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

Language Variation – European Perspectives VIII Edited by Hans Van de Velde Nanna Haug Hilton Remco Knooihuizen John Benjamins Publishing Company Studies in Language Variation 25 Language Variation – European Perspectives VIII Studies in Language Variation (SILV) issn 1872-9592 Volume 25 Language Variation – European Perspectives VIII Selected papers from the Tenth International Conference on Language Variation in Europe (ICLaVE 10), Leeuwarden, June 2019 Edited by Hans Van de Velde, Nanna Haug Hilton and Remco Knooihuizen The series aims to include empirical studies of linguistic variation as well as its description, explanation and interpretation in structural, social and cognitive terms. The series will cover any relevant subdiscipline: sociolinguistics, contact linguistics, dialectology, historical linguistics, anthropology/anthropological linguistics. The emphasis will be on linguistic aspects and on the interaction between linguistic and extralinguistic aspects — not on extralinguistic aspects (including language ideology, policy etc.) as such. For an overview of all books published in this series, please see benjamins.com/catalog/silv Editors Peter Auer Universität Freiburg Frans Hinskens Meertens Instituut & Radboud Universiteit Nijmegen Paul Kerswill University of York Editorial Board Arto Anttila Stanford University Gaetano Berruto Università di Torino Jenny Cheshire University of London Katie Drager University of Hawai‘i at Mānoa Katarzyna Dziubalska-Kołaczyk Adam Mickiewicz University in Poznań Jürg Fleischer Philipps-Universität Marburg Peter Gilles University of Luxembourg Brian D. Joseph The Ohio State University Johannes Kabatek Universität Zürich Pieter Muysken Radboud University Nijmegen Pia Quist University of Copenhagen Anne-Catherine Simon Université catholique de Louvain Sali A. Tagliamonte University of Toronto Øystein Alexander Vangsnes UiT The Arctic University of Norway Juan A. Villena Ponsoda Universidad de Málaga Language Variation – European Perspectives VIII Selected papers from the Tenth International Conference on Language Variation in Europe (ICLaVE 10), Leeuwarden, June 2019 Edited by Hans Van de Velde Fryske Akademy & Utrecht University Nanna Haug Hilton University of Groningen Remco Knooihuizen University of Groningen John Benjamins Publishing Company Amsterdam / Philadelphia doi 10.1075/silv.25 Cataloging-in-Publication Data available from Library of Congress: lccn 2021010868 (print) / 2021010869 (e-book) isbn 978 90 272 0885 9 (Hb) isbn 978 90 272 5982 0 (e-book) © 2021 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Company · https://benjamins.com 8 TM The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984. Table of contents Introduction 1 Remco Knooihuizen, Nanna Haug Hilton and Hans Van de Velde Chapter1 The volatile linguistic shape of ‘Town Frisian’/‘Town Hollandic’ 11 Arjen P. Versloot Chapter2 Is there an interlanguage speech acceptability deficit? 35 Rias van den Doel and Adriaan Walpot Chapter3 Revisiting the vowel mergers of East Anglia: Correlations of moan, mown and goose 53 Kerri-Ann Butcher Chapter4 Modeling regional variation in voice onset time of Jutlandic varieties of Danish 79 Rasmus Puggaard Chapter5 “Organically German”? Changing ideologies of national belonging 111 Janet M. Fuller Chapter6 Exploring an approach for modelling lectal coherence 135 Karen V. Beaman Chapter7 “I’m dead posh in school”: Attitudes and linguistic behaviour of Merseyside adolescents 161 Rachel Byrne Chapter8 Benim: A new pronoun in Swedish 181 Nathan J. Young vi Language Variation – European Perspectives VIII Chapter9 Identification of clusters of lexical areas using geographical factors: A case study in the Occitan language area 209 Clément Chagnaud, Guylaine Brun-Trigaud and Philippe Garat Chapter10 (Il)literacy and language change: Non-standard relative constructions in historical Basque 227 Dorota Krajewska and Eneko Zuloaga Chapter11 Dialect contact in the vowel system of Mišótika Cappadocian 247 Nicole Vassalou, Dimitris Papazachariou and Mark Janse Chapter12 Leaders of language change: Macro and micro perspectives 269 Meredith Tamminga Chapter13 Ethnic variation in real time: Change in Australian English diphthongs 291 James Grama, Catherine E. Travis and Simon Gonzalez Index 315 Chapter10 (Il)literacy and language change Non-standard relative constructions in historical Basque Dorota Krajewska and Eneko Zuloaga University of the Basque Country UPV/EHU In this paper we examine a corpus of sixteenthto nineteenth-century Basque private letters and administration documents from a historical sociolinguistic point of view. Because of the diglossic situation, such texts are rare but valuable in the historical corpus of Basque. In particular, we analyse relative clauses with the pronoun zein ‘which’. The construction was borrowed from Romance by highly literate bilinguals, but then spread, especially through formulaic language, to less literate writers too. We focus on the development of non-standard variants of this relative clause, which are common in our corpus, but not found in printed literary texts. We argue that the sociolinguistic context was crucial in the emergence, spread and syntactic change of the relative construction. Keywords: historical sociolinguistics, syntax, relative clauses, literacy, Basque 1. Introduction In this paper we adopt the viewpoint of historical sociolinguistics to reflect on Basque writing, written language, and syntactic change under language contact. Our analysis is based on a corpus of sixteenthto nineteenth-century Basque letters and administrative documents. Processes described here have parallels in other European languages, but the Basque case offers the opportunity to observe them in a particularly interesting diglossic context, where a prestigious and standardised language (French or Spanish) was the usual means of written communication, and another language (Basque) had not yet developed a standard. We focus on relative clauses (RC), which are complex constructions, typical of higher registers, and prone to cause problems for not-so-proficient writers. The most common relativisation strategy in modern Basque and in most historical sources is prenominal, and differs from relative clauses found in the surrounding https://doi.org/10.1075/silv.25.10kra © 2021 John Benjamins Publishing Company 228 Dorota Krajewska and Eneko Zuloaga Romance languages. However, through contact Basque innovated a construction with the pronoun zein ‘which’, which is the object of our study. Being considered an uninteresting calque, it has received little attention, and its use outside the “canonical” sources (such as printed books) has not been studied. We show that the relative pronoun zein is common in our corpus and that it became a feature of written registers. Moreover, we propose that several non-standard variants of the construction demonstrate that the diffusion of syntactic change happened in written interactions between writers, for example, by means of exchanges of letters, and mainly through formulaic language. The paper is structured as follows. Section2 introduces relativisation strategies in Basque and Romance. Section3 describes the sociolinguistic situation of Basque before the twentieth century and the influence it had on writing. Section4 presents our corpus study. Section5 provides an explanation of the syntactic changes observed in the corpus and compares them to phenomena attested in other European languages. 2. Basque and Romance relativisation strategies Basque has several finite relative constructions, which differ in the following features (De Rijk 1972; Oyharçabal 1987; Oyharçabal 2003; Krajewska 2017) (see Table1): a. embedded vs. appositive RC b. position of the RC: postnominal vs. prenominal c. treatment of the relativised noun: gap vs. pronoun d. choice of subordinator: -en, baitor both We thus have prenominal (1), postnominal (2), appositive (3), and zein relatives (4), illustrated below with examples from classical Basque sources: (1) [gizon-ek erran dukeite-n] hitz alfer guzi-a-z man-def.erg.pl say aux.3pl>3sg-sub word vain all-def-ins ‘all the vain words man can say’ (Leizarraga, 1571) (2) arbore [fruktu on-ik egi-ten ez-tu-en guzi-a] tree fruit good-part make-ipfv neg-aux.3sg>3sg-sub all-def.abs ‘every tree which does not produce good fruit’ (Leizarraga, 1571) (3) zuhaitz alferr-a [fruitu-rik iasai-ten ez-tu-en-a] tree lazy-def.abs fruit-part give-ipfv neg-aux.3sg>3sg-sub-def.abs ‘a lazy tree which does not give fruit’ (Axular, 1643) Chapter10. (Il)literacy and language change 235 4.3 Non-standard variants The most interesting aspect of the zein relative construction in our corpus is its non-standard variants, which are not found in canonical sources. They involve a form of the pronoun zein ‘which’ functioning as an invariable relative particle. Contrary to the standard construction, the pronoun does not take the case corresponding to the relativised argument. In the most common non-standard variant, the inessive pronoun– usually the indefinite zeinetan and sometimes the definite singular zeinean– relativises all syntactic functions, for example, the absolutive subject: (6) Iskriba-tzen dauzkitzut bi lerro hauk zu-ri gure write-ipfv aux.1sg>3pl<2sg two line these you-dat our osasun-a-ren marka-tze-ko, [zein-etan hanitz on health-def-gen inform-nmlz-rm which-indef.ines very good bai-ta Jainko-a-ri eskerr-ak] sub-be.3sg god-def-dat thank-def.abs.pl ‘I am writing these two lines to you to inform you on my health, which, thank God, is very good.’ (LeDauphin-45, private letter, 1757) In other variants we find zeinak, ambiguous between absolutive plural or ergative singular (used to relativise absolutive singular in (7)), or the absolutive definite zeina. (7) Egorri dinat bat Lantzekanetu-rekin, bertze-a hire send aux.1sg>3sg<2sg one Lantzekanetu-soc other-def.abs your kusin-a defuntu-ba-rekin, [zen-ak etorri cousin-def.abs dead-def-soc which-def.abs.pl/def.erg come bait-zen Misel-ekin Baiona-rat, gaso-a] sub-aux.pst.3sg Misel-soc Bayonne-adl poor-def.abs ‘I have sent you one with Lantzekanetu, and another one with your late cousin, who came with Misel to Bayonne, poor guy.’ (LeDauphin-20, private letter, 1757) There are also clauses that we label as “non-relative” because what looks like a subordinate relative does not modify anything. For example, in (8), as the translation shows, the syntax is rather odd: if ‘in which’ refers to the letter, a part of the sentence is missing (we would expect ‘in which we wrote that…’). 236 Dorota Krajewska and Eneko Zuloaga (8) Marka-tzen darotazu ez-tuzu-la Qanigon-ekin write-ipfv aux.2sg>3sg<1sg neg-aux.2sg>3sg-compl Qanigon-soc izatu letra-rik, bainan egin gintuben, [zein-etan have letter-part but make aux.pst.1pl>3sg which-indef.ines aizpa Nana Miarritz-en bai-tugu neskato Bodri-rekin] sister Nana Miarritz-ines sub-have.1pl>3sg maid Bodri-soc ‘You write that you haven’t got our letters from Qanigon, but we wrote one, in which sister Nana is maid in Biarritz with Bodri.’ (LeDauphin-38, private letter, 1757) Non-standard variants are most common in private letters, where over 40% of the examples are of this kind (Table4). Table4. Non-standard zein relative clauses in different text types Text type Tokens of zein RC Non-standard variants (%) Adm. doc. 55 20.0 Notaries 37 18.9 Adm. letters 70 7.1 Private letters 65 43.1 Table5 presents occurrences of each variant (including the non-relative), and the number of sources in which it appears. The variant with zeinetan is the most widespread non-standard option. Table5. Frequencies of variants of zein relative Zein RC type Tokens Texts Standard 176 88 Non-standard zeinetan 28 22 Non-standard zeina/zeinak 15 8 Non-relative 8 4 The frequency of non-standard relatives seems highest in the eighteenth century, and it decreases in the nineteenth century. This might reflect an increase in literacy and knowledge of Romance languages, which would cause using the standard zein relative. Non-standard variants were unsuccessful innovations: they are not attested in the twentieth century.3 3. However, in modern written Basque there is a non-standard relative construction with the invariant particle non ‘where’, which seems not unlike the variant with zeinetan, but its development and use require further study. Chapter10. (Il)literacy and language change 237 4.4 Zein relatives and formulaic language Zein relatives in our corpus are often found in epistolary formulae, which are defined by Wray (2005: 9) as “a sequence, continuous or discontinuous, of words or other elements, which is, or appears to be, prefabricated: that is, stored and retrieved whole from memory at the time of use, rather than being subject to generation or analysis by the language grammar”. Following the classification by Wray (2005) and Rutten and Van der Wal (2012), two classes are relevant here: text-constitutive (e.g. letter openings) and intersubjective formulae (e.g. greetings or expressions of health). In a small corpus such as ours it is difficult to decide what counts as a formula. We thus only include two classes frequent in the data: (a) letter openings (“I have received your letter in which …” and its variations, see Example(9)), and (b) the expression of the sender’s health and hope that the receiver is alright (“we have health, which we hope you also have”, “I want to let you know about my health, which is good” and similar phrases, Example(10)). (9) Rezebitu dut atzo zu-k eskribatu karta, receive aux.1sg>3sg yesterday you-erg write letter [zoin-tan erra-ten bai-tuzu jente paubri-ak which-indef.ines say-ipfv sub-aux.2sg>3sg people poor-def.abs.pl dire-la zu-ri eska-tzen dute-n-ak be.3pl-compl you-dat ask-ipfv aux.3pl>3sg-sub-def.abs.pl gastu-ak] payment-def.abs.pl ‘Yesterday I received the letter you wrote, in which you say that it is the poor people that asks you for money.’ (administration letter, 1817; Camino (2012)) (10) Osasun dugu Janko-a-ri esker, [zein-etan dezira-tzen health have.1pl>3sg god-def-dat thank which-indef.ines wish-ipfv bai-kinduke zuri-a hala ba-litz] sub-aux.pot.1pl>3sg your-def.abs so cond-be.hyp.3sg ‘We have health, thank God, which we wish you also have.’ (LeDauphin-46, private letter, 1757) Table6 shows that there is a correlation between the frequency of occurrence in a formula and letter type, on one hand, and type of relative, on the other hand. In private letters relatives appear more frequently in formulae than in administration letters. In Le Dauphin collection, this proportion is even higher (39%). Non-standard relatives are also more common as part of a formula than the standard construction. 238 Dorota Krajewska and Eneko Zuloaga Table6. Relative clauses in formulae according to the text type and construction Context Tokens Formulae Letter type Administrative 70 12.9% Private 65 29.2% Relative clause type Standard 102 14.7% Non-standard 33 39.4% 5. Discussion In this paper we focus on texts written in Basque despite the adverse diglossic situation. Their authors usually belonged to upper social strata and were taught how to write in Romance. Thus, literacy in Basque depended on literacy in another language. This fact is patent in spelling: for instance, writers often struggled to render Basque sibilants using Romance systems, which have fewer sibilant phonemes. Here we explore the consequences it had on syntax. Writing does not usually happen in void and knowing how to write is not limited to the command of the code. It also presupposes knowledge related to different aspects of writing, such as genre knowledge and knowledge related to the social context of writing (Beaufort & Iñesta 2014). For example, writing a letter requires understanding of the conventions ruling this type of texts and of how the given community usually writes them, as well as what is thought to be correct and what not. Genres can in fact be understood as “socially and culturally patterned ways of engaging in activity” (Lillis 2013: 70). As regards genres analysed here, in terms of Koch and Oesterreicher (2012 [1985]), private letters are closer to the language of immediacy, and formal documents produced by different institutions are examples of language of distance: they constitute a monologue, there is distance between the partners, and topics are fixed. Administrative correspondence falls somewhere in between. Nevertheless, the texts in our corpus were written before the language was standardised, and as a result they exhibit dialectal features and characteristics described for the language of immediacy by Koch & Oesterreicher (2012 [1985]: 454–455): congruence errors, problems with segmentation, low type-token ratio, overall lexical poverty, hesitation phenomena or markers of correction. Moreover, features of illiterate writing (see Montgomery 1995) are common: problems with spelling and word division or missing punctuation marks. This is because those texts were written in Basque by people who were literate in Romance. In a complex sociolinguistic environment, they could not rely on a fully standardised language and had to use a dialectal variety. Nevertheless, they looked for models to construct their writings. Chapter10. (Il)literacy and language change 239 Bilinguals could resort to norms of other languages. For instance, when a clerk prepared a Basque document, he could translate from Spanish or French the typical phrases of administrative language, and, in general, try to copy the style of the text. The same holds for other conventions, for example the use of epistolary formulae. Writers also treated Basque texts as models. For this to be possible, though, a certain tradition of Basque writing had to exist in the social environment. For instance, when composing a letter, the writer could reuse fragments of Basque letters, provided he had access to any. In general, learning of correspondence-writing conventions usually happened in practice: by reading or listening to letters and then imitation (see van der Wal & Rutten 2013 and references therein). This explains the similarity between texts written by different people. The zein relative construction is thought to have developed in Basque due to language contact (Lafitte 1991 [1944]; Haase 1992; Trask 1998). This might be a change from above spreading in writing from more literate bilinguals (usually belonging to higher social classes) to less proficient ones. Similar explanations have been proposed for changes in relativisation strategies in a number of languages (Romaine 1982; Rissanen 1999; Rutten & van der Wal 2014, among others). In Basque, this change could happen through imitating the Romance style, for instance, through importing epistolary formulae, when writing a text belonging to a genre not-so-entrenched in Basque. We will develop this idea in what follows. (11)–(14) are typical French eighteenth-to-twentieth-centuries letter openings with different types of RCs (with qui, dont and lequel). Example(14) resembles non-standard Basque relatives: apart from a resumptive pronoun elle, the pronoun has the non-standard form (auquel with a preposition instead of laquelle). Branca-Rosoff (1990) lists incorrect relatives among characteristics of letters of early twentieth-century soldiers. (11) Je répond a ta lettre que nous avons reçu il y a qu’elques jours dont nous avons étés très contents de voir que vous êtes tous en bonne santé (Bruneton-Governatori & Moreux 2015: 89) (12) Je vous aist crist cette Lettre pour vous faire asavoir de mes nousvelle qui sontres bonne dieu mersie (Martineau 2007: 205) (13) J’ai reçu hier 20 janvier ta lettre du 17 dans laquelle tu me parles du dessus de piano d’Adèle et où se trouve le petit mot de Jeannette (Vicari 2018: 14) (14) J’ai reçu ta longue lettre du 28 auquel elle m’a fait grand plaisir (Branca-Rosoff 1990: 23) Thus, formulae containing the zein RC were used by Basque writers, possibly less experienced, who were in need of prefabricated elements. Imitation was not always perfect, and often it was more like a “broken telephone” transmission, with 240 Dorota Krajewska and Eneko Zuloaga misunderstandings and modifications of the construction. One of the reasons for that could be the limited command of Romance languages, and particularly of formal registers. Bilingual speakers who used primarily spoken registers of Romance languages might have had problems with relatives in these languages, because they are rare in the spoken language and because Basque and Romance relativisation strategies differ. For English, for instance, Biber (1988) found that per 1,000 words there are on average 2.9 relative clauses in conversations, 4.6 in press reports, and 8.6 in official documents. Additionally, the French and Spanish equivalents of the zein RC– structures with lequel and el cual, respectively– pertain to formal written registers (Guiraud 1966; Auger 1995). With limited exposition to those constructions, bilinguals might not have understood their syntax. Speakers were more likely to be exposed to Romance constructions with invariant particles, which also differ from Basque usual relatives, but are simpler to use, and more common in the spoken language. The resulting non-standard Basque relatives are similar to lequel/el cual relatives because of the pronoun used, but function like Romance relatives with invariant particles. Formulaic language played a significant role in these processes. As regards the variant with the inessive zeinetan, there is an epistolary formula, in which the inessive argument is relativised (15). Its equivalent is common in other languages (13), and the first step was to translate it into Basque using the standard zein construction: (15) Recebitu dut çure guthun-a, [çoin-etan escriba-tzen receive aux.1sg>3sg your letter-def.abs which-indef.ines write-ipfv bai-terautaçu nola dezir handi-a ducie-la sub-aux.2sg>3sg<1sg how wish big-def.abs have.2pl>3sg-compl guci-ec baque eta arcordu hun bat-en eçar-te-ra bi all-def.erg.pl peace and agreement good one-gen bring-nmlz-adl two herri hoyen artian] community these.gen between ‘I have received your letter in which you write me that your greatest wish is to bring peace and good agreement between the two communities.’ (Etxart, administrative letter, 1616–1617) We think that such formulae– or even this very formula– were the source of the non-standard zeinetan RC. Writers reanalysed the inessive pronoun in a formula (15) as a particle introducing the clause and extended it to situations where another syntactic function is relativised. This reanalysis could have happened through “recycling” and slight modifications of parts of formulae. This is represented in (16). The first stage is the formula (16a). It is then shortened: from ‘in which you write that X’ we arrive at ‘in which X’. This clause is used in a sentence containing a reference to a letter (16b). Even Chapter10. (Il)literacy and language change 241 though syntactically it cannot function as a RC, semantically it resembles it. We find several such examples in our corpus: the ones we labelled as “non-relative” (see (8) above). In such contexts the pronoun zeinetan loses its locative value and becomes a semantically empty relativiser. Afterwards the construction can be used to modify any noun phrase in any circumstances (16c), as in (17). (16) a. I have received your letter, [in which you write that X] b. You write that you haven’t got any letters … [in which X] c. … [a noun phrase] [which X] (17) salutatzen zaituzte aizpa Nanak, [zeinetan Miarritzen greet-ipfv aux.2pl>3sg sister Nana-erg which-indef.ines Biarritz-ines bai-ta neskato] sub-be.3sg maid ‘sister Nana, who is a maid in Biarritz, sends greetings.’ (LeDauphin-15, private letter, 1757) Similar scenarios can be proposed for other, less frequent variants. The option with zeina can be linked to the health formula, in which the absolutive form of the pronoun is found: (18) Bada, ni-k ere satifa-tzen zaitut guri-a-z, [zeñ-a so I-erg also satisfy-ipfv aux.1sg>2sg our-def-ins which-def.abs bai-tugu haiñitz perfekt-a Jainko-a-ri esker] sub-have.1pl>3sg very perfect-def.abs god-def-dat thank ‘And so, I also inform you on our (health) which, thank God, we have perfect.’ (LeDauphin-49, private letter, 1757) The changes attested in Basque can be considered another example of diaphasic variation in relativisation strategies, common in many European languages, where relative pronouns that are inflected in the standard language are invariant in non-standard varieties. A particular case of this variation involves locative pronouns as general relativisers. In some languages such relativisers belong to the standard (e.g. pu ‘where’ in Modern Greek (Murelli 2011: 184)), but elsewhere they are limited to non-standard varieties. For example, wo ‘where’ can relativise all cases in some German varieties (Romaine 1984; Fleischer 2004). In Romance, something similar happens in non-standard Italian with dove ‘where’ (Cerruti 2016). The explanation proposed here for Basque non-standard relative clauses resembles the general scenario in the development of relative particles from locative pronouns proposed by Murelli (2011: 183). According to him, the locative interrogative is first used to relativise locative arguments, but it then becomes an “unspecific connector” linking main and subordinate clauses (which is not a RC syntactically, but there is a shared participant in both clauses). From there it extends to relative 242 Dorota Krajewska and Eneko Zuloaga clauses, relativising first inanimate and then also animate obliques, before extending to other syntactic functions. In Basque texts we also find something similar to the unspecific connector (in “non-relative” examples), though we do not have enough information on the order of extension (oblique to grammatical cases). The data discussed here provide an insight into why such changes can happen and how they can proceed in written language. In Basque, the reason appears to be related to language contact, incomplete bilingualism and inexpert writers’ reliance on formulaic language. Fixed phrases were also important for the development of relative constructions elsewhere, for example in the English whrelative (Romaine 1982; Rissanen 1999). Bergs (2005: 50) argues that who, whom or whose “do not simply enter the linguistic system as grammatical elements in one big swoop, but which originate as part and parcel of complex formulaic expressions which are then slowly (des-)integrated into the grammar.” Non-standard zein relatives also originated as a part of fixed phrases, but eventually became a productive relativisation device (the construction occurs also outside formulaic contexts). Moreover, the non-standard zein eventually spread to writing of people proficient in Romance. We could also ask why writers used the zein relative construction so often: as compared to Basque religious prose, this construction is more frequent in our corpus, especially in letters and notarial documents. We think that the construction, most probably not employed in informal oral interactions, became a feature associated with written language. Letters are, in general, closer to spoken discourse than many other genres, but they contain elements of more formal registers, too, and reflect people’s awareness of writing conventions (Martineau 2007; van der Wal & Rutten 2016). The use of formulae is a sign of this awareness. More generally, the spread of zein relative into the genres other than literary texts can be seen as an example of the creation of Basque written register. Importantly, with non-standard zein relatives this process seems to have happened in a “natural” way, in exchanges between writers, and not through influence of printed language. The proof of that is that those non-standard relatives are not found in printed Basque books. 6. Conclusion The Basque zein relative clause has been traditionally considered an uninteresting calque from Romance, but the particularities behind this borrowing were not discussed. Having analysed a corpus of genres understudied in Basque linguistics, such as administrative documents and private letters, we have shown that in order to understand how the construction was introduced into the language it is necessary to take into account the sociolinguistic situation and characteristics of textual genres. Chapter10. (Il)literacy and language change 243 We have reflected on reasons for the emergence and spread of zein relative, the mechanisms that underlie its diffusion to different genres and from more literate to less literate writers. Finally, we have explained the creation of non-standard variants of the construction. Similar processes to those observed in Basque were described for other European languages, but our case study provides an insight into how such processes take place in the context of diglossia and lack of standard variety. This sociolinguistic situation made people look for models in other languages (specifically, imitating Romance constructions), but also in Basque texts they had access to, such as letters. In this way, the relative construction with the pronoun zein ‘which’, common in literary texts, spread to administrative or legal language and private letters. In letters, epistolary formulae, partially fixed expressions usually translated from Romance, played a key role in these processes, because a few of them contained the zein relative. In a further development, non-standard relatives (e.g. the variant with inessive pronoun) emerged when, similarly to what occurred in other European languages, writers reanalysed the syntax of formulae with the standard zein relative. We propose that this innovation happened because of the limited command of Romance languages of some writers, who did not understand the syntax of the standard construction. The new relativisation strategy then became productive and spread to other writers. We argue that zein relative, standard and non-standard, eventually became a feature of the emerging Basque written register. The life of the non-standard relative construction, though interesting, was relatively short, as it did not make it into the twentieth century, and the reasons for it appear to be increasing knowledge of Romance languages in the population. Abbreviations and glossing conventions abs absolutive ins instrumental adl adlative ipfv imperfective aux auxiliary verb neg negation compl complementiser nmlz nominalisation cond conditional part partitive dat dative pl plural def definite pot potential erg ergative pst past gen genitive rm relational marker hyp hypothetical sg singular indef indefinite soc sociative ines inessive sub subordinator In glosses of finite verbs the sign “>” distinguishes ergative and absolutive arguments and “<” distinguishes dative ones. 244 Dorota Krajewska and Eneko Zuloaga Funding The research for the paper was made possible by the grants from the Spanish Ministry of Science and Innovation (FFI2016-76032-P and PGC2018-098995-B-I00). Support given by the research group on historical linguistics (IT1344-19) funded by the Basque Government is also acknowledged. References Amorrortu, Estibaliz. 2003. Basque sociolinguistics: Language, society, and culture. Reno: Center for Basque Studies. Auger, Julie. 1995. On the history of relative clauses in French and some of its dialects. In Henning Andersen (ed.), Historical linguistics, 1993: Selected papers from the 11th International Conference on Historical Linguistics, 19–32. Amsterdam & Philadelphia: John Benjamins. https://doi.org/10.1075/cilt.124.03aug Azkue, Resurrección María. 1891. Euskal izkindea. Bilbao: Astui. Azkue, Resurrección María. 1923–1925. Morfología vasca. Bilbao: Editorial Vasca. Azkue, Resurrección María. 1969 [1905–1906]. Diccionario vasco-español-francés. Bilbao: La Gran Enciclopedia Vasca. Beaufort, Anne & Anna Iñesta. 2014. Author profiles: Awareness, competence, and skills. In Eva-Maria Jakobs & Daniel Perrin (eds.), Handbook of writing and text production, 141–158. Berlin: Mouton De Gruyter. https://doi.org/10.1515/9783110220674.141 Bergs, Alexander. 2005. The role of the individual in language change from the point of view of Social Network Analysis. Logos and Language 6. 30–54. Biber, Douglas. 1988. Variation across speech and writing. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9780511621024 Blanche-Benveniste, Claire. 1990. Usages normatifs et non normatifs dans les relatives en français, en espagnol et en portugais. In Johannes Bechert, Giuliano Bernini & Claude Buridant (eds.), Toward a typology of European languages, 317–335. Berlin: Mouton de Gruyter. https://doi.org/10.1515/9783110863178.317 Branca-Rosoff, Sonia. 1990. Conventions d’écriture dans la correspondance des soldats. Mots 24(1). 21–36. https://doi.org/10.3406/mots.1990.1535 Bruneton-Governatori, Ariane & Bernard Moreux. 2015. Un modèle épistolaire populaire: Les lettres d’émigrés béarnais. In Daniel Fabre (ed.), Par écrit: Ethnologie des écritures quotidiennes, 79–103. Paris: Éditions de la Maison des sciences de l’homme. Camino, Iñaki. 2012. Bi testu llabur Aezkoa eta Zuberoako. Garralda (1828) & Larraine (1817). Fontes Linguae Vasconum 114. 61–72. Carter, Karen E. 2011. Creating Catholics: catechism and primary education in early modern France. Notre Dame: University of Notre Dame Press. Cerruti, Massimo. 2016. Changes from below, changes from above. Relative constructions in contemporary Italian. In Massimo Cerruti, Claudia Crocco & Stefania Marzo (eds.), Towards a new standard: Theoretical and empirical studies on the restandardization of Italian, 62–88. Berlin: Mouton de Gruyter. Chartier, Roger. 1997. Du livre au lire. Réseaux 1. 271–290. Chartier, Roger, Dominique Julia & Marie-Madeleine Compère. 1976. L’éducation en France du XVIe au XVIIIe siècle. Paris: Société d’Éducation d’Enseignement Supérieur.