of Scotland and
Ireland. Belfast: Queen’s University, 81–99.
377
Durham, Mercedes and Jonathan Morris. 2016. “An overview of sociolinguistics in
Wales,” in: Mercedes Durham and Jonathan Morris (eds.), Sociolinguistics in
Wales. London: Palgrave Macmillan, 3–28.
Eckert, Penelope and Sally McConnell-Ginet. 1992. “Communities of practice: Where
language, gender, and power all live,” in: Kira Hall, Mary Bucholtz and Birch
Moonwomon (eds.), Locating power: Proceedings of the 1992 Berkeley Women
and Language Conference. Berkeley: Berkeley Women and Language Group, 89–
99.
Edwards, John. 2010. Minority languages and group identity: cases and categories. Amsterdam ; Philadelphia: John Benjamins.
Evans, Daniel. 2015. “Teenage apathy for speaking Welsh reveals struggle to keep it as
a living language”. The Conversation. (http://theconversation.com/teenage-apathy-for-speaking-welsh-reveals-struggle-to-keep-it-as-a-living-language-48422)
(date of access: 18 Apr. 2016).
Evas, Jeremy and Daniel Cunliffe. 2016. “Behavioural economics and minority language
eservices — the case of Welsh,” in: Jonathan Morris and Mercedes Durham (eds.)
Sociolinguistics in Wales, Basingstoke: Palgrave MacMillan, 61-91.
Ferguson, Charles. 1972. “Diglossia,” in: Pier Paolo Gigliolo (ed.), Language and social
context. Selected readings. Harmondsworth: Penguin, 232–251.
Ferguson, Charles A. 1996. “Standardization as a form of language spread,” in: Thom
Huebner (ed.), Sociolinguistic perspectives: Papers on language in society, 1959–
1994. Oxford and New York: Oxford University Press, 189–199.
Fiedler, Sabine. 2012. “Der Elefant im Raum... The influence of English on German phraseology,” in: Cristiano Furiassi, Virginia Pulcini and Félix Rodríguez González
(eds.), The Anglicization of European lexis. Amsterdam ; Philadelphia: John Benjamins Publishing, 239–260.
Fife, James. 1986. “Literary vs. Colloquial Welsh: Problems of definition,” Word 37, 3:
141–151.
Filppula, Markku, Juhani Klemola and Heli Paulasto. 2008. English and Celtic in contact.
New York, London: Routledge Taylor & Francis Group.
Flohr, Harald. 2013. “The phenomenon of language contact – English influence on Irish
and Welsh found in the translations of Rowling’s Harry Potter and the Philosopher’s Stone,” Zeitschrift für Celtische Philologie 60, 1: 65–116.
Flores, Nelson and Jamie L. Schissel. 2014. “Dynamic Bilingualism as the norm: Envisioning a heteroglossic approach to standards-based reform,” TESOL Quarterly
48, 3: 454–479.
Fowkes, Robert A. 1945. “English idiom in modern Welsh,” WORD 1, 3: 239–248.
Fowkes, Robert A. 1948. “Prosody in the influence of English on Welsh,” WORD 4, 1:
37–41.
378
Fowkes, Robert A. 1949. “Initial mutation of loanwords in Welsh,” WORD 1, 5: 205–13.
Fowkes, Robert A. 1954. “On the gender of Early English loanwords in Welsh’,” WORD
1, 10: 66–70.
Fraser, Bruce. 1976. The verb-particle combination in English. New York: Academic
Press.
Furiassi, Cristiano, Virginia Pulcini and Félix Rodríguez González (eds.). 2012. The Anglicization of European lexis. Amsterdam ; Philadelphia: John Benjamins Publishing.
Gal, Susan. 2006. “Contradictions of standard language in Europe: Implications for the
study of practices and publics,” Social Anthropology 14, 2: 163–181.
García, Ofelia. 2009. Bilingual Education in the 21st Century: A Global Perspective. Oxford, UK: Wiley-Blackwell.
Gardner, D. and Mark Davies. 2007. “Pointing out frequent phrasal verbs: A corpus-based
analysis,” TESOL Quarterly 41, 2: 339–360.
Garrett, Peter, Nikolas Coupland and Angie Williams. 2003. Investigating language attitudes: Social meanings of dialect, ethnicity, and performance. Cardiff: University
of Wales Press.
Garrett, Peter. 2010. Attitudes to language. Cambridge: Cambridge University Press.
Gibbard, Alun. 2010. “Defnyddio’r Saesneg wrth siarad Cymraeg = dim problem” [Using
English while talking in Welsh = no problem], Golwg, 27 May 2010: 14–15.
Golwg360. 2014. “Pws: Ymateb ‘torcalonnus a phathetig’ Radio Cymru” [Pws: ’devastating
and
pathetic’
reaction
of
Radio
Cymru].
(http://golwg360.cymru/newyddion/cymru/162729-pws-ymateb-torcalonnus-aphathetig-radio-cymru) (date of access: 16 Jun. 2016).
Golwg360. 2018. “Radio Cymru 2 – ‘gorsaf newydd, nid arbrawf,’” [Radio Cymru 2 –
‘new
station,
not
an
experiment’].
(https://golwg360.cymru/newyddion/cymru/509404-radio-cymru-gorsafnewydd-arbrawf) (date of access: 29 Jan. 2018).
Görlach, Manfred. 2003. English words abroad. Amsterdam: John Benjamins Publishing
Company.
Gottlieb, Henrik. 2004. “Danish echoes of English,” Nordic Journal of English Studies 3,
2: 39–65.
Gottlieb, Henrik. 2012. “Phraseology in flux: Danish anglicisms beneath the surface,” in:
Cristiano Furiassi, Virginia Pulcini and Félix Rodríguez González (eds.), The Anglicization of European lexis. Amsterdam ; Philadelphia: John Benjamins Publishing, 169–198.
379
Gouws, Rufus H. 2009. “Dictionaries as Innovative tools in a New Perspective on Standardisation,” in: Henning Bergenholtz, Sandro Nielsen and Sven Tarp (eds.), Lexicography at a crossroads: Dictionaries and encyclopedias today, lexicographical
tools tomorrow. Bern: Peter Lang, 265–282.
Gries, Stefan T. 1999. “Particle movement: A cognitive and functional approach,” Cognitive Linguistics, 10: 105–145.
Grosjean, François. 2001. Life with Two Languages: An Introduction to Bilingualism.
Cambridge, Massachusetts: Harvard University Press.
Grosjean, François. 2010. Bilingual. Cambridge, Massachusetts: Harvard University
Press.
Grosjean, François. 2012. “An attempt to isolate, and then differentiate, transfer and interference,” International Journal of Bilingualism 16, 1: 11–21.
Gruffudd, Heini. 2015. “A oes rhywun yn darllen?” [Is anyone reading?], in: Delyth Prys
and Robat Trefor (eds.), Ysgrifau a chanllawiau cyfieithu [Essays and guidelines
on translation]. Caerfyrddin: Coleg Cymraeg Cenedlaethol, 115–126.
Gwyn Lewis, W. 2015. “Wales: Towards national strategy,” in: Peeter Mehisto and Fred
Genesee (eds.), Building bilingual education systems: forces, mechanisms and
counterweights. Cambridge: Cambridge University Press, 147–166.
Hartmann, R. R. K. and Gregory James. 2002. Dictionary of lexicography. New York:
Routledge.
Haspelmath, Martin. 2009. “Lexical borrowing: Concepts and issues,” in: Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the world’s languages: A comparative
handbook. Berlin: Mouton de Gruyter, 35–54.
Haugen, Einar. 1950. “The analysis of linguistic borrowing,” Language 26, 2: 210–231.
Haugen, Einar. 1953. The Norwegian language in America: A study in bilingual behavior.
Bloomington: Indiana University Press.
Heine, Bernd and Tania Kuteva. 2005. Language contact and grammatical change. Cambridge: Cambridge University Press.
Heinz, Sabine. 2003. Welsh dictionaries in the twentieth century: a critical analysis.
Muenchen: Lincom Europa.
Heinz, Sabine. 2010. “Major developments of the English literature in Ireland and Wales
and their interplay with history,” in: Liliana Sikorska (ed.), History is mostly repair and revenge. Discourses of/on history in literature in English. Frankfurt am
Main: Peter Lang, 25–42.
Hickey, Raymond (ed.). 2010. The Handbook of Language Contact. Oxford: WileyBlackwell.
380
Hickey, Tina M. 2007. “Children’s language networks in minority language immersion:
What goes in may not come out,” Language and Education 21, 1: 46–65.
Hickey, Tina M., Gwyn Lewis and Colin Baker. 2014. “How deep is your immersion?
Policy and practice in Welsh-medium preschools with children from different language backgrounds,” International Journal of Bilingual Education and Bilingualism 17, 2: 215–234.
Hincks, Rhisiart. 1993. “Cymraeg Glân Gloyw – Ysgrifennu a siarad Saesneg yn Gymraeg (i)” [Welsh of the best kind – Writing and speaking English in Welsh (i)],
Barn 369: 57.
Hincks, Rhisiart. 2000. The literary language as a scapegoat in Wales and Brittany. Aberystwyth: University of Wales.
Hincks, Rhisiart. 2007. Heb fenthygca cymaint a sill ar neb o ieithoedd y byd [Without
borrowing as much as a syllable from any of the world’s languages]. Aberystwyth:
Prifysgol Cymru; cyfres Astudiaethau Achlysurol, Rhif 3.
Hirata, Ryuichiro. 2012. Preposition stranding in Welsh. [Unpublished Ph.D. dissertation,
Prifysgol Bangor University.].
Hoffer, Bates L. 2002. “Language borrowing and language diffusion: An overview,” Intercultural Communication Studies 11, 2: 1–36.
Hoffer, Bates L. 2005. “Language borrowing and the indices of adaptability and receptivity,” Intercultural Communication Studies 14, 2: 53–72.
Huddlestone, Rodney and Geoffrey Pullum. 2000. Cambridge grammar of English. Cambridge: Cambridge University Press.
Hughes, J. Elwyn (ed.). 2011. Cyfansoddiadau a beirniadaethau Eisteddfod Genedlaethol 2011 [Winning literary entries and adjudications of the National Eisteddfod
of Wales 2011]. Wrecsam: Eisteddfod Genedlaethol Cymru.
Hunter, Jerry. 2006. “Welsh poetry [4] 20th century,” John Koch (ed.), Celtic culture: A
historical encyclopedia. Oxford: ABC-CLIO, 1775–1778.
ap Iwan, Emrys. 1939. “Plicio gwallt yr hanner Cymry” [Plucking off the hair of the half
Welsh],” in: David Myrddin Lloyd and Robert Ambrose Jones (eds.), Detholiad
o erthyglau a llythyrau Emrys ap Iwan : II. Llenyddol, ieithyddol [Selected articles
and letters of Emrys ap Iawn: II literary, linguistic]. Aberystwyth: Y Clwb Llyfrau
Cymreig, 103–125.
Johnson, Ian. 2013. “Revitalising Welsh in the Chubut Province, Argentina: The role of
the Welsh Language Project,” in: Michael Newton (ed.), Celts in the Americas.
Sydney: Cape Breton University Press, 145–160.
Jones, Berwyn Prys. 1988. “Official Welsh,” in: Martin J. Ball (ed.), The use of Welsh: A
contribution to sociolinguistics. Clevedon: Multilingual Matters, 175–182.
381
Jones, Ceri and Warren Chapman. 2000. “Dweud eich dweud – Arweiniad i Gymraeg
llafar ac idiomatig” [Have your say – a guide to spoken and idiomatic Welsh], in:
Steve Morris (ed.), Cylchrawn Ymchwil Cymraeg i Oedolion, 52–58.
Jones, Dafydd Glyn. 1988. “Literary Welsh,” in: Martin J. Ball (ed.), The use of Welsh:
A contribution to sociolinguistics. Clevedon: Multilingual Matters, 125–204.
Jones, Dafydd Glyn. 2013. “Welsh is a language just like any other”. Institute of Welsh
Affairs website. (http://www.clickonwales.org/2013/08/welsh-is-a-language-justlike-any-other/) (date of access: 31 May 2016).
Jones, Elin Haf Gruffydd. 2007. “The Territory of Television: S4C and the representation
of the ‘whole of Wales,’” in: Michael J. Cormack and Niamh Hourigan (eds.),
Minority language media: concepts, critiques and case studies. Multilingual Matters Series, Clevedon: Multilingual Matters, 188–211.
Jones, Hywel. 2008. “The changing social context of Welsh: A review of statistical
trends”. International Journal of Bilingual Education and Bilingualism, 11: 5, 541
– 557.
Jones, Hywel. 2010. “Welsh speakers: age profile and out-migration,” in: Delyth Morris
(ed.), Welsh in the twenty-first century. Cardiff: University of Wales Press, 118–
147.
Jones, Hywel M. 2012. “A statistical overview of the Welsh language”. Cardiff: Welsh
Language Board Cardiff. (http://www.poliglotti4.eu/docs/A_statistical_overview_of_the_Welsh_languagef2.pdf) (date of access: 18 Apr. 2016).
Jones, Mari C. 1998. Language obsolescence and revitalization: Linguistic change in two
sociolinguistically contrasting Welsh communities. Oxford: Clarendon Press.
Jones, Mari and Ishtla Singh. 2013. Exploring language change. London, New York:
Routledge.
Jones, M.C. 2006. “Jèrriais,” in: Keith Brown and Sarah Ogilvie (eds.), Concise encyclopedia of languages of the world. Oxford: Elsevier, 562–564.
Jones, Meirion Prys and Ceinwen Jones. 2014. “The Welsh language in education in the
UK, 2nd edition”. (https://files.eric.ed.gov/fulltext/ED565354.pdf) (date of access: 16 Mar. 2016).
Jones, Morris. 1979. “The present condition of the Welsh language,” in: Meic Stephens
(ed.), The Welsh Language Today. Llandysul: Gomer, 112–130.
Jones, Morris and Alan R. Thomas. 1977. The Welsh language: Studies in its syntax and
semantics. Cardiff: University of Wales Press for the Schools Council.
Jones, R.M. 2011. “A fydd Cymru Cymraeg mewn pryd?” [Will the Welsh speaking
Welsh be on time?]. (http://www.rmjones-bobijones.net/adennillyriaith.html)
(date of access: 23 May 2016).
382
Jones, Sylvia Prys. 2015. “Theori ac ymarfer cyfieithu yng Nghymru heddiw” [Theory
and practive of translation in today’s Wales], in: Delyth Prys and Robat Trefor
(eds.), Ysgrifau a chanllawiau cyfieithu [Essays and guidelines on translation].
Caerfyrddin: Coleg Cymraeg Cenedlaethol, 97–113.
Jørgensen, J. Normann. 2008. “Polylingual languaging around and among children and
adolescents,” International Journal of Multilingualism 5, 3: 161–176.
Kandler, Anne, Roman Unger and James Steele. 2010. “Language shift, bilingualism and
the future of Britain’s Celtic languages,” Philosophical Transactions of the Royal
Society B: Biological Sciences 365, 1559: 3855–3864.
Kaufmann, Judith. 2009. Cymdeithaseg cyfieithu. Dylanwad cyfieithu ar y pryd ar y defnydd o’r Gymraeg yng Ngwynedd [A sociology of translation. The effect of interpreting on the use of Welsh in Gwynedd]. [Unpublished Ph.D. dissertation, Prifysgol Bangor University.].
Keizer, Evelien. 2009. “Verb–preposition constructions in FDG,” Lingua 119, 8: 1186–
1211.
Kennedy, Arthur Garfield. 1920. The Modern English verb-adverb combination. Palo
Alto: Stanford University Press.
King, Gareth. 1993. Modern Welsh: a comprehensive grammar. New York: Routledge.
Klonowska, Marta. 2013. A critical analysis of Collins Spurrell Welsh Dictionary. [Unpublished M.A. thesis, Adam Mickiewicz University in Poznań.].
Konishi, Tomoshichi. 1958. “The growth of the verb-adverb combination in English: A
brief sketch,” in: Kazuo Araki, Taiichiro Egawa, Toshiko Oyama and Minoru Yasui (eds.), Studies in English grammar and linguistics: A miscellany in honour of
Takanobu Otsuka. Tokio: Kenyusha, 117–128.
Kroskrity, Paul V. 2016. “Language Ideologies and Language Attitudes”, in: Mark Aronoff (ed.) Oxford Bibliographies in Linguistics, available online (http://www.oxfordbibliographies.com/view/document/obo9780199772810/obo9780199772810-0122.xml?rskey=5s3wwB&result=64)
(date of access: 15 Aug. 2018)
Labov, William. 1966. The social stratification of English in New York City. Washington,
DC: Center for Applied Linguistics.
Labov, William. 1972. Sociolinguistic patterns. Philadelphia: University of Pennsylvania
Press.
Lakoff, George. 1987. Women, fire, and dangerous things. Chicago: University of Chicago Press.
Landry, Rodrigue and Réal Allard. 1991. “Can schools promote additive bilingualism in
minority group children?,” in: Lilliam M. Malavé and Georges Duquette (eds.),
Language, culture and cognition. Clavedon: Multilingual Matters, 198– 231.
383
Lane, Pia. 2014. “Minority language standardisation and the role of users,” Language
Policy 14, 3: 263–283.
Langacker, Ronald W. 1991. Foundations of cognitive grammar: descriptive application.
Volume 2. Stanford: Stanford University Press.
Langer, Nils and Agnete Nesse. 2012. “Linguistic purism,” in: Juan Manuel HernándezCampoy and Juan Camilo Conde-Silvestre (eds.), The handbook of historical sociolinguistics. Chichester: John Wiley & Sons, 607–625.
LaPiere, Richard Tracy. 1934. “Attitudes vs. Actions”. Social Forces, 13, 230–237.
Laufer, Batia and Stig Eliasson. 1993. “What causes avoidance in L2 learning: L1-L2
difference, L1-L2 similarity, or L2 complexity?” Studies in Second Language Acquisition, 15: 35–48.
Lewis, D. Geraint. 1995. Y llyfr berfau [The book of verbs]. Llandysul: Gwasg Gomer.
Lewis, E. Glyn. 1987. “Attitude to the planned development of Welsh,” International
Journal of the Sociology of Language 1987, 66: 11–26.
Lewis, Gwyn, Bryn Jones and Colin Baker. 2012. “Translanguaging: origins and development from school to street and beyond,” Educational Research and Evaluation
18, 7: 641–654.
Liebscher, Grit and Jennifer Dailey-O’Cain. 2017. “Contextualizing language attitudes:
An interactional perspective,” Language and Linguistics Compass 11:e12250, 114.
Lindner, Susan. 1981. A lexico-semantic analysis of verb-particle constructions with up
and out. San Diego: University of California.
Lipka, Leonhard. 1972. Semantic structure and word-formation: Verb-particle constructions in contemporary English. München: Wilhelm Fink Verlag.
Lipski, John M. 1990. The language of the Isleños: Vestigial Spanish in Louisiana. Baton
Rouge: Louisiana State University Press.
“Literary Welsh morphology”. 2018 In: Wikipedia: The free encyclopedia.
(https://en.wikipedia.org/w/index.php?title=Literary_Welsh_morphology&oldid=716103881) (date of access: 25 Jan. 2018).
Löffler, Marion. 2006. “Language (revival) movements in the Celtic countries [4]
Wales,” in: John Koch (ed.), Celtic culture: A historical encyclopedia. Oxford:
ABC-CLIO, 1099–1103.
Löffler, Marion. 2008. “English in Wales,” in: H. Momma and Michael Matto (eds.), A
companion to the history of the English language. Chichester, UK; Malden, MA:
Wiley-Blackwell.
Lyon, Jean and Nick Ellis. 1991. “Parental attitudes towards the Welsh language,” Journal of Multilingual and Multicultural Development 12, 4: 239–251.
384
Maio, Gregory R., James M. Olson, Mark M. Bernard and Michelle A. Luke. 2006. “Ideologies, Values, Attitudes, and Behavior,” in: John Delamater (ed.), Handbook of
Social Psychology. New York: Springer, 283–308.
Malmkjaer, Kirsten. 2004. Linguistics encyclopedia. New York: Routledge.
Mayr, Robert, Jonathan Morris, Ineke Mennen and Daniel Williams. 2017. “Disentangling the effects of long-term language contact and individual bilingualism: The
case of monophthongs in Welsh and English,” International Journal of Bilingualism 21, 3: 245–267.
McArthur, Tom (ed.). 1992. The Oxford Companion to the English language. New York:
Oxford University Press.
McEwan-Fujita, Emily. 2008. “Working at ‘9 to 5’ Gaelic: Speakers, contexts, and ideologies of an emerging minority language register,” in: Kendall A. King (ed.), Sustaining linguistic diversity: endangered and minority languages and language varieties. Washington, D.C.: Georgetown University Press, 81–94.
Mendis, Dushyanthi. 2010. “Formality in academic writing: The use/non-use of phrasal
verbs in two varieties of English,” in: Miguel F. Ruiz-Garrido, Juan Carlos Palmer
-Silveira, Inmaculada Fortanet-Gómez (eds.) English for professional and academic purposes. Amsterdam: Rodopi, 11–24.
Millar, Robert McColl. 2005. Language, nation, and power: an introduction. Basingstoke, Hampshire; New York: Palgrave Macmillan.
Milroy, James. 2001. “Language ideologies and the consequences of standardization,”
Journal of Sociolinguistics 5, 4: 530–555.
Milroy, James. 2006. “The Ideology of the Standard Language,” in: Carmen Llamas,
Louise Mullany and Peter Stockwell (eds.), The Routledge Companion to Sociolinguistics. London: Routledge, 133–139.
Milroy, James and Lesley Milroy. 1999. Authority in language: Investigating Standard
English. London: Routledge.
Morgan, D. Densil. 2000. “Yr Iaith Gymraeg a chrefydd” [The Welsh Language and religion], in: Geraint H. Jenkins and Mari A. Williams (eds.), “Eu hiaith a gadwant?” Y Gymraeg yn yr ugeinfent ganrif [“Did they keep their language?” Welsh
in the twentieth century]. Caerdydd: Gwasg Prifysgol Cymru, 358–380.
Morgan, Kenneth O. 1981. Rebirth of a nation: Wales, 1880-1980. Oxford: Oxford University Press.
Morris, Delyth. 2010a. “Introduction,” in: Delyth Morris (ed.), Welsh in the twenty-first
century. Cardiff: University of Wales Press, 1–6.
Morris, Delyth. 2010b. “Young people and their use of the Welsh language,” in: Delyth
Morris (ed.), Welsh in the twenty-first century. Cardiff: University of Wales Press,
80–98.
385
Morris, Jonathan. 2013. Sociolinguistic variation and regional minority language bilingualism: An investigation of Welsh-English bilinguals in North Wales. [Unpublished PhD dissertation, University of Manchester.].
Morris, Jonathan. 2014. “The influence of social factors on minority language engagement amongst young people: an investigation of Welsh-English bilinguals in
North Wales,” International Journal of the Sociology of Language 2014, 230: 65–
89.
Morris, Jonathan, Robert Mayr and Ineke Mennen. 2016. “The role of linguistic background on sound variation in Welsh and Welsh English,” in: Mercedes Durham
and Jonathan Morris (eds.), Sociolinguistics in Wales. London: Palgrave Macmillan, 241–271.
Morris Jones, Bob. 1993. Ar lafar ac ar bapur: cyflwyniad i’r berthynas rhwng yr iaith
lafar a’r iaith ysgrifenedig [In speech and on paper: an introduction to the relationship between the spoken and literary language]. Aberystwyth: Y Ganolfan
Astudiaethau Addysg.
Morris-Jones, John. 1921. An elementary Welsh grammar. Oxford: Clarendon Press.
Moseley, Christopher (ed.). 2010. Atlas of the world’s languages in danger. (3rd ed.)
Paris: UNESCO Publishing.
Müller, Nicole and Martin J. Ball. 2005. “Code-switching and diglossia,” in: Ball, Martin
J., Clinical sociolinguistics. Malden, MA; Oxford: Blackwell Publishing, 49–62.
Musk, Nigel. 2006. Performing bilingualism in Wales with the spotlight on Welsh: a study
of language policy and the language practices of young people in bilingual education = Perfformio dwyieithrwydd yng Nghymru gyda’r sbotolau ar y Gymraeg.
Linköping: Department of Language and Culture, Linköping University.
Musk, Nigel. 2010. “Code-switching and code-mixing in Welsh bilinguals’ talk: confirming or refuting the maintenance of language boundaries?,” Language, Culture and
Curriculum 23, 3: 179–197.
Musk, Nigel. 2012. “Performing bilingualism in Wales: Arguing the case for empirical
and theoretical eclecticism,” Pragmatics 22, 4: 651–669.
Myers-Scotton, Carol. 2002. Contact linguistics: Bilingual encounters and grammatical
outcomes. Oxford University Press.
Myers-Scotton, Carol. 2006. Multiple Voices: An Introduction to Bilingualism. Oxford:
Blackwell.
National Language Unit. 1978. Cyflwyno’r iaith lenyddol [Introducing the literary language]. Y Bontfaen: D.Brown a’i Feibion.
Navarro i Ferrando, Ignasi. 1998. A cognitive semantics analysis of the lexical units AT,
ON and IN, in English. [Unpublished Ph.D. dissertation, Universitat Jaume I, Castelló.].
386
Nelde, Peter H. and Peter J. Weber. 2000. “Forty years of evolution in contact linguistics,”
in: Olga Mišeska Tomić and Milorad Radovanović (eds.), History and perspectives of language study: Papers in honor of Ranko Bugarski. Amsterdam ; Philadelphia: John Benjamins Publishing, 185–204.
Nicoladis, Elena and Andra Gavrila. 2015. “Cross-linguistic influence in Welsh–English
bilingual children’s adjectival constructions,” Journal of Child Language 42, 04:
903–916.
Ó hIfearnáin, Tadhg and Noel Ó Murchadha. 2011. “The perception of Standard Irish as
a prestige target variety”, in: Tore Kristiansen and Nikolas Coupland (eds.),
Standard languages and language standards in a changing Europe. Oslo: Novus
Press, 97–104.
O’Dowd, Elizabeth M. 1998. Prepositions and particles in English : A discourse-functional account. New York: Oxford University Press.
Office for National Statistics. 2012. “2011 Census”. (http://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/bulletins/2011censuskeystatisticsforwales/2012-12-11#proficiency-in-welsh) (date of
access: 27 Apr. 2016).
O’Rourke, Bernadette. 2011. Galician and Irish in the European context: Attitudes towards weak and strong minority languages. Hampshire: Palgrave Macmillan.
O’Rourke, Bernadette. 2017. “Negotiating the standard in contemporary Galicia,” in: Pia
Lane, James Costa and Haley De Korne (eds.), Standardizing minority languages:
competing ideologies of authority and authenticity in the global periphery. New
York: Routledge, 84–100.
O’Rourke, Bernadette, Joan Pujolar and Fernando Ramallo. 2014. “New speakers of minority languages: the challenging opportunity – Foreword,” International Journal
of the Sociology of Language 2015, 231: 1–20.
Owen, John Aled. 2018. Factors influencing Welsh medium school pupils’ social use of
Welsh. [Unpublished Ph.D. dissertation, School of Education, Bangor University].
Pakendorf, Brigitte. 2013. “The interplay of language-internal variation and contact influence in language change,” in: Isabelle Léglise and Claudine Chamoreau (eds.),
The interplay of variation and change in contact settings. Amsterdam/Philadelphia: John Benjamins Publishing Company, 199–228.
Parafita Couto, Maria del Carmen, Margaret Deuchar and Marika Fusser. 2015. “How do
Welsh-English bilinguals deal with conflict? Adjective-noun resolution,” in: Gerald Stell and Kofi Yakpo (eds). Code-switching: Between Structural and Sociolinguistic Perspectives. Berlin: De Gruyter, 65-84.
Parina, Elena. 2010. “Loanwords in Welsh: Frequency analysis on the basis of Cronfa
Electroneg o Gymraeg,” in: Roncevic Brozovic, Maxim Fomin and Ranko
Matasovic (eds.), Celts and Slavs in Central and Southeastern Europe. Zagreb:
Studia Celto-Slavica III, 183–194.
387
Parliament of the United Kingdom. 1993. “Welsh Language Act”. (http://www.legislation.gov.uk/ukpga/1993/38/contents) (date of access: 27 Apr. 2016).
Parry-Williams, Thomas Herbert. 1923. The English element in Welsh. London: Honourable Society of Cymmrodorion.
Pelli, Mario. 1976. Verb-particle constructions in American English. Bern: Francke.
Percy, Carol and Ingrid Tieken-Boon van Ostade. 2016. “Prescription and tradition: Establishing standards across time and space,” in: Ingrid Tieken-Boon van Ostade
and Carol Percy (eds.), Prescription and tradition in language: Establishing
standards across time and space. Multilingual Matters, 1–22.
Perdek, Magdalena. 2010. “Getting through to phrasal verbs: A cognitive organization of
phrasal verb entries in monolingual pedagogical dictionaries of English,” in: Anne
Dykstra and Tanneke Schoonheim (eds.), Proceedings of the XIV Euralex International Congress. Afûk, Ljouwert: Fryske Akademy, 1390–1398.
Pérez-Milans, Miguel. 2015. “Language and identity in linguistic ethnography,” in: Sian
Preece (ed.), The Routledge handbook of language and identity. Abingdon:
Routledge: 83–97.
Phillips, John D. 2007. “Rare and endangered linguistic subsystems in Celtic and Welsh,”
Linguistica e Filologia 24: 153–182.
Poplack, Shana. 2017. Borrowing: Loanwords in the Speech Community and in the
Grammar. Oxford University Press.
Poplack, Shana and Nathalie Dion. 2012. “Myths and facts about loanword development,” Language Variation and Change 24, 03: 279–315.
Porto Requejo, Maria Dolores and Carmen Pena Díaz. 2008. “A cognitive approach to
some phrasal verbs in English for specific purposes,” Ibérica 16: 109-128.
Potowski, Kim. 2010. Language diversity in the USA. Cambridge: Cambridge University
Press.
Pritchard, Heledd. 2015. “New app gives Welsh speakers three different ways to read the
Bible
on
iPhone
or
iPad”.
Wales
Online.
(http://www.walesonline.co.uk/news/wales-news/new-app-gives-welsh-speakers-9804783) (date of access: 1 May 2016).
Prys, Delyth. 2009. “The development and acceptance of electronic resources for Welsh,”
in: Verena Lyding (ed.), Proceedings of the second colloquium on lesser used
languages and computer linguistics (LULCL II). Bozen-Bolzano: EURAC Research, 33–41.
Prys, Delyth. 2015. “Geiriaduron, termiaduron ac adnoddau defnyddiol eraill” [Dictionaries, terminology dictionaries and other useful resources], in: Delyth Prys and
Robat Trefor (eds.), Ysgrifau a chanllawiau cyfieithu [Essays and guidelines on
translation]. Caerfyrddin: Coleg Cymraeg Cenedlaethol, 39–72.
388
Prys, Delyth, Gruffydd Prys and Bryn Dewi Jones. 2016. “Cysill Ar-lein: A corpus of
written contemporary Welsh compiled from an on-line spelling and grammar
checker,” in: Nicoletta Calzolari and Khalid Choukri (eds.), Proceedings of the
Tenth International Conference on Language Resources and Evaluation (LREC
2016). Paris: European Language Resources Association (ELRA), 3261–3264.
Prys, Delyth, Gruffydd Prys and Dewi Bryn Jones. 2015. “Quantifying the use of digital
Welsh-language language resources,” in: Zygmunt Vetulani and Joseph Mariani
(eds.) Proceedings of the 7th Language & Technology Conference: Human language technologies as a challenge for computer science and linguistics. Poznań:
Fundacja Uniwersytetu im. Adama Mickiewicza w Poznaniu, 355–359.
Prys, Delyth and Robat Trefor (eds.). 2015. Ysgrifau a chanllawiau cyfieithu [Essays and
guidelines on translation]. Caerfyrddin: Coleg Cymraeg Cenedlaethol.
Pulcini, Virginia, Cristiano Furiassi and Félix Rodríguez González. 2012. “The lexical
influence of English on European languages: From words to phraseology,” in:
Cristiano Furiassi, Virginia Pulcini and Félix Rodríguez González (eds.), The Anglicization of European lexis. Amsterdam ; Philadelphia: John Benjamins Publishing, 1–26.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik. 1985. A comprehensive grammar of the English language. London: Longman.
Redknap, Catrin. 2006. “Welsh-medium and bilingual education and training: steps towards a holistic strategy”, in: Redknap, Catrin, W. Gwyn Lewis, Sian Rhiannon
and Williams Janet Laugharne. Welsh-medium and Bilingual Education. Bangor:
School of Education, University of Wales, 1–20.
Robert, Elen. 2009. “Accommodating ‘new’ speakers? An attitudinal investigation of L2
speakers of Welsh in south-east Wales,” International Journal of the Sociology of
Language 2009, 195: 93–116.
Robert, Elen. 2011. “Standardness and the Welsh language,” in: Tore Kristiansen and
Nikolas Coupland (eds.), Standard languages and language standards in a changing Europe. Oslo: Novus Press, 135–144.
Robert, Elen. 2013. Welsh lexical planning and the use of lexis in institutional settings.
[Unpublished Ph.D. dissertation, Cardiff University.].
Roberts, Megan Eluned and Richard Matthew Jones. 1974. Cyfeiriadur i’r athro iaith: [A
Reference book for teachers of language] Pt. 1. Cardiff: University of Wales
Press.
Rodgers, Elena. 2017. “Towards a typology of discourse-based approaches to language
attitudes,” Language & Communication 56: 82–94.
Ross, Malcolm. 2007. “Calquing and metatypy,” Journal of Language Contact 1, 1: 116–
143.
389
Ross, Malcolm D. 1996. “Contact-induced change and the comparative method: Cases
from Papua New Guinea,” in: Mark Durie and Malcolm D. Ross (eds.), The comparative method reviewed: regularity and irregularity in language change. New
York: Oxford University Press, 180–217.
Rottet, Kevin. 2000. “The calquing of phrasal verbs in language contact,” in: Julie Auger
and Andrea Word-Allbritton (eds.), The CVC of sociolinguistics: contact, variation, and culture. Working Papers in Linguistics, Bloomington, Indiana: IULC,
109–126.
Rottet, Kevin. 2005. “Phrasal verbs and English influence in Welsh,” Word 56, 1: 39–70.
Rottet, Kevin and Steve Morris. In preparation. Comparative Stylistics of Welsh and English. Cardiff: University of Wales Press.
Rudzka-Ostyn, Brygida. 2003. Word power: phrasal verbs and compounds : a cognitive
approach. Berlin: Mouton de Gruyter.
Rundell, Michael (ed.). 2007. Macmillan Phrasal Verbs Plus. Oxford: Macmillan Education.
S4C. 2014. “Dyfodol teledu Cymraeg: The future of Welsh language television”.
(https://www.s4c.cymru/keynote/pdf/e_s4c-future-of-welsh-language-broadcasting.pdf) (date of access: 16 Jun. 2016).
S4C. 2015. “S4C annual report & statement of accounts for the 12 month period to
31 March 2015”. (http://www.s4c.cymru/abouts4c/annualreport/acrobats/s4c-annual-report-2015.pdf) (date of access: 29 May 2016).
Sankoff, Gillian. 2001. “Linguistic Outcomes of Language Contact,” in: Peter Trudgill,
John Chambers and Natalie Schilling-Estes (eds.), Handbook of Sociolinguistics.
Oxford: Basil Blackwell, 638–668.
Sarnoff, Irving. 1970. “Social Attitudes and the Resolution of Motivational Conflict,” in:
Marie Jahoda, and Neil Warren (eds.), Attitudes. Harmondsworth: Penguin, 279–
284.
Sayers, Dave. 2009. Reversing Babel: declining linguistic diversity and the flawed attempts to protect it [Unpublished Ph.D. dissertation, University of Essex.].
Schieffelin, Bambi B., Kathryn A. Woolard, and Paul V. Kroskrity (eds.). 1998. Language ideologies: Practice and theory. New York: Oxford University Press.
Schiffman, Harold F. 1998. “Diglossia as a sociolinguistic situation,” in: Florian Coulmas
(ed.), The Handbook of sociolinguistics. Oxford, UK: Blackwell Publishing Ltd,
205–217.
Screen, Benjamin 2016. “What place for technology in the translation of Welsh?
Language technology, language planning and the professional translation of
Welsh”. Paper presented at 2nd Poznan Conference of Celtic Studies, Poznań,
Poland, 5-6 July 2016. http://orca.cf.ac.uk/92389/ (date of access: 21 July 2018).
390
Seliger, Herbert W. and Elana Goldberg Shohamy. 1989. Second language research
methods. Oxford: Oxford University Press.
Selleck, Charlotte. 2016. “Re-negotiating ideologies of bilingualism on the margins of
education,” Journal of Multilingual and Multicultural Development 37, 6: 551–
563.
Selleck, Charlotte. 2012. A comparative ethnographic study of students’ experiences and
perceptions of language ideologies in bilingual Welsh/English education: Inclusive policy and exclusionary practice [Unpublished Ph.D. dissertation, University
of Cardiff.].
Selleck, Charlotte. 2013. “Inclusive policy and exclusionary practice in secondary education in Wales,” International Journal of Bilingual Education and Bilingualism
16, 1: 20–41.
Siencyn, Lleucu. 2014. “Why we shouldn’t mind our Welsh language too closely”. Wales
Online. (http://www.walesonline.co.uk/whats-on/arts-culture-news/lleucu-siencyn-shouldnt-mind-welsh-7607347) (date of access: 18 Apr. 2016).
Silverstein, Michael. 1979. “Language structure and linguistic ideology,” in: Paul R.
Clyne, William F. Hanks and Carol L. Hofbauer, The Elements: A Parasession on
Units and Levels, Chicago: Chicago Linguistic Society, 193–247.
Sinclair, John (ed.). 1991. Collins COBUILD Dictionary of Phrasal Verbs. London:
HarperCollins Publishers.
Siyanova, Anna and Norbert Schmitt. 2007. “Native and nonnative use of multi-word vs.
one-word verbs,” IRAL - International Review of Applied Linguistics in Language
Teaching 45, 2: 119–139.
Snell, K. D. M. 1998. The regional novel in Britain and Ireland: 1800-1990. Cambridge:
Cambridge University Press.
Soares da Silva, Augusto. 2014. “Measuring and comparing the use and success of loanwords in Portugal and Brazil: A corpus-based and concept-based sociolectometrical approach,” in: Eline Zenner and Gitte Kristiansen (eds.), New perspectives
on lexical borrowing: Onomasiological, methodological and phraseological innovations. Boston/Berlin: Walter de Gruyter, 101–142.
Speelman, Dirk, Stefan Grondelaers and Dirk Geeraerts. 2003. “Profile-based linguistic
uniformity as a generic method for comparing language varieties,” Computers and
the Humanities 37, 3: 317–337.
Sroka, K.A. 1972. The syntax of English phrasal verbs. The Hague: Mouton.
Stammers, Jonathan R. and Margaret Deuchar. 2012. “Testing the nonce borrowing hypothesis: Counter-evidence from English-origin verbs in Welsh,” Bilingualism:
Language and Cognition 15, 03: 630–643.
Stenson, Nancy. 1997. “Language contact and the development of Irish directional phrase
idioms,” in: Anders Ahlqvist and Věra Čapková (eds.), Dán do oide: essays in
391
memory of Conn R. Ó Cléirigh. Dublin: Institiúid Teangeolaíochta Éireann, 559–
577.
Stewart, William A. 1968. “A sociolinguistic typology for describing national multilingualism,” in: Joshua A. Fishman (ed.), Readings in the sociology of language. The
Hague and Paris: Mouton & Co, 531–545.
Swales, John and Christine Feak. 2004. Academic writing for graduate students: essential
tasks and skills. Ann Arbor: University of Michigan Press.
Televnaja, Julja, Krista Bennett, Christian Hempelmann and Katrina E. Triezenberg.
2004. “Semantic representation of English phrasal verbs,” in: Mieczyslaw A.
Klopotek, Slawomir T. Wierzchon and Krzysztof Trojanowski (eds.), Intelligent
information processing and web mining: Proceedings of the International IIS:
IIPWM ́04 Conference Held in Zakopane, Poland, May 17-20, 2004. Berlin, Heidelberg, New York: Springer-Verlag, 465–469.
Thim, Stefan. 2012. Phrasal verbs: The English verb-particle construction and its history.
Berlin and New York: Walter de Gruyter.
Thomas, Alan R. 1973. The linguistic geography of Wales: a contribution to Welsh dialectology. Cardiff: University of Wales Press.
Thomas, Alan R. 1987. “A spoken standard for Welsh: description and pedagogy,” International Journal of the Sociology of Language 1987, 66: 99–114.
Thomas, Alan R. (ed.). 2000. The Welsh dialect survey. Cardiff: University of Wales
Press.
Thomas, Beth and Peter Wynn Thomas. 1989. Cymraeg, Cymrag, Cymreg: cyflwyno’r
tafodieithoedd [Welsh, Welsh, Welsh: introducing the dialects]. Caerdydd:
Gwasg Taf.
Thomas, Ceinwen H. 1967. “Adolygiad o Gymraeg Byw” [A review of Cymraeg Byw],
Llen Cymru 2, 9: 242–249.
Thomas, Ceinwen H. 1982. “Registers in Welsh”, International Journal of the Sociology
of Language 35: 87–115.
Thomas, George. 1991. Linguistic purism. London and New York: Longman.
Thomas, Peter Wynn. 1996. Gramadeg y Gymraeg [The grammar of Welsh]. Gwasg Prifysgol Cymru.
Thomason, Sarah Grey and Terrence Kaufman. 1988. Language contact, creolization and
genetic linguistics. Berkeley: University of California Press.
Thomason, Sarah. 2001. Language contact. Edinburgh: Edinburgh University Press.
Thorne, David A. 2000. Gafael mewn gramadeg [Get a grip on grammar]. Llandysul:
Gwasg Gomer.
392
Tomos, Megan. 2006. “Welsh prose literature [4] the short story,” in: John Koch (ed.),
Celtic culture: A historical encyclopedia. Oxford: ABC-CLIO, 1784–1785.
Trask, Robert Lawrence. 1993. A dictionary of grammatical terms in linguistics. London
and New York: Routledge.
Treffers-Daller, J. 2010. “Borrowing,” in: Mirjam Fried, Jan-Ola Östman, J.-O. and Jef
Verschueren (eds.) Variation and Change: Pragmatic Perspectives. John Benjamins, Amsterdam/Philadelphia, 17–35.
Treffers-Daller, Jeanine. 2012. “Grammatical collocations and verb-particle constructions in Brussels French: a corpus-linguistic approach to transfer,” International
Journal of Bilingualism 16, 1: 53–82.
Trosborg, Anna. 1997. Text typology and translation. Amsterdam, Philadelphia: John
Benjamins Publishing.
Tyler, Andrea and Vyvyan Evans. 2003. The semantics of English prepositions: Spatial
scenes, embodied meaning, and cognition. Cambridge: Cambridge University
Press.
Uribe-Jongbloed, Enrique. 2015. “Issues of identity in minority language media production in Colombia and Wales,” Journal of Multilingual and Multicultural Development: 1–13.
Urla, Jacqueline, Estibaliz Amorrortu, Ane Ortega and Jone Goirigolzarri. 2017. “Basque
standardization and the new speaker: Political praxis and the shifting dynamics of
authority and value,” in: Pia Lane, James Costa and Haley De Korne (eds.), Standardizing minority languages: competing ideologies of authority and authenticity
in the global periphery, New York: Routledge, 24–46.
Veselinović, Elvira. 2006. “How to put up with cur suas le rud and the bidirectionality of
contact,” in: Hildegard Tristram (ed.), The Celtic Englishes IV. Potsdam: Potsdam
University Press, 173–190.
VOICE Project. 2007. “VOICE Transcription Conventions [2.1]”. (http://www.univie.ac.at/voice/voice.php?page=transcription_general_information) (date of access: 15 May 2017).
Walsh, Olivia. 2016. Linguistic purism: Language attitudes in France and Quebec. Amsterdam, Philadelphia: John Benjamins Publishing Company.
Weibel, Birgit. 2007. “Phrasal verbs in learner English: a corpus-based study of German
and Italian students” [Unpublished Ph.D. dissertation, Albert-Ludwigs-Universität, Freiburg.].
Weinreich, Uriel. 1953. Languages in contact: findings and problems. The Hague: Mouton & Co.
Welsh Assembly Government. 2007. “Defining schools according to Welsh medium provision”.
(http://dera.ioe.ac.uk/7302/2/defining-schools-welsh-medi2_Redacted.pdf) (date of access: 22 May 1017).
393
Welsh
Government.
2006.
“Yr
Arddulliadur”
[The
style
(https://gov.wales/docs/dcells/publications/131128-translation-servicestyleguide-cy.pdf) (date of access: 21st July 2018).
guide].
Welsh Government. 2013. “Adolygiad o Gymraeg ail iaith yng Nghyfnodau Allweddol 3
a 4: Adroddiad ac Argymhellion” [Review of Welsh as second language in Key
Stages
3
and
4:
Report
and
Recommendations].
(https://gov.wales/docs/dcells/publications/130926-review-of-welsh-secondlancy.pdf ) (date of access: 15th August 2018).
Welsh Government. 2015. “WJEC GCE AS/A LEVEL in Welsh (First Language)”.
(http://www.wjec.co.uk/qualifications/qualification-resources.html?subject=Welsh&level=GCEASA) (date of access: 18 May 2016).
Welsh Government. 2017. “Cymraeg 2050: A million Welsh speakers”. (http://www.assembly.wales/laid%20documents/gen-ld11108/gen-ld11108-e.pdf) (date of access: 20 Jan. 2018).
Welsh Government and Welsh Language Commissioner. 2015. “National Survey for
Wales, 2013-14: Welsh Language Use Survey”. (http://gov.wales/statistics-andresearch/welsh-language-use-survey/?lang=en) (date of access: 4 Apr. 2016).
Wierszycka, Joanna. 2014. “Phrasal Verbs in learner English: a semantic approach. A
study based on a POS tagged spoken corpus of learner English,” Research in Corpus Linguistics 1, 1: 81–93.
Williams, Colin H. 2010. “From act to action in Wales”, in: Delyth Morris (ed.), Welsh
in the twenty-first Century, Cardiff: University of Wales Press, 36–60.
Williams, Eddie. 2009. “Language attitudes and identity in a North Wales town: “something different about Caernarfon”?”. International Journal of the Sociology of
Language 195: 63-91.
Williams, Glynn C. 1994. “Language standardisation, correctness and society,” Euskera:
Euskaltzaindiaren lan eta agiriak= Trabajos y actas de la Real Academia de la
Lengua Vasca= Travaux et actes de l’Academie de la Langue basque 39, 1: 81–
89.
Williams, Miriam Hedd. 2016. “Penderfyniad S4C i orfodi is-deitlau Saesneg ar ei
rhaglenni” [S4C’s decision to impose English subtitles in its programmes].
(http://poblcaerdydd.com/penderfyniad-s4c-i-orfodi-is-deitlau-saesneg-ar-eirhaglenni/) (date of access: 16 Jun. 2016).
Williams, Stephen J. 1980. Elfennau gramadeg Cymraeg [Elements of Welsh grammar].
Caerdydd: Gwasg Prifysgol Cymru.
Williamson, David. 2014. “Wales is no longer a nation of churchgoers but faith is alive”.
Wales Online. (http://www.walesonline.co.uk/news/news-opinion/wales-nolonger-nation-churchgoers-7025782) (date of access: 1 May 2016).
Willis. 2016. “Cyfieithu iaith y caethweision yn Uncle Tom’s Cabin a darluniadau o
siaradwyr ail iaith mewn llenyddiaeth Gymraeg,” [Translating the language of
394
slaves in Uncle Tom’s Cabin and images of second language speakers in Welsh
literature] Llên Cymru 39: 56–72.
Winford, Donald. 2003. An Introduction to Contact Linguistics. Malden: Blackwell.
Winford, Donald. 2010. “Contact and Borrowing,” in: Raymond Hickey (ed.), The Handbook of Language Contact. Oxford: Wiley-Blackwell, 170–187.
Winters, Margaret E. 2010. “Introduction: On the emergence of diachronic cognitive linguistics,” in: Margaret E. Winters, Heli Tissari and Kathryn Allan (eds.), Historical Cognitive Linguistics. Walter de Gruyter, 3–30.
Wooldridge, Dawn. 2011. “Gwella Cysill at defnydd cyfieithwyr: adnabod ymyrraeth gan
yr iaith Saesneg mewn testunau Cymraeg [Improving Cysill for use by translators;
recognizing interference by English in Welsh texts] [Unpublished MRes thesis,
Prifysgol Bangor University.].
Zelinsky-Wibbelt, Cornelia (ed.). 1993. The semantics of prepositions: From mental processing to natural language processing. Berlin, New York: Walter de Gruyter.
Zenner, Eline and Gitte Kristiansen (eds.). 2014. New perspectives on lexical borrowing:
Onomasiological, methodological and phraseological innovations. Boston/Berlin: Walter de Gruyter.
Zenner, Eline, Dirk Speelman and Dirk Geeraerts. 2012. “Cognitive sociolinguistics
meets loanword research: Measuring variation in the success of anglicisms in
Dutch,” Cognitive Linguistics 23, 4: 749 – 792.
Zwicky, Arnold. 1978. “On markedness in morphology,” Die Sprache 24, 2: 22–37.
395
Appendix A – The corpus
Corpus texts
Novels
Y Llyfrgell
Adenydd glöyn byw
Afallon
Tair rheol anhrefn
Craciau
Abbreviation
LLYFR
AGB
AFALL
TR
CRAC
Periodicals
Weekly magazines/newspapers
Golwg 44/2014
Golwg 45/2014
Golwg 46/2014
Golwg 47/2014
Y Cymro 11.07.2014
Y Cymro 18.07.2014
Y Cymro 25.07.2014
Y Cymro 8.08.2014
G 44
G 45
G 46
G 47
C 11/ 7
C 18/7
C 25/7
C 8/8
Monthlies, quarterlies, yearlies
Barn, 617/2014
Y Wawr, summer 2014
Y Faner Newydd, summer 2014
Y Casglwr, 113/2014
LOL 46/2014
Cristion Gorff 185, August 2014
Gwyliedydd 189, June/July 2014
BARN
W
FN
Cas
LOL
CR
Gwyl
Local bulletins (papurau bro)
Clonc 325, 2014(Ceredigion)
Cwlwm 364, 2014 (Carmarthenshire)
Yr Angor 369, 2014 (Ceredigion)
Y Ddolen 396, 2014 (Ceredigion)
Llais Aeron 371, 2014 (Ceredigion)
Blewyn Glas 405, 2014 (Powys)
Y Tincer 370, 2014 (Ceredigion)
CL
CW
YA
YD
LLA
BG
T
396
Sample of the corpus
397
Appendix B – The questionnaire
FFURFLEN GANIATÂD CYFRANOGWR
Ymchwilydd:
Marta Listewnik
Canolfan Astudiaethau Celtaidd,
Prifysgol Adam Mickiewicz yn Poznań, Gwlad Pwyl,
mlistewnik@wa.amu.edu.pl
Enw’r cyfranogwr:................................................................
Mae’r ymchwilydd a enwir uchod wedi rhoi gwybodaeth ddigonol i mi am yr ymchwil
rwyf wedi gwirfoddoli i gymryd rhan ynddi. Rwy’n deall bod hawl gennyf i dynnu'n ôl
o’r ymchwil unrhyw bryd. Rwy’n deall hefyd y perchir fy hawliau o ran peidio â datgelu
pwy ydwyf a chyfrinachedd.
Rwy’n cytuno i’r cyfweliad/trafodaeth gael ei recordio.
Llofnod y sawl sy’n cymryd rhan:
……………………………….
………………………………
Dyddiad
…………………………………
Gwneir copi deublyg o’r ffurflen hon. Dylai’r sawl sy’n cymryd rhan gadw un copi a’r
ymchwilydd y llall.
398
MANYLION PERSONOL
Enw
......................................................
Oedran
☐ 18-24 ☐ 25-34 ☐ 35-44 ☐ 45-54 ☐ 55-64 ☐ 65+
Rhyw ☐ gwryw
☐ benyw
Ble ydych chi’n byw?......................................................................................................
Statws gwaith
☐ llawn/rhan-amser
☐ di-waith/chwilio am waith
☐ myfyriwr/myfyrwraig llawn amser ☐ wedi ymddeol/arall ddim yn gweithio'n barhaol
Beth yw’r lefel uchaf o addysg ffurfiol y gwnaethoch chi ei gorffen?
☐
☐
☐
☐
☐
TGAU,
lefel-O/TAU,
Tystysgrif
Ysgol,
NVQ
lefel
1
neu
2
neu
Lefel A/AS, Tystysgrif Ysgol Uwch, GNVQ, Diploma Cenedlaethol BTEC, NVQ lefel 3 neu
Gradd Fagloriaeth, Diploma Addysg Uwch/Bellach, TAR (PGCE), HND, NVQ lefel 4 neu
Gradd
Feistr,
Doethuriaeth
/
PhD,
NVQ
lefel
5
neu
Dim un o’r uchod
gyfwerth
gyfwerth
gyfwerth
gyfwerth
Ers pryd ydych chi’n gallu siarad Cymraeg?
☐
Ers pan oeddwn i’n 2 flwydd oed neu’n iau
☐
Ers pan oeddwn i’n 4 blwydd oed neu’n iau
☐
Ers yr ysgol gynradd
☐
Ers yr ysgol uwchradd
☐
Dysgais i Gymraeg yn oedolyn
Ym mha ardal neu ardaloedd oeddech chi’n byw pan ddysgoch chi siarad Gymraeg?
……………………………………………………………………………………………………
Pa iaith ydych chi’n ei siarad bob dydd
gartref?
☐ Cymraeg yn unig
☐ Cymraeg yn bennaf
☐ Saesneg yn bennaf
yn y brifysgol/ yn y gwaith (os berthnasol)?
☐ Cymraeg yn unig
☐ Cymraeg yn bennaf
☐ Saesneg yn bennaf
☐ yr un faint o Gymraeg a Saesneg
☐ Saesneg yn unig
☐ iaith arall yn bennaf
☐ yr un faint o Gymraeg a Saesneg
☐ Saesneg yn unig
☐ iaith arall yn bennaf
yn eich cymdogaeth (e.e. mewn siop, mewn tafarn)?
☐ Cymraeg yn unig
☐ Cymraeg yn bennaf
☐ yr un faint o Gymraeg a Saesneg
☐ Saesneg yn bennaf
gyda ffrindiau agos?
☐ Cymraeg yn unig
☐ Saesneg yn bennaf
☐ Saesneg yn unig
☐ Cymraeg yn bennaf
☐ iaith arall yn bennaf
☐ yr un faint o Gymraeg a Saesneg
☐ Saesneg yn unig
☐ iaith arall yn bennaf
399
HOLIADUR
Enw.............................
Dyddiad...............
RHAN A
Dyma 12 brawddeg wedi’u tynnu o bapurau newydd neu wefannau cyhoeddus. Cylchwch bob gair/ymadrodd rydych chi’n ystyried yn addas i lenwi’r bwlch. Croeso i chi
rannu unrhyw sylwadau (e.e. opsiynau arall, gwahaniaethau yn yr ystyr).
1 Mae corff swyddog diogelwch o Gymru gafodd ei ladd yn Irac wedi _____________ i wledydd
Prydain.
☐ dychwelyd
☐ dod yn ôl
SYLWADAU:......................................................................................................................
2 Er gwaethaf y llawdriniaeth brys i'w wyneb bydd rhaid i’r bachgen gael sawl llawdriniaeth arall,
er mwyn ychwanegu croen newydd at ei wyneb wrth iddo _________________.
☐ dyfu i fyny
☐ dyfu
SYLWADAU:......................................................................................................................
3 Nid pawb, am wahanol resymau, sydd yn gwybod beth yn union yw eisteddfod a beth sydd yn
___________________ mewn eisteddfod.
☐ mynd ymlaen
☐ digwydd
SYLWADAU:......................................................................................................................
4 Chafodd neb ei anafu yn y digwyddiad, ac mae ymchwilwyr yn ceisio _____________ sut yn
union y dechreuodd y tân.
☐ darganfod
☐ gweithio allan
☐ gweithio mas
SYLWADAU:......................................................................................................................
5 Yn dilyn ymateb gwych i'w nofel gyntaf, mae sawl un wedi bod yn _____________ nofel nesaf.
☐ edrych ymlaen at ei ☐ edrych ymlaen i’w
☐ disgwyl yn eiddgar am ei
SYLWADAU:......................................................................................................................
6 Os oes rhieni ar gael i __________ yng ngardd yr ysgol ar brynhawn dydd Iau, Mawrth 17eg,
byddwn yn ddiolchgar iawn.
☐ helpu allan
☐ helpu mas
☐ helpu
☐ gynorthwyo
400
SYLWADAU:......................................................................................................................
7 Does dim ots a ydych chi'n siarad Cymraeg neu beidio, mae'r dosbarthiadau yn ddwyieithog a
byddwch yn ___________________ yn sydyn yn y ddwy iaith.
☐ dysgu’r caneuon ☐ pigo'r caneuon i fyny ☐ pigo'r caneuon lan ☐ cael crap ar y caneuon
SYLWADAU:......................................................................................................................
8 Mae’n rhaid i ni herio peryglon gwahaniaethu ac eithafiaeth o bob math a pheidio byth â sefyll
o’r neilltu pan fyddwn yn ___________________ a chasineb.
☐ dod ar draws rhagfarn
☐ taro ar ragfarn
SYLWADAU:......................................................................................................................
9 Er i’r cyhoeddwr gael cynnig grant gan y Cyngor Llyfrau ar gyfer cyhoeddi, penderfynodd
___________________ â'r fenter heb y grant.
☐ fwrw ymlaen
☐ gario ymlaen
☐ barhau
SYLWADAU:......................................................................................................................
10 Nid yw'r ateb mor syml â dweud wrth y ddwy ochr i ________________a thrafod y sefyllfa.
☐ eistedd
☐ eistedd i lawr
SYLWADAU:......................................................................................................................
11 Caiff 6,185 o oleuadau eu ___________________ o ganol nos tan y wawr.
☐ troi i ffwrdd
☐ diffodd
☐ troi bant
☐ troi off
SYLWADAU:......................................................................................................................
12 Mae angen cynnal mwy o ymarferiadau milwrol i roi neges glir nad ydym yn
___________________.
☐ sefyll o'r neilltu
☐ eistedd yn ôl
SYLWADAU:......................................................................................................................
401
RHAN B
Dyma setiau o frawddegau yn perthyn i bedair sefyllfa wahanol. Dychmygwch eich
bod yn y sefyllfa. A fyddech chi’n defnyddio’r ymadroddion mewn trwm yn y cyddestun?
Ticiwch y blwch pe byddech.
Wedyn, os ceir dau opsiwn tafodieithol, tanlinellwch yr hwn byddech chi’n ei ddewis.
SEFYLLFA 1. Rydych chi’n cael sgwrs â’ch ffrind am faterion pob dydd .
☐
Faint o bobol wnaeth droi i fyny/ droi lan i’r cyfarfod?
☐
Mae’r ffôn wedi rhedeg allan o / rhedeg mas o fatri eto.
☐
Wnaeth rhywun ffeindio allan/ ffeindio mas bod nhw’n gwerthu tocynnau o hyd.
☐
Bob tro wyt ti’n troi’r teledu ymlaen, maen nhw’n siarad am y gemau.
☐
Torrodd y car i lawr a doedd hi ddim yn gallu fforddio'r costau cynnal.
☐
All dy chwaer edrych ar ôl Mabon benwythnos yma?
SEFYLLFA 2. Rydych chi’n mewn cyfarfod cyhoeddus gydag aelod o Gynulliad yn trafod prob-
lemau eich ardal.
☐
Rwyf wedi siarad â’r staff Gwasanaethau Ambiwlans, ac wedi ffeindio allan/ ffeindio
mas byddai’n haws iddyn nhw ddefnyddio’r wybodaeth pe bai GPS ar gael.
☐
Mae henoed sy'n edrych ar ôl plant yn dioddef yn ariannol.
☐
Dyw hi ddim yn anarferol i bedair neu bump ambiwlans droi i fyny/ droi lan â chleifion
o fewn hanner awr.
☐
Mae’r llywodraeth wedi colli ysbrydoliaeth a rhedeg allan o/rhedeg mas o stem.
☐
Pan wnes i droi y radio lleol ymlaen, roedd pob cân yn Saesneg a’r unig Gymraeg a
glywais mewn mwy na hanner awr oedd un llinell mewn newyddion.
☐
Mae cerbydau yn torri i lawr a dydy casgliadau gwastraff bwyd ddim yn digwydd.
SEFYLLFA 3. Rydych chi’n ysgrifennu llyfr ffeithiol poblogaidd sy'n cyflwyno gwybodaeth am yr
amgylchedd.
☐
Mae llawer un yn pryderu byddwn yn rhedeg allan o/ rhedeg mas o'r olew.
☐
Pe gallwn ffeindio allan/ffeindio mas mwy am effaith y cemegyn, byddwn ni’n gwybod
sut i ymdopi â’r broblem.
402
☐
Os bydd ffermwyr yn troi'r dŵr ymlaen heb angen gwneud hynny, gall y tir wynebu
syched.
☐
Mae ffermio coco yn golygu edrych ar ôl y coed coco a chynaeafu'r codau, eplesu a
sychu'r ffa a'u pacio mewn sachau yn barod i gael eu troi yn eich hoff far o siocled.
☐
Mae’n werth cadw rheiddiaduron yn isel am gyfnodau hirach yn hytrach na’u troi i fyny/
troi lan am rai oriau.
☐
Mae sylweddau nad yw’n bosibl eu torri i lawr yn cael eu galw’n sylweddau hirbarhaus.
SEFYLLFA 4. Rydych chi’n paratoi taflen wybodaeth am wasanaeth iechyd yn y sir .
☐
Os ydych chi’n edrych ar ôl rhywun ag anghenion gofal a chymorth a fyddai'n methu
ymdopi heb eich help chi, rydych yn ofalwr.
☐
Gallwch ffeindio allan/ ffeindio mas rhagor am hyfforddiant a gweithgareddau dan y
ddolen isod.
☐
Rydym yn chwilio am wirfoddolwyr ag egni a syniadau – croeso mawr i bobl o bob oed
ac o bob rhan o'r ardal droi i fyny/ droi lan.
☐
Os ydych wedi rhedeg allan o/rhedeg mas o‘r feddyginiaeth, ewch i fferyllfa sydd
cymryd rhan yn y rhaglen.
☐
Gall cleifion droi’r larwm ymlaen pan fo angen.
☐
Mae gan staff yr ysbyty bŵer i symud cerbydau sydd wedi'u gadael neu sydd wedi torri
i lawr.
403
Appendix C – Interview Schedule
Regular questions
1.
Ydych chi’n ymwybodol o ddefnyddio geiriau Saesneg neu fenthyciadau Saesneg yn
eich Cymraeg llafar?
2.
Ydych chi’n siarad Cymraeg mewn sefyllfaoedd ffurfiol? (os yn berthnasol:) Oes
geiriau fyddech chi’n eu hosgoi wrth siarad yn ffurfiol?
3.
Ydych chi’n ysgrifennu yn Gymraeg? (os yn berthnasol:) Pa mor aml?
4.
Fyddech chi’n ceisio osgoi benthyciadau wrth ysgrifennu?
5.
Pan ydych chi’n ysgrifennu, ydy hi’n digwydd eich bod chi’n methu meddwl am air
Cymraeg dych chi angen? (os yn berthnasol:) Ydych chi’n chwilio amdano? (os yn
berthnasol:) Ble ydych chi’n chwilio?
6.
Ydych chi wedi gwneud y canlynol o fewn y mis diwethaf?
-
Darllen llyfr yn Gymraeg
-
Darllen cylchgronau neu bapurau newydd yn Gymraeg
-
Gwrando ar y radio yn Gymraeg
-
Gwylio’r teledu yn Gymraeg
-
Defnyddio meddalwedd Gymraeg ar y cyfrifiadur (e.e. Microsoft Office Cymraeg, Cysill neu Cysgeir)
-
Gwylio fideo yn Gymraeg ar-lein (e.e. YouTube, S4C Clic, BBC iPlayer)
-
Defnyddio Geiriadur Cymraeg (os ydych – pa un?)
404
Follow-up interview
1. Oes gynnoch chi unrhyw sylwadau ynglŷn â’r ymadroddion yn yr holiadur?
Other questions in the follow-up interview – themes
2. Asking participants to justify their lexical choices. The form of the question depended
on the overall results of the questionnaire e.g. “Ydy’r ymadroddion fel ‘na yn swnio’n
naturiol yn gyffredinol?”, “Felly, mae rhai o’r ymadroddion yn ymddangos yn anaddas
i chi yn y cywair mwy ffurfiol?”
3. Group-specific questions regarding:
– use of borrowings from English in classroom (teachers);
– use of borrowings from English in journalistic writing (journalists);
– use of borrowings from English in creative writing (writers);
– use of borrowings from English in in public space (social activists).
4. Asking about the speakers’ opinion on whether Welsh speakers should avoid the influence of English in their language or accept it.
5. Asking about the personal experience of the speakers as learners (L2 speakers) or
students at school/university.
6. Questions on standard Welsh. “Oes angen Cymraeg safonol?”?, “O ble mae Cymraeg
safonol yn dod?”
Note that due to time constrains it was not possible to cover all the themes with each of
the participants or in some cases the participants did not offer any opinions.
405