A real-time window on 19th-century vernacular French: The Rcits du franais qubcois d'autrefois

This article describes the construction of a corpus of spoken French with a time depth of a century and a half, the Récits du français québécois d’autrefois (RFQ). The folktales, local legends, and interviews constituting the RFQ were produced by speakers born between 1846 and 1895. They spoke the French of 19th-century rural Québec, a variety shown to be replete with the vernacular structures and inherent variability of contemporary dialects. The authors review the advantages and drawbacks associated with this type of diachronic material, and argue that, exploited judiciously, it effectively represents an earlier stage of spoken French. They show how systematic comparison of the RFQ with contemporary vernaculars can help pinpoint the existence, date, and direction of language change. (Apparent-time construct, Canadian French, corpus construction, diachronic data, language change, linguistic variability, ne deletion, real time analysis, vernacular)* I N T R O D U C T I O N The study of linguistic change has always been hampered by the dearth of appropriate real-time data reflecting earlier stages. Most older corpora of spoken language are plagued by shallow time depth, and written texts may not be representative of oral speech, where most changes originate. The lack of reliable diachronic evidence is at least partly responsible for the widespread assumption that salient features of contemporary vernaculars are recent innovations.1 This problem is particularly relevant to varieties of Canadian French, whose distincLanguage in Society 36, 707–734. Printed in the United States of America DOI: 10.10170S0047404507070662 © 2007 Cambridge University Press 0047-4045007 $15.00 707 tive and often stigmatized features are routinely ascribed to language change, whether caused by intense contact with English or by centuries of isolation from the supposedly standardizing influence of Metropolitan French. This article describes a novel use of underutilized recordings of moribund folklore to confront these issues: the construction of a corpus of spoken French with a time depth of a century and a half, which we call the Récits du français québécois d’autrefois (RFQ). The stories, local legends, and interviews constituting the RFQ were produced by speakers born between 1846 and 1895. They spoke the French of 19th-century rural Québec, a variety that, as we will show, is replete with the vernacular structures and inherent variability of contemporary dialects. As such, we argue that it effectively represents an earlier stage of spoken French. Systematic comparison of the RFQ with 20th-century vernaculars can help pinpoint the existence, date, and direction of language change. The construction of this corpus and the larger research program of which it forms part fall within the tradition of historical sociolinguistics. The aim is to situate spoken vernaculars, and in particular their nonstandard features, within the context of their trajectory of development. Comparison with appropriate earlier stages enables us to determine whether these features are retentions or innovations, and if the latter, whether contact-induced or internal developments. Much effort and ingenuity has been invested in locating earlier sources of the spoken language: in plays, diaries, letters, journals, dialect surveys, metalinguistic texts, and overseas varieties (cf. Ayres-Bennett 1994, 2000, 2004; Mougeon & Béniak 1994; Lodge 1996, 2003; Poplack 2000; Wüest 2002; Martineau & Mougeon 2003; Dufter & Stark 2005; Poplack et al. in preparation). Studies based upon them have yielded valuable insights into what the French spoken in the past must have been like; their authors have nonetheless been quick to caution against the facile equation of these documents with the oral usage of the time, especially for purposes of dating variant forms. Perhaps the greatest problem involves the issue of negative evidence. It is not clear how to interpret the absence of a form: because it was not in use, because the writer did not happen to use it, because it was not appropriate to the genre, or because normative conventions blocked literary usage of (nonstandard) variants that may nonetheless have proliferated in speech. Attestations can also be problematic, since French writers often employed literary conventions to represent, for comic effect, dialectal features that may not have been part of anyone’s vernacular (e.g., Wüest 2002; Ayres-Bennett 2004). In addition, social information about the writers and the community is often lacking. It is in this context, as we will show in the last section of this article, that the oral recordings constituting the RFQ are particularly valuable. But these materials must also be used with caution. Potentially detracting from their utility as an earlier stage are the possibilities that the original speakers recorded by the folklorists were not representative of 19th-century Québécois, and that their speech was not representative of the French of the time, whether because of the prevaS H A N A P O P L A C K A N D A N N E S TA M A N D 708 Language in Society 36:5 (2007) lence in the recordings of ritualized expressions characteristic of performed speech genres, or superposed variants incurred by the constraints of the data elicitation procedures. An additional question, applicable to all linguistic studies carried out in real time, is whether two cohorts of speakers, born more than a century apart (1846 and 1965), may meaningfully be compared. We address each of these issues, arguing that, exploited judiciously, the RFQ is a valuable source of data on 19th-century Québec French. In what follows, we first outline the methodological assumptions underlying this work, then describe the context in which the materials constituting the RFQ were collected, since this is what makes them especially valuable for the study of linguistic variation and change. We next detail the construction of the RFQ and the methods employed to render it linguistically useful for our purposes. Finally, we illustrate the utility of the corpus for the study of linguistic change with an analysis of one of the most widely documented features of the French language – the variable deletion, or as some would have it, reinsertion – of the negative particle ne. Despite the prodigious amount of scholarly attention accorded this variable, and the variety of texts examined, there is remarkably little consensus on when the process of ne deletion gained ground. The oral data of the RFQ constitute a valuable adjunct to this discussion. M E T H O D O L O G I C A L U N D E R P I N N I N G S The apparent-time construct Efforts to reconstruct the spoken language of an earlier century inevitably rely on the assumption that speakers’ vernaculars remain stable over their lifetimes. While this assumption has been widely accepted by students of language change (Labov 1966, 1994; Bailey et al. 1991; Lightfoot 1999; Anderson & Lightfoot 2002; Bailey 2002), only recently has the kind of cross-sectional and longitudinal research necessary to test it empirically become available (G. Sankoff 2005). Such evidence consists of panel studies, successive interviews of the same individuals at (at least) two points in time. Taken together, the results of the panel studies currently available provide strong confirmation that adult vernaculars generally remain stable over the lifetime (e.g., Labov 1994; Labov & Auger 1998). Where change has been reported, it has almost without exception involved the phonology (e.g., Gordon et al. 2004; but see Baugh 1996), teenagers (Cukor-Avila 2000), variables involving high levels of social awareness (Labov 1994:112), and0or situations of dialect contact (Gordon et al. 2004:261– 62). None of these conditions is pertinent here. And it is usually the case that only some, rather than all individuals were involved (Baugh 1996; G. Sankoff et al. 2006). Older speakers in particular were found to participate only marginally in the developments taking place around them, even where the community as a whole was involved in vigorous change (Labov 1994). With regard to morphosyntax, nearly all studies have reported stability (e.g., Daveluy 1987; Lessard A R E A L T I M E W I N D O W O N 1 9 T H C E N T U R Y V E R N A C U L A R F R E N C H Language in Society 36:5 (2007) 709 1989; Bailey et al. 1991; Cukor-Avila 2000; Bailey 2002; G. Sankoff et al. 2006). In sum, while we now know that change across the lifespan can occur, the available evidence suggests that it is the exception, not the rule. Where attested, it tends to involve especially those members of the community who engaged in long-term accommodation to a second dialect, and to be resisted by older speakers. The primacy of the vernacular Much of the variability of interest to sociolinguists involves nonstandard or stigmatized linguistic forms, and their trajectory over time. Because these are concentrated in the informal speech style known as the vernacular (though generally avoided in formal contexts and in writing), our comparative enterprise relies crucially on the possibility of tapping into such forms. Labov 1966, 1984 defines the vernacular as the style used unreflectingly when speakers are not monitoring their speech. In contrast to the more formal superposed variants acquired later in life and used sporadically as circumstances dictate, the rules governing linguistic variation in the vernacular, which is acquired in preadolescent years, are regular. Indeed, Labov (1984:29) characterizes the vernacular as “the most systematic data for linguistic analysis.” Moreover, as the first dialect acquired, the vernacular can be expected to remain the most stable, forming the basis of historical continuity among successive language states. Thus the vernacular affords the truest indication of how far a change has progressed (Coveney 1996:278). The elicitation techniques employed by the folklorists who collected the data on which the RFQ is based were particularly conducive to the use of this speech style, as we detail in ensuing sections. The RFQ speakers acquired their vernaculars in the latter half of the 19th century; to the extent that they have remained stable, we can take them as a reflection of the speech of the time. T H E O R I G I N A L C O L L E C T I O N S The folklore tradition in French Canada The recordings from which we constructed the RFQ had been collected throughout French Canada in the 1940s and 1950s by the folklorists Luc Lacourcière, Carmen Roy, and their associates (Lacourcière 1946; Roy 1981), as part of an ethnographic tradition flourishing since the beginning of the century. Driven by the romantic ideals of documenting for posterity the “beauties as well as the value of bygone traditions” 2 (Lacourcière & Savard 1951:86), their goal was to create an authentic picture of French Canadian folklife, exemplified by tales, legends, songs, jokes, and assorted terminology. To achieve this goal, the folklorists turned to habitants, or country people, specifically targeting the “elite of the illiterate” (Lacourcière 1959:32), whom they considered to constitute the most direct conduit to oral tradition. Most were men over the age of 70 (Lacourcière & Savard 1950:64), veterans of the Québec logging and fishing industries, and S H A N A P O P L A C K A N D A N N E S TA M A N D 710 Language in Society 36:5 (2007) familiar with the associated lore. Women were less in demand in this connection (Lacourcière 1959:33), although fortunately for our purposes, some had been recorded. The folklorists sought them in isolated rural areas of Québec, wherever the art of storytelling had survived (Lacourcière 1961:6). Scant information is available regarding the educational levels or occupations of the original storytellers, but as far as we can reconstruct, there was little heterogeneity among them. On the contrary, the folklorists’ avowed research interests practically limited participants to insular farmers, loggers, and fishermen with little or no formal education, corresponding to the non-mobile older rural males (Chambers & Trudgill 1980) coveted by traditional dialectologists for their conservative characteristics. It is unclear just how large the original collections were, since no log is available. By Lacourcière’s own tally, the Archives de Folklore collection was huge: in 1966, it contained 20,000 recordings of 6000 individuals distributed over 4000 locales across French Canada (Lacourcière 1966:226–27), collected by some 40 fieldworkers (Lacourcière 1962:254). Roy’s data, concentrated in the Gaspé region of Québec, were far more modest, collected personally, and not so well documented. They made their recordings on disks or magnetic ribbon (which Lacourcière claims to have pioneered for these purposes); these were later copied onto reels. The folklorists’ field methods were exemplary (and in fact served as a model for much subsequent sociolinguistic work), consisting for the most part of participant observation achieved over “long and repeated stays in the region” (Lacourcière & Savard 1953:99). These permitted them to build what Lacourcière called a “climate of collaboration,” which he describes as “a product of trust. It implies lengthy dealings, intimate contact. It can be achieved, but only if you cast aside all prejudice, share in popular life, melt into it, so to speak, until any foreign characteristic is completely erased” (Lacourcière & Savard 1953:99).