(Il)literacy and language change: Non-standard relative constructions in historical Basque

Language Va ia ion – Eu opean Pe spec i es VIII Edi ed by Hans Van de Velde Nanna Haug Hil on Remco Knooihuizen John Benjamins Publishing Company S udies in Language Va ia ion 25 Language Va ia ion – Eu opean Pe spec i es VIII S udies in Language Va ia ion (SILV) issn 1872-9592 Volume 25 Language Va ia ion – Eu opean Pe spec i es VIII Selec ed pape s om he Ten h In e na ional Con e ence on Language Va ia ion in Eu ope (ICLaVE 10), Leeuwa den, June 2019 Edi ed by Hans Van de Velde, Nanna Haug Hil on and Remco Knooihuizen The se ies aims o include empi ical s udies o linguis ic a ia ion as well as i s desc ip ion, explana ion and in e p e a ion in s uc u al, social and cogni i e e ms. The se ies will co e any ele an subdiscipline: sociolinguis ics, con ac linguis ics, dialec ology, his o ical lin- guis ics, an h opology/an h opological linguis ics. The emphasis will be on linguis ic aspec s and on he in e ac ion be ween linguis ic and ex alinguis ic aspec s — no on ex alinguis- ic aspec s (including language ideology, policy e c.) as such. Fo an o e iew o all books published in his se ies, please see benjamins.com/ca alog/sil Edi o s Pe e Aue Uni e si ä F eibu g F ans Hinskens Mee ens Ins i uu & Radboud Uni e si ei Nijmegen Paul Ke swill Uni e si y o Yo k Edi o ial Boa d A o An ila S an o d Uni e si y Gae ano Be u o Uni e si à di To ino Jenny Cheshi e Uni e si y o London Ka ie D age Uni e si y o Hawai‘i a Mānoa Ka a zyna Dziubalska-Kołaczyk Adam Mickiewicz Uni e si y in Poznań Jü g Fleische Philipps-Uni e si ä Ma bu g Pe e Gilles Uni e si y o Luxembou g B ian D. Joseph The Ohio S a e Uni e si y Johannes Kaba ek Uni e si ä Zü ich Pie e Muysken Radboud Uni e si y Nijmegen Pia Quis Uni e si y o Copenhagen Anne-Ca he ine Simon Uni e si é ca holique de Lou ain Sali A. Tagliamon e Uni e si y o To on o Øys ein Alexande Vangsnes UiT The A c ic Uni e si y o No way Juan A. Villena Ponsoda Uni e sidad de Málaga Language Va ia ion – Eu opean Pe spec i es VIII Selec ed pape s om he Ten h In e na ional Con e ence on Language Va ia ion in Eu ope (ICLaVE 10), Leeuwa den, June 2019 Edi ed by Hans Van de Velde F yske Akademy & U ech Uni e si y Nanna Haug Hil on Uni e si y o G oningen Remco Knooihuizen Uni e si y o G oningen John Benjamins Publishing Company Ams e dam / Philadelphia doi 10.1075/sil .25 Ca aloging-in-Publica ion Da a a ailable om Lib a y o Cong ess: lccn 2021010868 (p in ) / 2021010869 (e-book) isbn 978 90 272 0885 9 (Hb) isbn 978 90 272 5982 0 (e-book) © 2021 – John Benjamins B.V. No pa o his book may be ep oduced in any o m, by p in , pho op in , mic o ilm, o any o he means, wi hou w i en pe mission om he publishe . John Benjamins Publishing Company · h ps://benjamins.com 8 TM The pape used in his publica ion mee s he minimum equi emen s o he Ame ican Na ional S anda d o In o ma ion Sciences – Pe manence o Pape o P in ed Lib a y Ma e ials, ansi z39.48-1984. Table o con en s In oduc ion 1 Remco Knooihuizen, Nanna Haug Hil on and Hans Van de Velde Chap e 1 The ola ile linguis ic shape o ‘Town F isian’/‘Town Hollandic’ 11 A jen P. Ve sloo Chap e 2 Is he e an in e language speech accep abili y de ici ? 35 Rias an den Doel and Ad iaan Walpo Chap e 3 Re isi ing he owel me ge s o Eas Anglia: Co ela ions o moan, mown and goose 53 Ke i-Ann Bu che Chap e 4 Modeling egional a ia ion in oice onse ime o Ju landic a ie ies o Danish 79 Rasmus Puggaa d Chap e 5 “O ganically Ge man”? Changing ideologies o na ional belonging 111 Jane M. Fulle Chap e 6 Explo ing an app oach o modelling lec al cohe ence 135 Ka en V. Beaman Chap e 7 “I’m dead posh in school”: A i udes and linguis ic beha iou o Me seyside adolescen s 161 Rachel By ne Chap e 8 Benim: A new p onoun in Swedish 181 Na han J. Young i Language Va ia ion – Eu opean Pe spec i es VIII Chap e 9 Iden i ica ion o clus e s o lexical a eas using geog aphical ac o s: A case s udy in he Occi an language a ea 209 Clémen Chagnaud, Guylaine B un-T igaud and Philippe Ga a Chap e 10 (Il)li e acy and language change: Non-s anda d ela i e cons uc ions in his o ical Basque 227 Do o a K ajewska and Eneko Zuloaga Chap e 11 Dialec con ac in he owel sys em o Mišó ika Cappadocian 247 Nicole Vassalou, Dimi is Papazacha iou and Ma k Janse Chap e 12 Leade s o language change: Mac o and mic o pe spec i es 269 Me edi h Tamminga Chap e 13 E hnic a ia ion in eal ime: Change in Aus alian English diph hongs 291 James G ama, Ca he ine E. T a is and Simon Gonzalez Index 315 Chap e 10 (Il)li e acy and language change Non-s anda d ela i e cons uc ions in his o ical Basque Do o a K ajewska and Eneko Zuloaga Uni e si y o he Basque Coun y UPV/EHU In his pape we examine a co pus o six een h- o nine een h-cen u y Basque p i a e le e s and adminis a ion documen s om a his o ical sociolinguis ic poin o iew. Because o he diglossic si ua ion, such ex s a e a e bu aluable in he his o ical co pus o Basque. In pa icula , we analyse ela i e clauses wi h he p onoun zein ‘which’. The cons uc ion was bo owed om Romance by highly li e a e bilinguals, bu hen sp ead, especially h ough o mulaic language, o less li e a e w i e s oo. We ocus on he de elopmen o non-s anda d a i- an s o his ela i e clause, which a e common in ou co pus, bu no ound in p in ed li e a y ex s. We a gue ha he sociolinguis ic con ex was c ucial in he eme gence, sp ead and syn ac ic change o he ela i e cons uc ion. Keywo ds: his o ical sociolinguis ics, syn ax, ela i e clauses, li e acy, Basque 1. In oduc ion In his pape we adop he iewpoin o his o ical sociolinguis ics o e lec on Basque w i ing, w i en language, and syn ac ic change unde language con ac . Ou analysis is based on a co pus o six een h- o nine een h-cen u y Basque le e s and adminis a i e documen s. P ocesses desc ibed he e ha e pa allels in o he Eu opean languages, bu he Basque case o e s he oppo uni y o obse e hem in a pa icula ly in e es ing diglossic con ex , whe e a p es igious and s anda dised language (F ench o Spanish) was he usual means o w i en communica ion, and ano he language (Basque) had no ye de eloped a s anda d. We ocus on ela i e clauses (RC), which a e complex cons uc ions, ypical o highe egis e s, and p one o cause p oblems o no -so-p o icien w i e s. The mos common ela i isa ion s a egy in mode n Basque and in mos his o ical sou ces is p enominal, and di e s om ela i e clauses ound in he su ounding h ps://doi.o g/10.1075/sil .25.10k a © 2021 John Benjamins Publishing Company 228 Do o a K ajewska and Eneko Zuloaga Romance languages. Howe e , h ough con ac Basque inno a ed a cons uc ion wi h he p onoun zein ‘which’, which is he objec o ou s udy. Being conside ed an unin e es ing calque, i has ecei ed li le a en ion, and i s use ou side he “ca- nonical” sou ces (such as p in ed books) has no been s udied. We show ha he ela i e p onoun zein is common in ou co pus and ha i became a ea u e o w i en egis e s. Mo eo e , we p opose ha se e al non-s anda d a ian s o he cons uc ion demons a e ha he di usion o syn ac ic change happened in w i - en in e ac ions be ween w i e s, o example, by means o exchanges o le e s, and mainly h ough o mulaic language. The pape is s uc u ed as ollows. Sec ion2 in oduces ela i isa ion s a egies in Basque and Romance. Sec ion3 desc ibes he sociolinguis ic si ua ion o Basque be o e he wen ie h cen u y and he in luence i had on w i ing. Sec ion4 p esen s ou co pus s udy. Sec ion5 p o ides an explana ion o he syn ac ic changes ob- se ed in he co pus and compa es hem o phenomena a es ed in o he Eu opean languages. 2. Basque and Romance ela i isa ion s a egies Basque has se e al ini e ela i e cons uc ions, which di e in he ollowing ea- u es (De Rijk 1972; Oyha çabal 1987; Oyha çabal 2003; K ajewska 2017) (see Table1): a. embedded s. apposi i e RC b. posi ion o he RC: pos nominal s. p enominal c. ea men o he ela i ised noun: gap s. p onoun d. choice o subo dina o : -en, bai - o bo h We hus ha e p enominal (1), pos nominal (2), apposi i e (3), and zein ela i es (4), illus a ed below wi h examples om classical Basque sou ces: (1) [gizon-ek e an dukei e-n] hi z al e guzi-a-z man-de .e g.pl say aux.3pl>3sg-sub wo d ain all-de -ins ‘all he ain wo ds man can say’ (Leiza aga, 1571) (2) a bo e [ uk u on-ik egi- en ez- u-en guzi-a] ee ui good-pa make-ip neg-aux.3sg>3sg-sub all-de .abs ‘e e y ee which does no p oduce good ui ’ (Leiza aga, 1571) (3) zuhai z al e -a [ ui u- ik iasai- en ez- u-en-a] ee lazy-de .abs ui -pa gi e-ip neg-aux.3sg>3sg-sub-de .abs ‘a lazy ee which does no gi e ui ’ (Axula , 1643) Chap e 10. (Il)li e acy and language change 235 4.3 Non-s anda d a ian s The mos in e es ing aspec o he zein ela i e cons uc ion in ou co pus is i s non-s anda d a ian s, which a e no ound in canonical sou ces. They in ol e a o m o he p onoun zein ‘which’ unc ioning as an in a iable ela i e pa icle. Con a y o he s anda d cons uc ion, he p onoun does no ake he case co e- sponding o he ela i ised a gumen . In he mos common non-s anda d a ian , he inessi e p onoun– usually he inde ini e zeine an and some imes he de ini e singula zeinean– ela i ises all syn ac ic unc ions, o example, he absolu i e subjec : (6) Isk iba- zen dauzki zu bi le o hauk zu- i gu e w i e-ip aux.1sg>3pl<2sg wo line hese you-da ou osasun-a- en ma ka- ze-ko, [zein-e an hani z on heal h-de -gen in o m-nmlz- m which-inde .ines e y good bai- a Jainko-a- i eske -ak] sub-be.3sg god-de -da hank-de .abs.pl ‘I am w i ing hese wo lines o you o in o m you on my heal h, which, hank God, is e y good.’ (LeDauphin-45, p i a e le e , 1757) In o he a ian s we ind zeinak, ambiguous be ween absolu i e plu al o e ga i e singula (used o ela i ise absolu i e singula in (7)), o he absolu i e de ini e zeina. (7) Ego i dina ba Lan zekane u- ekin, be ze-a hi e send aux.1sg>3sg<2sg one Lan zekane u-soc o he -de .abs you kusin-a de un u-ba- ekin, [zen-ak e o i cousin-de .abs dead-de -soc which-de .abs.pl/de .e g come bai -zen Misel-ekin Baiona- a , gaso-a] sub-aux.ps .3sg Misel-soc Bayonne-adl poo -de .abs ‘I ha e sen you one wi h Lan zekane u, and ano he one wi h you la e cousin, who came wi h Misel o Bayonne, poo guy.’ (LeDauphin-20, p i a e le e , 1757) The e a e also clauses ha we label as “non- ela i e” because wha looks like a sub- o dina e ela i e does no modi y any hing. Fo example, in (8), as he ansla ion shows, he syn ax is a he odd: i ‘in which’ e e s o he le e , a pa o he sen ence is missing (we would expec ‘in which we w o e ha …’). 236 Do o a K ajewska and Eneko Zuloaga (8) Ma ka- zen da o azu ez- uzu-la Qanigon-ekin w i e-ip aux.2sg>3sg<1sg neg-aux.2sg>3sg-compl Qanigon-soc iza u le a- ik, bainan egin gin uben, [zein-e an ha e le e -pa bu make aux.ps .1pl>3sg which-inde .ines aizpa Nana Mia i z-en bai- ugu neska o Bod i- ekin] sis e Nana Mia i z-ines sub-ha e.1pl>3sg maid Bod i-soc ‘You w i e ha you ha en’ go ou le e s om Qanigon, bu we w o e one, in which sis e Nana is maid in Bia i z wi h Bod i.’ (LeDauphin-38, p i a e le e , 1757) Non-s anda d a ian s a e mos common in p i a e le e s, whe e o e 40% o he examples a e o his kind (Table4). Table4. Non-s anda d zein ela i e clauses in di e en ex ypes Tex ype Tokens o zein RC Non-s anda d a ian s (%) Adm. doc. 55 20.0 No a ies 37 18.9 Adm. le e s 70 7.1 P i a e le e s 65 43.1 Table5 p esen s occu ences o each a ian (including he non- ela i e), and he numbe o sou ces in which i appea s. The a ian wi h zeine an is he mos wide- sp ead non-s anda d op ion. Table5. F equencies o a ian s o zein ela i e Zein RC ype Tokens Tex s S anda d 176 88 Non-s anda d zeine an 28 22 Non-s anda d zeina/zeinak 15 8 Non- ela i e 8 4 The equency o non-s anda d ela i es seems highes in he eigh een h cen u y, and i dec eases in he nine een h cen u y. This migh e lec an inc ease in li e acy and knowledge o Romance languages, which would cause using he s anda d zein ela i e. Non-s anda d a ian s we e unsuccess ul inno a ions: hey a e no a es ed in he wen ie h cen u y.3 3. Howe e , in mode n w i en Basque he e is a non-s anda d ela i e cons uc ion wi h he in a ian pa icle non ‘whe e’, which seems no unlike he a ian wi h zeine an, bu i s de elop- men and use equi e u he s udy. Chap e 10. (Il)li e acy and language change 237 4.4 Zein ela i es and o mulaic language Zein ela i es in ou co pus a e o en ound in epis ola y o mulae, which a e de- ined by W ay (2005: 9) as “a sequence, con inuous o discon inuous, o wo ds o o he elemen s, which is, o appea s o be, p e ab ica ed: ha is, s o ed and e ie ed whole om memo y a he ime o use, a he han being subjec o gene a ion o analysis by he language g amma ”. Following he classi ica ion by W ay (2005) and Ru en and Van de Wal (2012), wo classes a e ele an he e: ex -cons i u i e (e.g. le e openings) and in e sub- jec i e o mulae (e.g. g ee ings o exp essions o heal h). In a small co pus such as ou s i is di icul o decide wha coun s as a o mula. We hus only include wo classes equen in he da a: (a) le e openings (“I ha e ecei ed you le e in which …” and i s a ia ions, see Example(9)), and (b) he exp ession o he sende ’s heal h and hope ha he ecei e is al igh (“we ha e heal h, which we hope you also ha e”, “I wan o le you know abou my heal h, which is good” and simila ph ases, Example(10)). (9) Rezebi u du a zo zu-k esk iba u ka a, ecei e aux.1sg>3sg yes e day you-e g w i e le e [zoin- an e a- en bai- uzu jen e paub i-ak which-inde .ines say-ip sub-aux.2sg>3sg people poo -de .abs.pl di e-la zu- i eska- zen du e-n-ak be.3pl-compl you-da ask-ip aux.3pl>3sg-sub-de .abs.pl gas u-ak] paymen -de .abs.pl ‘Yes e day I ecei ed he le e you w o e, in which you say ha i is he poo people ha asks you o money.’ (adminis a ion le e , 1817; Camino (2012)) (10) Osasun dugu Janko-a- i eske , [zein-e an dezi a- zen heal h ha e.1pl>3sg god-de -da hank which-inde .ines wish-ip bai-kinduke zu i-a hala ba-li z] sub-aux.po .1pl>3sg you -de .abs so cond-be.hyp.3sg ‘We ha e heal h, hank God, which we wish you also ha e.’ (LeDauphin-46, p i a e le e , 1757) Table6 shows ha he e is a co ela ion be ween he equency o occu ence in a o mula and le e ype, on one hand, and ype o ela i e, on he o he hand. In p i- a e le e s ela i es appea mo e equen ly in o mulae han in adminis a ion le - e s. In Le Dauphin collec ion, his p opo ion is e en highe (39%). Non-s anda d ela i es a e also mo e common as pa o a o mula han he s anda d cons uc ion. 238 Do o a K ajewska and Eneko Zuloaga Table6. Rela i e clauses in o mulae acco ding o he ex ype and cons uc ion Con ex Tokens Fo mulae Le e ype Adminis a i e 70 12.9% P i a e 65 29.2% Rela i e clause ype S anda d 102 14.7% Non-s anda d 33 39.4% 5. Discussion In his pape we ocus on ex s w i en in Basque despi e he ad e se diglossic si - ua ion. Thei au ho s usually belonged o uppe social s a a and we e augh how o w i e in Romance. Thus, li e acy in Basque depended on li e acy in ano he lan- guage. This ac is pa en in spelling: o ins ance, w i e s o en s uggled o ende Basque sibilan s using Romance sys ems, which ha e ewe sibilan phonemes. He e we explo e he consequences i had on syn ax. W i ing does no usually happen in oid and knowing how o w i e is no lim- i ed o he command o he code. I also p esupposes knowledge ela ed o di e en aspec s o w i ing, such as gen e knowledge and knowledge ela ed o he social con ex o w i ing (Beau o & Iñes a 2014). Fo example, w i ing a le e equi es unde s anding o he con en ions uling his ype o ex s and o how he gi en communi y usually w i es hem, as well as wha is hough o be co ec and wha no . Gen es can in ac be unde s ood as “socially and cul u ally pa e ned ways o engaging in ac i i y” (Lillis 2013: 70). As ega ds gen es analysed he e, in e ms o Koch and Oes e eiche (2012 [1985]), p i a e le e s a e close o he language o immediacy, and o mal doc- umen s p oduced by di e en ins i u ions a e examples o language o dis ance: hey cons i u e a monologue, he e is dis ance be ween he pa ne s, and opics a e ixed. Adminis a i e co espondence alls somewhe e in be ween. Ne e heless, he ex s in ou co pus we e w i en be o e he language was s anda dised, and as a esul hey exhibi dialec al ea u es and cha ac e is ics desc ibed o he language o immediacy by Koch & Oes e eiche (2012 [1985]: 454–455): cong uence e o s, p oblems wi h segmen a ion, low ype- oken a io, o e all lexical po e y, hesi a- ion phenomena o ma ke s o co ec ion. Mo eo e , ea u es o illi e a e w i ing (see Mon gome y 1995) a e common: p oblems wi h spelling and wo d di ision o missing punc ua ion ma ks. This is because hose ex s we e w i en in Basque by people who we e li e a e in Romance. In a complex sociolinguis ic en i onmen , hey could no ely on a ully s anda dised language and had o use a dialec al a- ie y. Ne e heless, hey looked o models o cons uc hei w i ings. Chap e 10. (Il)li e acy and language change 239 Bilinguals could eso o no ms o o he languages. Fo ins ance, when a cle k p epa ed a Basque documen , he could ansla e om Spanish o F ench he ypical ph ases o adminis a i e language, and, in gene al, y o copy he s yle o he ex . The same holds o o he con en ions, o example he use o epis ola y o mulae. W i e s also ea ed Basque ex s as models. Fo his o be possible, hough, a ce ain adi ion o Basque w i ing had o exis in he social en i onmen . Fo ins ance, when composing a le e , he w i e could euse agmen s o Basque le - e s, p o ided he had access o any. In gene al, lea ning o co espondence-w i ing con en ions usually happened in p ac ice: by eading o lis ening o le e s and hen imi a ion (see an de Wal & Ru en 2013 and e e ences he ein). This explains he simila i y be ween ex s w i en by di e en people. The zein ela i e cons uc ion is hough o ha e de eloped in Basque due o language con ac (La i e 1991 [1944]; Haase 1992; T ask 1998). This migh be a change om abo e sp eading in w i ing om mo e li e a e bilinguals (usually be- longing o highe social classes) o less p o icien ones. Simila explana ions ha e been p oposed o changes in ela i isa ion s a egies in a numbe o languages (Romaine 1982; Rissanen 1999; Ru en & an de Wal 2014, among o he s). In Basque, his change could happen h ough imi a ing he Romance s yle, o in- s ance, h ough impo ing epis ola y o mulae, when w i ing a ex belonging o a gen e no -so-en enched in Basque. We will de elop his idea in wha ollows. (11)–(14) a e ypical F ench eigh een h- o- wen ie h-cen u ies le e openings wi h di e en ypes o RCs (wi h qui, don and lequel). Example(14) esembles non-s anda d Basque ela i es: apa om a esump i e p onoun elle, he p o- noun has he non-s anda d o m (auquel wi h a p eposi ion ins ead o laquelle). B anca-Roso (1990) lis s inco ec ela i es among cha ac e is ics o le e s o ea ly wen ie h-cen u y soldie s. (11) Je épond a a le e que nous a ons eçu il y a qu’elques jou s don nous a ons é és ès con en s de oi que ous ê es ous en bonne san é (B une on-Go e na o i & Mo eux 2015: 89) (12) Je ous ais c is ce e Le e pou ous ai e asa oi de mes nous elle qui son es bonne dieu me sie (Ma ineau 2007: 205) (13) J’ai eçu hie 20 jan ie a le e du 17 dans laquelle u me pa les du dessus de piano d’Adèle e où se ou e le pe i mo de Jeanne e (Vica i 2018: 14) (14) J’ai eçu a longue le e du 28 auquel elle m’a ai g and plaisi (B anca-Roso 1990: 23) Thus, o mulae con aining he zein RC we e used by Basque w i e s, possibly less expe ienced, who we e in need o p e ab ica ed elemen s. Imi a ion was no al- ways pe ec , and o en i was mo e like a “b oken elephone” ansmission, wi h 240 Do o a K ajewska and Eneko Zuloaga misunde s andings and modi ica ions o he cons uc ion. One o he easons o ha could be he limi ed command o Romance languages, and pa icula ly o o mal egis e s. Bilingual speake s who used p ima ily spoken egis e s o Romance languages migh ha e had p oblems wi h ela i es in hese languages, because hey a e a e in he spoken language and because Basque and Romance ela i isa ion s a egies di e . Fo English, o ins ance, Bibe (1988) ound ha pe 1,000 wo ds he e a e on a e age 2.9 ela i e clauses in con e sa ions, 4.6 in p ess epo s, and 8.6 in o icial documen s. Addi ionally, he F ench and Spanish equi alen s o he zein RC– s uc u es wi h lequel and el cual, espec i ely– pe ain o o mal w i en eg- is e s (Gui aud 1966; Auge 1995). Wi h limi ed exposi ion o hose cons uc ions, bilinguals migh no ha e unde s ood hei syn ax. Speake s we e mo e likely o be exposed o Romance cons uc ions wi h in a ian pa icles, which also di e om Basque usual ela i es, bu a e simple o use, and mo e common in he spoken language. The esul ing non-s anda d Basque ela i es a e simila o lequel/el cual ela i es because o he p onoun used, bu unc ion like Romance ela i es wi h in a ian pa icles. Fo mulaic language played a signi ican ole in hese p ocesses. As ega ds he a ian wi h he inessi e zeine an, he e is an epis ola y o mula, in which he ines- si e a gumen is ela i ised (15). I s equi alen is common in o he languages (13), and he i s s ep was o ansla e i in o Basque using he s anda d zein cons uc ion: (15) Recebi u du çu e gu hun-a, [çoin-e an esc iba- zen ecei e aux.1sg>3sg you le e -de .abs which-inde .ines w i e-ip bai- e au açu nola dezi handi-a ducie-la sub-aux.2sg>3sg<1sg how wish big-de .abs ha e.2pl>3sg-compl guci-ec baque e a a co du hun ba -en eça - e- a bi all-de .e g.pl peace and ag eemen good one-gen b ing-nmlz-adl wo he i hoyen a ian] communi y hese.gen be ween ‘I ha e ecei ed you le e in which you w i e me ha you g ea es wish is o b ing peace and good ag eemen be ween he wo communi ies.’ (E xa , adminis a i e le e , 1616–1617) We hink ha such o mulae– o e en his e y o mula– we e he sou ce o he non-s anda d zeine an RC. W i e s eanalysed he inessi e p onoun in a o mula (15) as a pa icle in oducing he clause and ex ended i o si ua ions whe e ano he syn ac ic unc ion is ela i ised. This eanalysis could ha e happened h ough “ ecycling” and sligh modi ica- ions o pa s o o mulae. This is ep esen ed in (16). The i s s age is he o mula (16a). I is hen sho ened: om ‘in which you w i e ha X’ we a i e a ‘in which X’. This clause is used in a sen ence con aining a e e ence o a le e (16b). E en Chap e 10. (Il)li e acy and language change 241 hough syn ac ically i canno unc ion as a RC, seman ically i esembles i . We ind se e al such examples in ou co pus: he ones we labelled as “non- ela i e” (see (8) abo e). In such con ex s he p onoun zeine an loses i s loca i e alue and becomes a seman ically emp y ela i ise . A e wa ds he cons uc ion can be used o modi y any noun ph ase in any ci cums ances (16c), as in (17). (16) a. I ha e ecei ed you le e , [in which you w i e ha X] b. You w i e ha you ha en’ go any le e s … [in which X] c. … [a noun ph ase] [which X] (17) salu a zen zai uz e aizpa Nanak, [zeine an Mia i zen g ee -ip aux.2pl>3sg sis e Nana-e g which-inde .ines Bia i z-ines bai- a neska o] sub-be.3sg maid ‘sis e Nana, who is a maid in Bia i z, sends g ee ings.’ (LeDauphin-15, p i a e le e , 1757) Simila scena ios can be p oposed o o he , less equen a ian s. The op ion wi h zeina can be linked o he heal h o mula, in which he absolu i e o m o he p onoun is ound: (18) Bada, ni-k e e sa i a- zen zai u gu i-a-z, [zeñ-a so I-e g also sa is y-ip aux.1sg>2sg ou -de -ins which-de .abs bai- ugu haiñi z pe ek -a Jainko-a- i eske ] sub-ha e.1pl>3sg e y pe ec -de .abs god-de -da hank ‘And so, I also in o m you on ou (heal h) which, hank God, we ha e pe ec .’ (LeDauphin-49, p i a e le e , 1757) The changes a es ed in Basque can be conside ed ano he example o diapha- sic a ia ion in ela i isa ion s a egies, common in many Eu opean languages, whe e ela i e p onouns ha a e in lec ed in he s anda d language a e in a ian in non-s anda d a ie ies. A pa icula case o his a ia ion in ol es loca i e p onouns as gene al ela i - ise s. In some languages such ela i ise s belong o he s anda d (e.g. pu ‘whe e’ in Mode n G eek (Mu elli 2011: 184)), bu elsewhe e hey a e limi ed o non-s anda d a ie ies. Fo example, wo ‘whe e’ can ela i ise all cases in some Ge man a ie- ies (Romaine 1984; Fleische 2004). In Romance, some hing simila happens in non-s anda d I alian wi h do e ‘whe e’ (Ce u i 2016). The explana ion p oposed he e o Basque non-s anda d ela i e clauses e- sembles he gene al scena io in he de elopmen o ela i e pa icles om loca i e p onouns p oposed by Mu elli (2011: 183). Acco ding o him, he loca i e in e og- a i e is i s used o ela i ise loca i e a gumen s, bu i hen becomes an “unspeci ic connec o ” linking main and subo dina e clauses (which is no a RC syn ac ically, bu he e is a sha ed pa icipan in bo h clauses). F om he e i ex ends o ela i e 242 Do o a K ajewska and Eneko Zuloaga clauses, ela i ising i s inanima e and hen also anima e obliques, be o e ex end- ing o o he syn ac ic unc ions. In Basque ex s we also ind some hing simila o he unspeci ic connec o (in “non- ela i e” examples), hough we do no ha e enough in o ma ion on he o de o ex ension (oblique o g amma ical cases). The da a discussed he e p o ide an insigh in o why such changes can happen and how hey can p oceed in w i en language. In Basque, he eason appea s o be ela ed o language con ac , incomple e bilingualism and inexpe w i e s’ eliance on o - mulaic language. Fixed ph ases we e also impo an o he de elopmen o ela i e cons uc ions elsewhe e, o example in he English wh- ela i e (Romaine 1982; Rissanen 1999). Be gs (2005: 50) a gues ha who, whom o whose “do no simply en e he linguis ic sys em as g amma ical elemen s in one big swoop, bu which o igina e as pa and pa cel o complex o mulaic exp essions which a e hen slowly (des-)in eg a ed in o he g amma .” Non-s anda d zein ela i es also o igina ed as a pa o ixed ph ases, bu e en ually became a p oduc i e ela i isa ion de ice ( he cons uc ion occu s also ou side o mulaic con ex s). Mo eo e , he non-s anda d zein e en ually sp ead o w i ing o people p o icien in Romance. We could also ask why w i e s used he zein ela i e cons uc ion so o en: as compa ed o Basque eligious p ose, his cons uc ion is mo e equen in ou co pus, especially in le e s and no a ial documen s. We hink ha he cons uc- ion, mos p obably no employed in in o mal o al in e ac ions, became a ea u e associa ed wi h w i en language. Le e s a e, in gene al, close o spoken discou se han many o he gen es, bu hey con ain elemen s o mo e o mal egis e s, oo, and e lec people’s awa eness o w i ing con en ions (Ma ineau 2007; an de Wal & Ru en 2016). The use o o mulae is a sign o his awa eness. Mo e gene ally, he sp ead o zein ela i e in o he gen es o he han li e a y ex s can be seen as an example o he c ea ion o Basque w i en egis e . Impo an ly, wi h non-s anda d zein ela i es his p ocess seems o ha e happened in a “na u al” way, in exchanges be ween w i e s, and no h ough in luence o p in ed language. The p oo o ha is ha hose non-s anda d ela i es a e no ound in p in ed Basque books. 6. Conclusion The Basque zein ela i e clause has been adi ionally conside ed an unin e es ing calque om Romance, bu he pa icula i ies behind his bo owing we e no dis- cussed. Ha ing analysed a co pus o gen es unde s udied in Basque linguis ics, such as adminis a i e documen s and p i a e le e s, we ha e shown ha in o de o unde s and how he cons uc ion was in oduced in o he language i is necessa y o ake in o accoun he sociolinguis ic si ua ion and cha ac e is ics o ex ual gen es. Chap e 10. (Il)li e acy and language change 243 We ha e e lec ed on easons o he eme gence and sp ead o zein ela i e, he mechanisms ha unde lie i s di usion o di e en gen es and om mo e li e a e o less li e a e w i e s. Finally, we ha e explained he c ea ion o non-s anda d a ian s o he cons uc ion. Simila p ocesses o hose obse ed in Basque we e desc ibed o o he Eu opean languages, bu ou case s udy p o ides an insigh in o how such p ocesses ake place in he con ex o diglossia and lack o s anda d a ie y. This sociolinguis ic si ua ion made people look o models in o he languages (speci ically, imi a ing Romance cons uc ions), bu also in Basque ex s hey had access o, such as le e s. In his way, he ela i e cons uc ion wi h he p onoun zein ‘which’, common in li e a y ex s, sp ead o adminis a i e o legal language and p i a e le e s. In le e s, epis ola y o mulae, pa ially ixed exp essions usually ansla ed om Romance, played a key ole in hese p ocesses, because a ew o hem con ained he zein ela i e. In a u he de elopmen , non-s anda d ela i es (e.g. he a ian wi h inessi e p onoun) eme ged when, simila ly o wha occu ed in o he Eu opean languages, w i e s eanalysed he syn ax o o mulae wi h he s anda d zein ela i e. We p opose ha his inno a ion happened because o he limi ed command o Romance languages o some w i e s, who did no unde s and he syn ax o he s anda d cons uc ion. The new ela i isa ion s a egy hen be- came p oduc i e and sp ead o o he w i e s. We a gue ha zein ela i e, s anda d and non-s anda d, e en ually became a ea u e o he eme ging Basque w i en egis e . The li e o he non-s anda d ela i e cons uc ion, hough in e es ing, was ela i ely sho , as i did no make i in o he wen ie h cen u y, and he easons o i appea o be inc easing knowledge o Romance languages in he popula ion. Abb e ia ions and glossing con en ions abs absolu i e ins ins umen al adl adla i e ip impe ec i e aux auxilia y e b neg nega ion compl complemen ise nmlz nominalisa ion cond condi ional pa pa i i e da da i e pl plu al de de ini e po po en ial e g e ga i e ps pas gen geni i e m ela ional ma ke hyp hypo he ical sg singula inde inde ini e soc socia i e ines inessi e sub subo dina o In glosses o ini e e bs he sign “>” dis inguishes e ga i e and absolu i e a gumen s and “<” dis inguishes da i e ones. 244 Do o a K ajewska and Eneko Zuloaga Funding The esea ch o he pape was made possible by he g an s om he Spanish Minis y o Science and Inno a ion (FFI2016-76032-P and PGC2018-098995-B-I00). Suppo gi en by he esea ch g oup on his o ical linguis ics (IT1344-19) unded by he Basque Go e nmen is also acknowledged. Re e ences Amo o u, Es ibaliz. 2003. Basque sociolinguis ics: Language, socie y, and cul u e. Reno: Cen e o Basque S udies. Auge , Julie. 1995. On he his o y o ela i e clauses in F ench and some o i s dialec s. In Hen- ning Ande sen (ed.), His o ical linguis ics, 1993: Selec ed pape s om he 11 h In e na ional Con e ence on His o ical Linguis ics, 19–32. Ams e dam & Philadelphia: John Benjamins. h ps://doi.o g/10.1075/cil .124.03aug Azkue, Resu ección Ma ía. 1891. Euskal izkindea. Bilbao: As ui. Azkue, Resu ección Ma ía. 1923–1925. Mo ología asca. Bilbao: Edi o ial Vasca. Azkue, Resu ección Ma ía. 1969 [1905–1906]. Dicciona io asco-español- ancés. Bilbao: La G an Enciclopedia Vasca. Beau o , Anne & Anna Iñes a. 2014. Au ho p o iles: Awa eness, compe ence, and skills. In E a-Ma ia Jakobs & Daniel Pe in (eds.), Handbook o w i ing and ex p oduc ion, 141–158. Be lin: Mou on De G uy e . h ps://doi.o g/10.1515/9783110220674.141 Be gs, Alexande . 2005. The ole o he indi idual in language change om he poin o iew o Social Ne wo k Analysis. Logos and Language 6. 30–54. Bibe , Douglas. 1988. Va ia ion ac oss speech and w i ing. Camb idge: Camb idge Uni e si y P ess. h ps://doi.o g/10.1017/CBO9780511621024 Blanche-Ben enis e, Clai e. 1990. Usages no ma i s e non no ma i s dans les ela i es en ançais, en espagnol e en po ugais. In Johannes Beche , Giuliano Be nini & Claude Bu idan (eds.), Towa d a ypology o Eu opean languages, 317–335. Be lin: Mou on de G uy e . h ps://doi.o g/10.1515/9783110863178.317 B anca-Roso , Sonia. 1990. Con en ions d’éc i u e dans la co espondance des solda s. Mo s 24(1). 21–36. h ps://doi.o g/10.3406/mo s.1990.1535 B une on-Go e na o i, A iane & Be na d Mo eux. 2015. Un modèle épis olai e populai e: Les le es d’émig és béa nais. In Daniel Fab e (ed.), Pa éc i : E hnologie des éc i u es quo idi- ennes, 79–103. Pa is: Édi ions de la Maison des sciences de l’homme. Camino, Iñaki. 2012. Bi es u llabu Aezkoa e a Zube oako. Ga alda (1828) & La aine (1817). Fon es Linguae Vasconum 114. 61–72. Ca e , Ka en E. 2011. C ea ing Ca holics: ca echism and p ima y educa ion in ea ly mode n F ance. No e Dame: Uni e si y o No e Dame P ess. Ce u i, Massimo. 2016. Changes om below, changes om abo e. Rela i e cons uc ions in con empo a y I alian. In Massimo Ce u i, Claudia C occo & S e ania Ma zo (eds.), To- wa ds a new s anda d: Theo e ical and empi ical s udies on he es anda diza ion o I alian, 62–88. Be lin: Mou on de G uy e . Cha ie , Roge . 1997. Du li e au li e. Réseaux 1. 271–290. Cha ie , Roge , Dominique Julia & Ma ie-Madeleine Compè e. 1976. L’éduca ion en F ance du XVIe au XVIIIe siècle. Pa is: Socié é d’Éduca ion d’Enseignemen Supé ieu .