«Variation and contact in French spoken outside France» Françoise Gadet, Université de Paris-10

Paper for the Symposium Language Contact and the Dynamics of Language: Theory and Implications 10-13 May 2007 Leipzig «Variation and contact in French spoken outside France» Françoise Gadet, Université de Paris-10 Everywhere around the world, except in France 1, the French language is «in contact» with one or several language(s). And everywhere, it is in an institutional or de facto dominated position (even, in some American situations e. g. Newfoundland, New England or Missouri - it is so dramatically dominated as to be in a process of attrition or even approaching death). In this communication, I shall consider the relationships between variation and interference. What are the effects of contact on the scope of variation in the French language, taken in a very wide meaning? 1. Theoretical considerations For the description of facts relating to contact in French, I shall adopt a quite unusual position for a sociolinguist, i. e. the point of view of writing a grammar. More specifically, I shall speak of my experience as a participant in a project to create a Reference Grammar of French, where I am in charge of the documentation concerning all facts of all types of possible variation in French. 1.1. The Reference Grammar of French, and the place intended to variation My participation in this huge project (including 50 collaborators, 2500 pages intended, to be published in 2009) requires me to take into account, as can be expected, diatopic, diastratic, and diaphasic variation in France itself, as well as the discursive genres and the consequences of the written/spoken dimension (if any, as far as the syntactic level is concerned). Among dimensions of variation, what appears to have the most decisive consequences on linguistic forms is diatopic variation, and especially diatopy outside France, in relation with the complex history of French around the world. French can offer the products of extremely 1 It could be said that French is also «in contact» in France, as 1) there are no isolated language, nowhere around the world, and 2) in France, a lot of speakers speaks a language other than French. But the situations of French in and outside France are certainly quite different, and it does not imply the same things, for a language, to be in a dominant or in a dominated position. 1

diversified historical situations : two main periods of expansion, through emigration and through colonization (17 e and 19 e century) ; first-degree diaspora from France, and second-degree diaspora from Quebec ; different local events and histories ; different conditions of conservation, with different isolation conditions (mainland, islands, or isolates) ; and finally, different modalities of contact with different local languages. Creoles, the extreme pole of a continuum in French varieties, won t be considered here. What I shall argue here is that, even if the ultimate aim for a grammar is to present grammatical data, it is not possible to evaluate them without weighing and balancing sociolinguistic and syntactic arguments, concerning every phenomenon, every case, and every situation. We shall discuss here only some of the grammatical areas in which variation is to be found. And we shall establish, in every case, the reasons why they are of interest for a reflection on contact. In these variant areas, we shall try to depart from what can be said to arise from inside French itself (endogenous, internal or structural variation Chaudenson et al 1993 would say intrasystemic), and what could, eventually, be said to be consequences of contact (exogenous variation, or intersystemic). In other words, I have to discuss the best decision as to which phenomena have to be included in the grammar, and which don t concern a grammar, but only a sociolinguistic description. 1.2. Variation and contact My argument in this communication will concern the possible amount of variability in a language, and the effects of contact on it. Looked at as «panlectal», variation in the French language spoken outside France shows at the same time convergent and divergent features, the second kind being those we are going to consider now. My discussion will be from a syntactic point of view, being mostly concerned in subtile forms of interferences, like convergence and surdetermination of phenomena. Everything in a language cannot be variable : the grammatical zones including phenomena of variation are not dispatched everywhere in the grammar. This observation led the linguists Roman Jakobson and André Martinet, in the fifties, to conceive a model differentiating between zones in a language : they opposed invariant (or mostly so) hard core zones («noyau dur»), and weak points in a system («points de faiblesse ou de fragilité du système»), which are the variant areas, where reconfigurations are more or less constantly ongoing. In French, we observe that from 17th century till now, processes of reconfigurability concern mostly the same points. If there is a part of optimality (or, less restrictively, auto-regulation), it cannot be taken as the leading force, as social considerations constantly impact upon linguistic variation. 2

If the term «variation» needs to be specified, so does «contact», even if it is conventionalized. We can adopt the tentative definition in Thomason s text : «contact is a source of linguistic change if it is less likely that a given change would have occurred outside a specific contact situation», p. 9. And, concerning syntactic facts, we can add : - a distinction between interference (concerning a rule) and borrowing (concerning a form) ; - a definition of convergence, as, for example, «the loss of a native form without counterpart in the other language, to the benefit ot a form having a counterpart in the dominant language» ; - a distinction between what is dysfunctional (a loss without compensation), and what is a reorganization or restructuring, the first one appearing only in a dying language. In fact, contact happens not between languages, but between speakers, or rather, as the sociologist Erving Goffman would put it, as a consequence of the multiple ordinary adjustments between speakers, in different language situations (what he called footing). It is the reason why we have to take into account the diversity between speakers through a community. This position is in opposition to that held by the linguist generally said to be the theoriser of language contact, Weinreich, who wrote (1953 : 41) : «In the interference of two grammatical patterns it is ordinarily the one which uses relatively free and invariant morphemes in its paradigms which serves as the model for imitation». Clearly, Weinreich was mostly looking for structural principles, and was not much concerned with sociolinguistic considerations. We shall then have to discuss the balance between grammatical and sociolinguistic arguments. My questions will then be : 1) Does grammatical borrowing happen (see e.g. Johanson 2002, according to whom syntax is not very open to borrowing)? 2) If yes, in which syntactic areas does it happen? 3) Are these areas the same in intrasystemic variation, and in effects of contact? If no, what are the differences? 4) Can we go as far as the hypothesis of borrowing acting on the regulation of the system? 1.3. Which geographical zones In the grammar, I take into account all varieties of French spoken in the world, including African. But for the sake of this communication, most of my examples come from studies on North American French : Quebec, Ontario, Acadie (New Brunswick, Prince Edward Island, Nova Scotia), Newfoundland, West of Canada, New England, Missouri and Louisiana. In these instances, the language in contact is always English, typologically not very far from French. As I adopt a syntactic point of view, I first consider a linguistic phenomenon, and not sociolinguistic 3

situations, which are quite different (e. g. Quebec well-established situation vs Missouri remanent French). I won t speak here of African French, even though it would be interesting from the point of view of the grammatical points concerned. But the discussion is difficult enough without introducing other factors of diversity, like the extra-family and home transmission 2. I am aware this position is neither obvious nor easy to hold, and I shall return to this point in my concluding remarks. 2. Examining some specific grammatical zones We are now going to consider some of the variant areas of French, selecting them on account of different linguistic properties : the verb system is the most strongly structured, clitic pronouns are a little bit less, prepositions are close to lexicon (a system, but more open), forms like back or comme are not included in systems ; and que-phrases concern the syntax of complex structures. 2.1. Some specificities of the verbal system (examples from different North American varieties) We are dealing here first with the generalization of avoir as an auxiliary (except with some verbs, like naître and mourir). Avoir generalization could be viewed as an influence of English (Mougeon showed in Ontario that the less the speaker is exposed to French, the more (s)he will generalize avoir). But this phenomenon is also to be found in every ordinary spoken variety around the world, in particular in Quebec and in France (examples from «français populaire» in Bauche, 1920). The rarefaction of the subjunctive, which will only be found after il faut que, or after some very frequent verbs like vouloir, could also be understood as an influence of English. Conclusion : In all these cases, there is a convergence between English and French. But as 1) they represent internal simplification, 2) there are also found elsewhere, we won t consider them as borrowings. 2.2. Word order of the clitic pronouns There is a tendency to eliminate whatever does not follow SVO word order, and especially preverbal clitics, as in il parle à nous-autres, rather than il nous parle (from Louisiana French, Valdman et al 2005, but to be found in many other places) See also je vais là, rather than j y vais (Ontario). Can this be considered an influence of English, where clitics and nouns follow the same word order? Other examples are certainly borrowings, like ma mère me fait manger z en (Newfoundland, King 1989, from eat some ). 2 I avoid the term «native speaker», which I feel too problematic to be freely handled. 4

Conclusion : clitics show two contradictory trends : reinforcement (mainly for subjects, as in nousautres) and deletion. As both can be understood as two different ways of simplification, and as both are also found elsewhere (for ex. in Belgium), there is no need to think in terms of borrowing, except as a reinforcement process. 2.3. Infinitive constructions with subject (Belgium) The non standard form passe-moi le journal pour moi lire (literally give me the newspaper for me to read it ) constitutes a recurrent case of discussion among grammarians. The first linguist who described this, Louis Remacle in 1952, coined it as Belgian, and a product of contact between French and the Germanic language spoken in Belgium, Flemish. Today, we know better that this form is not specifically Belgian (which is not really a surprise, as languages usually don t follow political borders). It is, at least, to be found in a big North-East quarter of France (also North, Lorraine, Alsace). However, it is also to be found in Louisiana, with examples like il a fait un gros plancher pour nousautres danser dessus (corpus Stäbler). But there is no possible relationship between Belgium and Louisiana, as most of the emigrants from the first diaspora came from the West of France (Normandie, Bretagne, Poitou). According to Neumann-Holzschuh 2005, this form would be one way of avoiding the strongly irregular conjugation of subjunctive. Brasseur 1998 also finds forms like i se ramassiont dans les maisons veiller (Newfoundland), which he calls a «substitute infinitive», or «a verbal form as bare, as reduced as possible». Conclusion : It is not necessary to consider the form pour moi lire to be an effect of contact, it can also be understood as : 1) simplified (no need to conjugate the subjunctive), 2) more explicit, because it formulates the subject. It could be reinforced by contact with a germanic language. 2.4. Prepositions (examples taken in different Canadian varieties) For some phenomena, the discussion is still open, between archaïsm, structural reorganisation, and borrowing from English. Prepositions constitute one of the most variable areas in French, even in FF (see the numerous hesitations between à, de and zero), as it supposes a major reorganization from the case-based mother tongue, latin. - prepositions with regime. Several usages of prepositions are certainly borrowings, some of them without obvious structural reasons (aller sur un voyage, go on a trip ), others with understandable structural reasons, when the French system offers no easy solution, as in l appareil est dessus ou sur (both coming from it is on, Ontario, Mougeon in Chaudenson et al 1993 : in FF, we would have to say l appareil est sous tension). Other forms are difficult to decide : j ai attendu 10 mn pour l autobus ( wait for ), could be a transfer, but it underlines a «lack» in French (imprecision : transitive constructions concerning very different kinds of objects). We can add that there is, in popular 5

French, a long tradition of reinforcing prepositions (see chercher après quelqu un), which make attendre pour not strange at all. So, it could be a surdetermination (intrasystemic weakness + intersystemic reinforcement). Conclusion : both phenomena exist, borrowing and surdetermination, in different examples. Prepositions appear particularly open to interference. - preposition stranding. Preposition stranding, which in FF is heavily constrained (to some preposition, or according to animate/inanimate), is much more open in North-American French (Roberge 1998, Arrighi 2005). This openness is often given as a borrowing from English, but many examples do not work in the same way in English and in Acadian French. In AF, preposition stranding never happens with à or de, except in some examples from Prince Edward Island French, an obsolescent variety where borrowing is heavy : qui c est que t as été à Ottawa au travers de - King & Roberge 1990 is very probably an interference. For Roberge 1998, «holes» are not of same origin in English and in French (F : resumptive pronouns which can be omitted ; E : product of a translation). Conclusion : the explanation by interference with English is only possible at a superficial level. In 17th century French, the possibilities where also more open : viens-tu avec? ( do you come with?, Brunot/Bruneau, for a person). Former possibilities of the language where later refused through normative principles. Knowing that preposition stranding is not a very frequent phenomenon among languages, the difficulty would be to explain why it should have been an object of borrowing from French to English, even in intensive contact conditions. 2.5. Relatives, and deletions of que The deletion of que is often given as an influence of English (the man I saw ; I think he came), and until recently it was often given as a characteristics of American French. But newly developped corpus linguistics allows to observe it is also a frequent phenomenon in France, as a colloquial feature for different kinds of que : tu veux je vienne? (example from Bauche, from Paris in 1920) ; c est maintenant tu l entends? ça fait longtemps elle est sortie (corpus FG, Parisian popular French). Other examples given in King 1991 are most certainly true borrowings, even sometimes with doubling : l argent which qu il a donné (Prince Edward Island French). Conclusion : obsolescent varieties must once again be taken apart, where true borrowings are to be found. But the usual deletion of que is neither a borrowing nor an interference. 2.6. The particle back (examples from Acadian French, New Brunswick) 6

The existence of a particle back in Acadian French is generally understood as an effect of intensive contact with English : it is specifically found in chiac 3. But Tremblay 2005 has shown that this is not the only hypothesis, and probably not the best one, for at least 3 reasons : 1) the distribution of back is not the same in English and in Acadian French (wider use in Acadian French, with examples like je voulais pas back aller) ; 2) the meaning is not the same in both languages (once again, it is wider in Acadian French) ; and 3) back does not behave like the other particles directly borrowed to English by chiac version of French (e. g. up or out, in finder out, on les frig pas up see Chevalier et al 2005). Conclusion : back can be said to contribute to solving a problem in French (ambiguity of revenir come for the second time, or go in the reverse direction ). Following Tremblay 2005, hypothesis of a relexification of a former arrière particle (like in et cil vont arrière à la nef, they go back to the ship ), which disappeared from Standard French at the end of the Middle Age, but not from the dialects : it can have emigrated to Canada with the emigrants. 2.7. A specific use of comme (Moncton) Specific usages of comme are found in chiac, but are also known in Quebec. For Perrot 1992, and Chevalier & Cossette 2002, examples of comme in New Brunswick can globally be understood by every francophone, even if (s)he is not a North American (more precisely, (s)he will have a global understanding, not necessarily a sophisticated one 4 ), like in ça fait comme dix minutes qu on parle (Perrot 1992). The discussion is still open for relationship to English like. Conclusion : Chevalier & Cossette argue against an influence of English : the putative influence of like can t be categorically refused, but it remain to explain why it is comme and not like (never found in this usage) which is used, in a variety which did borrow but, so, whatever, anyway, well, who cares 5, and so on. 3. Conclusions : what lessons can be drawn The synthesis we can draw concerns a confrontation between structural and sociolinguistic principles. 3.1. Concerning contact-induced change 3 Chiac is a mixed language from Moncton, in New Brunswick : it is a youth language, French-based but strongly anglicised, of course in the lexicon, like in j ai watché le show à la TV - but also in some grammatical features. 4 When reception is concerned, comprehension can intervene at different levels. See examples of «passé surcomposé» (Franco-provençal area), like in je l ai eu fait mais je le fais plus depuis longtemps : 1st level (global) : «past», 2 nd level (fine) : «remote past aspect». 5 The borrowing of discourse markers, conjunctions, and logical particles is a clear indication of intensive confrontation of languages (attestations of but and so, for example, as well in Ontarian, Acadian and Cadian French, but not in Quebecian). Weinreich 1953 already gave them as free for symbolic investissment, as they don t constitute a real need. 7

For studying contact, several observations can be made, from my sociolinguisticosyntactic point of view. 1) Sociolinguistic arguments appear to be more important than structural ones, as agency is a leading force in the process of borrowing and change. I follow Thomason 1997, for whom, if structural and sociolinguistic trends appear to be in opposition, then the sociolinguistic principle will be the leading force 6. Contact concerns mostly ordinary spoken varieties, most of the time used and transmitted in a network, outside formal teaching and writing, therefore developped outside norm pressures and normative conceptions. These varieties are in themselves very variant indeed, and we always have to wonder who is the speaker in the corpus, what type (s)he belongs to (frequent vs occasional speaker, educated vs non educated, ), what is his/her network(s), as well as, which was not at all considered in this communication, stylistic variation (different ways of speaking in different settings and situations). Particularly from Acadian and Ontarian situations, we learn the difference between a frequent user of French and an occasional speaker, restricted to home usage. The linguistic outcome won t be the same in Quebec, where French is quite well-established and where it is possible to live all day long in French, and e. g. in Prince Edward Island, where French is in a process of final attrition. 2) Sociolinguistc considerations concerns as much representations and ideology than actual socio-demographic rather quantifiable facts. Speakers have personal feelings about their language(s), e. g. their loyalty toward their mother tongue, or the desire to resist or even to fight the penetration of the dominant language, and these feelings are decisive to their linguistic attitude ; 3) To establish what is innovative in a language, the term of comparaison should never be the standard language («Standard French» or «Reference French»), but the type of language ordinarily spoken in different places (among which, France), in the same ordinary circumstances (colloquial, or «popular» French). And, as far as historical data are concerned, what can be reconstructed of what kind of French must have been spoken by emigrants, to become what it is today in those areas. So, before assuming that any phenomenon is an innovation, we have to confront it to our knowledge of the history of the language, as well as to its geographical variation. 3.2. Concerning languages, language change, and the possibility to study them Beside internal and sociolinguistic principles of explanation, we met principles which could be proposed as possible universals (concerning, for example, considerations on how human beings interact in ordinary circumstances). Thus we are led to question the idea there can 6 For example, p. 5 : «in each case the failure to fit the predicted linguistic situation is explained by a confounding social factor». 8

be an endogene evolution of a language : there is no point in wondering what could be the French language s endogenous evolution, as there is no strictly endogenous evolution, and languages always evolve in social circumstances, which necessarily implies contact between speakers. If people s language is not different, their style, or their «way of speaking» surely is. We crossed this problem through Weinreich s structural position, but we encounter the same problems with usual terms like «drift» or «trends» ; «needs» or «optimality» are more ambiguous, between structure and speaker. My final and concluding point will be to wonder why French grammarians (at least French, I don t want to over-generalize too hastily) are so quick to explain an unusual phenomenon (which is, most of the time, nothing more that non-standard) through interference, even when it is obviously in contradiction with empirical data. One possible answer is that interference is a way of saving the idea of homogeneity, and, perhaps, of purity of the language. Grammarians have a prfound dislike for variation, as it is so far away from their conception of a language, that they will prefer elements of explanation coming from outside, largely forgetting interaction between speaking agents. Bibliography BAUCHE Henri, 1920, Le langage populaire, Paris, Payot. BRASSEUR Patrice (dir), 1998, Français d Amérique. Variation, créolisation, normalisation, Actes du colloque Les français d Amérique du Nord en situation minoritaire, Université d Avignon, Centre d études canadiennes CECAV. BRASSEUR Patrice, 1998, «De l ellipse du pronom personnel aux formes verbales non marquées dans les parlers acadiens», in P. Brasseur (dir), 75-91. BRASSEUR Patrice & Anika FALKERT (Dir), 2005, Français d Amérique : approches morphosyntaxiques, Paris, L Harmattan. BRUNOT Ferdinand & Charles BRUNEAU, 1956, Précis de grammaire historique de la langue française, Paris. CHAUDENSON Robert, 1998, «Variation, koïnèisation, créolisation : français d Amérique et créoles» in Brasseur (Dir), 163-79. CHAUDENSON Robert, Raymond MOUGEON & Edouard BENIAK, 1993, Vers une approche panlectale de la variation du français, Paris, Didier-Erudition. CHEVALIER Gisèle & Isabelle COSSETTE, 2002, «comme, tic ou marqueur d oralité?», Port Acadie, Revue interdisciplinaire en Etudes acadiennes, 65-87. CHEVALIER Gisèle & Michael LONG, 2005, «Finder out, pour qu on les frig pas up, comment c qu i workont out : les verbes à particule en chiac», in Brasseur/Falkert (dit), 201-12. GOFFMAN Erving, 1981, Forms of talk, Philadelphia, University of Pennsylvania Press. JOHANSON Lars, 2002, «Do Languages Die of Structuritis? On the Role of Code-Copying in Language Endangerment», Rivista di Linguistica, 14 / 2, pp. 249-70. KING Ruth, 1991, «WH-words, WH-questions and Relative Clauses in Prince Edward Island Acadian French», Canadian Journal of Linguistics/Revue canadienne de linguistique 36, 65-85. KING Ruth & Yves ROBERGE, 1990, «Preposition stranding in Prince Edward Island French», Probus 2, 351-69. 9

NEUMANN-HOLZSCHUH Ingrid, 2005, «Le subjonctif en français acadien», in Brasseur/Falkert eds, 125-44. PERROT Marie-Eve, 1992, «Fonctionnement du marqueur comme à partir d un corpus chiac (région de Moncton, Canada)», in Approches énonciatives de l énoncé complexe, Louvain/Paris, Peeters, 21-30. REMACLE Louis, 1952, Syntaxe du parler wallon de la Gleize, Paris, les Belles Lettres, t. 1. ROBERGE Yves, 1998, «Les prépositions orphelines dans diverses variétés de français d Amérique du Nord» in Patrice Brasseur (Dir), 49-60. STÄBLER Cynthia, 1994, La vie dans le temps et asteur, Corpus de français de Louisiane, Tübingen, Niemeyer. THOMASON Sarah, 1997, «On the unpredictability of contact effects», web site Sarah Thomason. THOMASON Sarah, 2007, «Social and linguistic factors as predictors of contact-induced change», this conference. TREMBLAY Mireille, 2005, «Back en français acadien : archaïsme ou innovation?», in Brasseur/Falkert (dir), 263-73. VALDMAN Albert, Julie AUGER & Deborah PISTON-HATLEN, 2005, Le français en Amérique du Nord, état présent, Québec, Presses de l'université Laval. VINET Marie-Thérèse, 1996, 10