1. Introduction
The expression of future temporal reference (FTR) has been widely studied across spoken French varieties. For centuries, grammarians (Maupas Reference Maupas1607, Antonini Reference Antonini1753) described the choice between the two main variants – the periphrastic future (PF), as in ça va être cette année ‘it's going to be this year’, and the inflected future (IF), as in ça sera au mois d'octobre ‘it will be in October’ – as influenced by temporal distance; periphrasis was argued to denote proximity (i.e., le futur proche). Empirical studies of most varieties of Acadian French (King and Nadasdi Reference King and Nadasdi2003, Comeau Reference Comeau2015) support this description. By contrast, the strongest predictor in Laurentian (Poplack and Turpin Reference Poplack and Turpin1999, Wagner and Sankoff Reference Wagner and Sankoff2011) and Continental (Roberts Reference Roberts2012) French varieties studied so far has been sentential polarity: negative clauses strongly favour the inflected future. Thus, varieties of spoken French, both within Canada (Acadian vs. Laurentian) and globally, appear to be divided between two types of systems with respect to the linguistic constraints operating on the future temporal reference variable.
In this article, we examine FTR in a corpus of French interviews collected in the mid-2000s in Vimeu, a rural area of northern France where Picard, a Gallo-Romance language, still enjoys relative vitality. By analyzing FTR in this contact variety, we not only expand on the very few variationist studies of this variable in Continental French, but also tap into the role that Picard may play on French morphosyntactic variation in the area. To determine the factors that condition FTR variant choice, we analyze spoken French data from Picard–French bilinguals and French monolinguals and consider a number of social and linguistic factors proposed in the literature. We show that despite parallels with Roberts’ (Reference Roberts2012) study of Continental French (e.g., similar rates of variants and effect of socio-stylistic factors), FTR in Vimeu French is constrained not by sentential polarity but by temporal distance, with imminent events strongly favouring the PF.
In section 2, we present the dependent variable and review the literature pertaining to its distribution and to the linguistic and social factors claimed to condition its usage in French. In section 3, we offer an account of FTR in Picard, the substrate language in the area, and describe the French corpus under study. In section 4, we circumscribe the variable context and outline the factors (both linguistic and social) which we hypothesize may have an effect on variant choice in Vimeu French. Section 5 presents our results, first for social factors then for linguistic factors, followed by a short discussion of the apparent stylistic effect of negative ne retention. A comparison across varieties for the only significant linguistic constraint, temporal distance, is provided in section 6, followed by concluding remarks.
2. Future temporal reference (FTR) in French
The expression of future temporal reference (that is, to express that an eventuality will occur posterior to the moment of speech) in French has been the subject of much debate. While the inflected future and futurate present were found as early as Latin (Fleischman Reference Fleischman1982), the first attestations of the periphrastic construction to express futurity in French, after having functioned as a verb of spatial movement, date back to the thirteenth century (Champion Reference Champion1978: 25). According to Wilmet (Reference Wilmet1970: 191), it is during the Middle French period, when Francien was further differentiating itself from other oïl varieties (e.g., Norman, Picard, Poitevin, or Saintongeais), that the competition between the periphrastic future and the inflected future accelerated; the periphrastic form was claimed to present the future as a continuity of the present.
Beginning in the seventeenth century, grammarians (Maupas Reference Maupas1607, Antonini Reference Antonini1753) sought to account for the diverse functions of FTR forms and associated the periphrastic construction with proximity to the moment of speech. A 2009 paper by Poplack and Dion provides a synthesis of grammarian and remarqueurs’ commentary pertaining to the FTR variable dating from 1530 to 1999 (though the bulk of the sources are from the nineteenth and twentieth centuries). While they uncover that many grammarians disagree with respect to the functions of the variants, they found the most agreed-upon association to be between the periphrastic future and proximity, which likely gave rise to its enduring label, namely le futur proche (or prochain) (Poplack and Dion Reference Poplack and Dion2009: 568).
Today, most varieties make use of three main forms: the inflected future, a synthetic form whereby a suffix attaches to the lexical verb, as in (1); the periphrastic future, which involves the semi-auxiliary aller ‘to go’ followed by the lexical verb, as in (2); and the futurate present, where the present indicative form conveys future temporal reference, as in (3).
(1) J'te montrerai ça tout à l'heure. (Michel G., 199)Footnote 1
‘I'll show that to you in a bit.’
(2) Là vous allez partir au mois de… fin juin? (Guy D., 231)
‘So you're going to leave in … at the end of June?’
(3) On part en Écosse fin août (Lysiane D., 528)
‘We leave for Scotland at the end of August.’
While all three variants are found in most varieties of modern spoken French, their relative proportions differ greatly from one variety to the next, as shown in Table 1. Within Canada, Laurentian French varieties – Quebec French and its offshoots in Ontario, western Canada and beyond – show the highest proportions of the PF variant, ranging from as low as 73 per cent in Ottawa-Hull to as high as 86.5 per cent in Hawkesbury. There is considerable diversity among varieties of Acadian French – spoken primarily in Canada's Atlantic Provinces (New Brunswick, Newfoundland and Labrador, Nova Scotia, and Prince Edward Island) and in eastern parts of Quebec. The varieties spoken in New Brunswick generally show higher rates of the PF (77.1% and 75.2%) than those spoken in other Atlantic provinces (as low as 41.4% in Prince Edward Island).Footnote 2
Overall distribution of FTR variants in spoken French
Once the comparison is extended to varieties spoken outside of Canada, we see an even greater range in terms of rates of the variants (e.g., 27.3% PF in 1950s to 58.8% in the 1980s). In contrast with Continental French, Martinique French displays rates similar to those found in Canadian French varieties (i.e., 72.3% PF). Fleischman comments that “[t]hroughout Romance this trend toward periphrastic futures appears to be more pronounced in the overseas regions (American Spanish, Canadian French) and, obviously, in the creoles, than in the respective continental dialects” (Fleischman Reference Fleischman1982: 101–102). Thus, it may be that the higher rates of the PF in varieties spoken beyond continental Europe reflect a phenomenon not unique to French, but one that is found more generally in Romance languages. Despite such rate disparities between varieties, which may also be attributed to internal distinctions across varieties or to differences in corpora collection (e.g., formality, conversation topic), we notice somewhat consistent trends in the linguistic constraints which operate on the variable, as detailed in sections 2.1 and 2.2.
2.1 FTR in Canadian French varieties
Contrary to the association between the periphrastic future and proximity of the eventuality to the moment of speech as purported by grammarians, studies which have examined FTR in Laurentian French spoken corpora (Deshaies and Laforge Reference Deshaies and Laforge1981, Emirkanian and Sankoff Reference Emirkanian and Sankoff1986, Poplack and Turpin Reference Poplack and Turpin1999, Blondeau Reference Blondeau, Papen and Chevalier2006, Grimm and Nadasdi Reference Grimm and Nadasdi2011, Wagner and Sankoff Reference Wagner and Sankoff2011) found that another constraint governs choice between the future forms: sentential polarity (i.e., whether a context is affirmative or negative). In all Laurentian varieties studied to date, negative contexts have been found to highly favour the inflected future while affirmative contexts favour the periphrastic future. The correlation between polarity and the future forms, first noted in Ile-aux-Coudres (Seutin Reference Seutin1975), was repeatedly confirmed for Laurentian French varieties spoken in Quebec City (Deshaies and Laforge Reference Deshaies and Laforge1981), Montreal (Emirkanian and Sankoff Reference Emirkanian and Sankoff1986, Blondeau Reference Blondeau, Papen and Chevalier2006, Wagner and Sankoff Reference Wagner and Sankoff2011), Ottawa-Hull (Poplack and Turpin Reference Poplack and Turpin1999, Poplack and Dion Reference Poplack and Dion2009) and Hawkesbury, ON (Grimm Reference Grimm2010, Grimm and Nadasdi Reference Grimm and Nadasdi2011), among others. However, these results are not confirmed in all varieties of Canadian French.
The other main Canadian French dialect group, Acadian French, is distinct from Laurentian French in a number of ways, including the FTR variable. For instance, in some of the most conservative varieties spoken in Nova Scotia (e.g., the Baie Sainte-Marie and Pubnico varieties), negation is expressed with point ‘not’ more frequently than with pas (Flikeid Reference Flikeid, Mougeon and Beniak1994: 295), a tendency also found in Picard (Auger Reference Auger, Landrecies and Petit2003c: 22).Footnote 6 With regard to the FTR variable, the pattern for most Acadian varieties is clear: sentential polarity does not influence variant choice. This finding has been confirmed for the conservative variety of Baie Sainte-Marie, located in southwest Nova Scotia (Comeau Reference Comeau2015), as well as in the varieties spoken in Prince Edward Island and Newfoundland (King and Nadasdi Reference King and Nadasdi2003). Yet, mixed results emerge from Chiasson-Léger's (Reference Chiasson-Léger2014) recent study of Acadian varieties spoken in New Brunswick, specifically in the southeastern city of Moncton, and in the northeastern town of Shippagan, located in the majority French region known as la péninsule acadienne. In the southeast, she notes that both sentential polarity and temporal distance constrain variant choice whereas the northeastern variety more closely resembles Laurentian French in that there is no temporal distance effect, rather it is polarity which conditions the variable. These conflicting results point to a perhaps unsurprising fact of Acadian varieties: they are not homogeneous. In fact, while we can only hypothesize as to the origins of the polarity constraint in New Brunswick Acadian French, its absence from the more conservative varieties (Baie Sainte-Marie, Prince Edward Island, and Newfoundland) suggests that it may result from contact with speakers of varieties in which the polarity constraint is operative (e.g., Laurentian French). Further research may shed light on the actuation of this constraint, the discussion of which remains beyond the scope of the present article.
2.2 FTR in other French varieties
Beyond the Canadian context, variationist work on French FTR has been limited to Martinique and, most relevant for our study, Continental France. In his analysis of Martinique French, Roberts (2014) found that the constraints governing the use of FTR closely parallel the Acadian varieties studies to date in that the only predictor of variant choice is temporal distance. The findings are not identical in the two varieties; he found a correlation between the IF and distant eventualities, but none between the PF and proximity, the strongest effect found in Acadian French. As in most varieties of Acadian French (with the exception of some varieties spoken in New Brunswick), sentential polarity does not constrain variant choice in Martinique. The fact that temporal distance is the strongest constraint in both varieties suggests that it may 1) have had a stronger effect than previously thought on seventeenth-century French and Gallo-Romance varieties, 2) have been operative in actual spoken language, and consequently 3) have legitimately attracted the attention of early grammarians and remarqueurs.
Roberts’ (Reference Roberts2012) study of Continental French focuses on the Beeching's (Reference Beeching2002) corpus of spontaneous spoken French, a collection of different types of oral data (interviews, announcements, etc.) collected between 1980 and 1990 in four areas of France: Lot, Minervois, Paris, and Brittany. Roberts’ results reveal a linguistic system similar to what has been found in Laurentian French: sentential polarity is the greatest predictor of variant choice with negative contexts highly favouring the inflected future variant. This link between Laurentian and Continental French raises questions as to the origins of the polarity constraint and to its possibly widespread nature. Did it originate in seventeenth-century France or did it develop independently on either side of the Atlantic? This question will most likely remain unanswered until variationist work on older corpora is undertaken. In any case, what emerges from the comparison of these studies is a largely bipartite division with regard to the FTR variable: in some varieties (conservative Acadian, Martinique), temporal distance constrains choice of the variants while in others (Laurentian and Continental), it is the polarity constraint.
3. FTR in a contact variety of continental French
In this article, we examine FTR in another variety of Continental French: that which is spoken in the rural Vimeu area of Picardie. By examining a single regional variety of Continental French and by controlling the type of data included in the corpus (i.e., sociolinguistic interviews), we seek to determine whether this variety displays variant distribution and conditioning similar to the four Continental French varieties examined by Roberts. Furthermore, the choice of community under investigation, the Vimeu area, allows us to consider whether long-standing contact with Picard, a language with high rates of the IF, may have influenced the FTR system in Vimeu French.
Located in the Picardie region of France, the Vimeu area occupies the westernmost part of the Somme département along the Bay of Somme and is bordered to the north by the Somme River and to the south by the Bresle River and Seine Maritime (Upper Normandy). According to INSEE (2002), the area is predominantly rural; urban areas are comprised of small towns, the largest of which, Friville-Escarbotin, had a population of only 7,000 in 1999.
Relative to other Picard-speaking areas of France, Vimeu represents an area of high regional language maintenance and vitality. According to a 1999 survey (Blot, Éloy, and Rouault Reference Blot, Eloy and Rouault2004), a greater proportion of adults heard Picard growing up or spoke it with friends and family in the Somme (27.3%) than in the Picard-speaking départements of Pas-de-Calais (22.2%), Nord (10.0%), Oise (3.7%), or Aisne (2.7%). Within the Somme, Vimeu stands out as particularly vibrant for the Picard language: “Today all observers appear to agree that Picard is disappearing as a first language […], though it seems to have fared better in those parts which were less severely affected by the 1914–18 war, such as the Vimeu area along the coast south of St. Valéry” (Coveney Reference Coveney1996/2002, following Carton Reference Carton and Tabouret-Keller1981). Today, even in Vimeu, most public activities are conducted in French and regular Picard use is limited to older speakers. Although the Picard language remains strongly tied to factory work and hunting, a hobby popular among Vimeu male residents, there are essentially no remaining monolingual speakers of Picard. The fact that this regional language still enjoys some vitality in the area (Carton Reference Carton and Tabouret-Keller1981, Auger Reference Auger, Pusch and Raible2002, Blot et al. Reference Blot, Eloy and Rouault2004), unlike in neighbouring Nord-Pas-de-Calais cities, and that it has been in intense contact with French makes Vimeu quite interesting linguistically.
3.1 Vimeu Picard and French morphosyntax: the case of FTR
Traditional Vimeu Picard and colloquial French share many morphosyntactic features, such as subject doubling (4) and the use of the avoir/avoér ‘to have’ auxiliary with verbs of movement and pronominals (5). These features, which distinguish both varieties from so-called ‘Standard French’, are central to the emerging Vimeu Picard standard (Auger Reference Auger, Landrecies and Petit2003c); the use of subject doubling has even been referred to as a ‘badge of Picard identity’ (Auger Reference Auger, Joseph, DeStefano, Jacobs and Lehiste2003a, Coveney Reference Coveney2005). Yet, metalinguistic comments made by some of our speakers about French teachers’ negative attitude toward these very features in school (Auger and Villeneuve Reference Auger and Villeneuve2015) indicate that the use of some Picard-like variants is somewhat stigmatized in ‘good’ French.
(4) Capieu il avoait ouvért sin carrieu. (Auger Reference Auger2003b: 385)
(Fr.: Capieu avait ouvert sa fenêtre.)
‘Capieu he had opened his window’
(5) Iz ont vnus s'présenter. (Auger Reference Auger, Landrecies and Petit2003c: 21)
(Fr.: Ils sont venus se présenter.)
‘They came to introduce themselves’
(6) I n’ont pus rien à manger. (Auger and Villeneuve Reference Auger, Villeneuve, Meyerhoff and Nagy2008: 242)
(Fr.: Ils n'ont plus rien à manger.)
‘They don't have anything to eat anymore’
With respect to a number of other morphosyntactic variables, the traditional Picard variant is the one found in ‘Standard French’. For instance, Auger and Villeneuve (Reference Auger, Villeneuve, Meyerhoff and Nagy2008) have shown that the deletion of negative ne, which is widely documented in colloquial spoken French (see Ashby Reference Ashby2001 for Continental French, and Poplack and St-Amand Reference Poplack, St-Amand, Martineau, Mougeon, Nadasdi and Tremblay2009 for Laurentian varieties in Canada, among many others), is rarely found in Picard (6). The expression of FTR in Picard displays a similar pattern: the relatively frequent PF reported for most French varieties (see Roberts Reference Roberts2012 for Continental French, and Poplack and Dion Reference Poplack and Dion2009 for Laurentian varieties) is rarely found in that language; it is the IF which is typically used to express FTR. In fact, although French grammarians have commented on the periphrastic form for centuries, Champion's (Reference Champion1978) survey of Gallo-Romance sources found that the PF “is mentioned just in passing, if at all, and it apparently plays no important role in the verbal system of these dialects” (1978: 51). In fact, the only evidence of PF he noted from Picard-speaking France (i.e., Chaurand Reference Chaurand1968: 205) comes from the Aisne, a département with low Picard vitality. Similarly, our own survey of Picard grammars and conversation guides did not yield a single mention of the PF (8); only the IF (7) is described.Footnote 7
(7) Espérons qu'o fra biétôt la paix. (Vasseur, Lettes 120, January 1946)
‘Let's hope that we will soon make peace’
(8) O va meume li foaire eune féte au mouos d'moai. (Vasseur, Lettes 132, March 1946)
‘We will even throw him a party in May’
Yet, the PF is not completely absent from written Picard. Once we consider diachronic data from textual sources, we do find some tokens of the PF. Auger and Villeneuve's (Reference Auger and Villeneuve2015) analysis of 393 tokens of FTR in Gaston Vasseur's Lettes à min cousin Polyte ‘Letters to my cousin Polyte’, a weekly chronicle published in the Bresles et Vimeuse newspaper between 1938 and 1971 – data is comprised of a 10-month sample for each of 1946, 1956 and 1966 – reveals an average of 16.5 % PF in the mid-twentieth century (cf. Table 2). Despite the semblance of a slow increase in PF (χ2(2) = 3.136; p = 0.208), the IF remains the default marker of futurity in Picard over this 20-year period.Footnote 8
Overall distribution of FTR variants in written Vimeu Picard (from Auger and Villeneuve Reference Auger and Villeneuve2015)
Due to the overwhelming rates of the IF in Picard and to the long-standing bilingualism and language contact situation in Vimeu, we sought to answer a number of research questions with respect to the variety of French spoken in the Vimeu area:
1. What is the distribution of FTR variants in Vimeu French? How do these rates compare with those reported for 1980s Continental French (Roberts Reference Roberts2012)? In light of Fleischman's (Reference Fleischman1982) observation regarding FTR in Europe and overseas, we hypothesize that Vimeu French will display lower rates of the PF than Laurentian and Martinique French varieties, but comparable to other varieties of Continental French (or slightly lower, given speakers’ exposure to the low PF rates of Vimeu Picard).
2. Is FTR in Vimeu subject to the same socio-stylistic conditioning as other varieties of French; that is, increased rates of the IF correlated with higher socioeconomic categories and with the use of other formal variants (e.g., vouvoiement, negative ne)? We anticipate a socio-stylistic effect similar to that found in both Canadian and Continental French: the IF will be favoured by speakers with higher levels of education and will correlate with other formal linguistic variants.
3. Does the bilingualism status of the speaker influence variant choice? Specifically, does proficiency in a language which appears to prefer the IF lead to bilinguals exhibiting lower rates of the PF that their monolingual French counterparts? Indeed, we hypothesize that Vimeu Picard's preference for the IF and low overall rate of the PF may create a favourable environment for correspondingly lower overall rates of the PF in Vimeu French generally, and more specifically in Picard–French bilingual speakers.
4. Lastly, which language-internal (linguistic) factors constrain the variable expression of FTR in this variety? Is temporal distance a predictor of variant choice, as suggested by grammarians referring to the PF as le futur proche ‘the near future’, or is sentential polarity a better predictor, as in Canada's Laurentian French varieties? With regard to the linguistic conditioning, we anticipate that our results will, like Roberts’ for Continental French, show that sentential polarity conditions variant choice (i.e., negative contexts will favour the IF).
3.2 The 2006–2007 Vimeu French corpus
Data for the present study are extracted from a corpus of French sociolinguistic interviews with monolingual French and bilingual Picard–French speakers collected by the first author during the summers of 2006 and 2007; conversation topics included childhood, schooling, family life, language use, and local customs. Because of the difficulties of obtaining a random sample in France (Blanc and Biggs Reference Blanc and Biggs1971) and Vimeu people's professed wariness of strangers (Vasseur Reference Vasseur1969), subjects were contacted using the second order relationship method (Milroy Reference Milroy1987). The fact that J. Auger had remained in contact with Vimeu residents since her initial fieldwork in the late 1990s ensured the success of the ‘friend of a friend’ recruitment method. To avoid eliciting Picard, the language variety elicited by Auger, all 31 speakers were told that the interview would be conducted entirely in French. This method guaranteed the elicitation of what Carton (Reference Carton and Tabouret-Keller1981: 17) describes as français d'intention ‘intended French’. Each 45- to 60-minute-long interview took place at the subjects’ home and was digitally recorded; data for this study are extracted from a sub-corpus of over 18.5 hours of recorded speech from 31 speakers.
Speakers from the 2006–2007 Vimeu French corpus are classified according sex and divided into two age groups: adults (25 to 54 years old) and elders (55 years and older). These two age groups were selected as a reflection of a societal shift in the level of exposure to regional languages: people born in the 1950s were less likely to have spoken a regional language with their parents as children than people born in the 1930s (Clanché Reference Clanché2002). In his study of regional French in Lille, Pooley (Reference Pooley1996, Reference Pooley2004) found that distinctive Picard features are virtually absent from the speech of speakers born after 1953. In Vimeu, exposure to Picard was likewise more limited for older adults, which may give rise to fewer Picard-like variants in adults’ French language. Beyond the potential influence of Picard on French in the Vimeu community, the effect of bilingualism at the individual level is also assessed by comparing the French output of Picard–French bilinguals to that of French monolinguals. The effect of level of education is also measured through a three-way division of the speaker sample: no baccalauréat (the French equivalent of a high school diploma), baccalauréat, and university education.
4. Methodology
4.1 The envelope of variation for FTR in Vimeu French
We defined the variable context as all tokens that unambiguously express future temporal reference, excluding reported speech, code-switches, metalinguistic comments, and false starts and hesitations. Following previous studies, we also excluded cases in which the verb aller is used as a verb of motion (9), examples in which a future form is used to express an imperative (10), lexicalized expressions, such as qu'en dira-t-on ‘gossip’, and the various forms of the discourse marker on va dire ‘let's say’ (N = 18).
(9) C'est pas grave, on va aller habiter à Mons-Boubert. (Annick M., 423)
‘It's okay, we are going to go live in Mons-Boubert’
(10) Vous l’laisserez à [Lysiane D.]. (Béatrice D., 700)
‘Give it back to Lysiane’
As in previous studies, PF and IF forms were often used to express “a conjectural tone” (Emirkanian and Sankoff Reference Emirkanian and Sankoff1986) or to refer to timeless truths or hypotheticals, as shown in (11).
(11) Mais en ville, tu parleras pas picard ici. (Albert D., 294)
‘But in town, one doesn't speak Picard here’
Since they do not belong to the same dependent variable as actual expressions of futurity, these gnomic futures (N = 109) are excluded from the current study.
4.2 Social and linguistic factors
Based on the sample from the Vimeu French corpus, we were able to assess the potential effect of a number of social factors on variant choice: sex, age (25–54 years-old or 55 and older), education level (no baccalauréat, baccalauréat, or university degree) and bilingualism status (French monolingual or Picard–French bilingual). The inclusion of sex and age allows us to test whether there is a potential change in progress with regard to the FTR variable. The consideration of education level on FTR helps us identify any socio-stylistic effect, comparable to what was found in Laurentian French (Emirkanian and Sankoff Reference Emirkanian and Sankoff1986, Grimm and Nadasdi Reference Grimm and Nadasdi2011, Wagner and Sankoff Reference Wagner and Sankoff2011) and in Continental French (Roberts Reference Roberts2012). While we hypothesize that contact with Picard might have had an effect on FTR in Vimeu French at the community level, the comparison between monolingual French speakers with bilingual Picard–French speakers allows us to further test the potential effect of Picard on individuals.
We also coded for the following linguistic factors: temporal distance, sentential polarity, adverbial specification, and subject type, each detailed in sections 4.2.1 to 4.2.4.
4.2.1 Temporal distance
Based on the fact that temporal distance has been reported for both Acadian and Martinique French varieties and since it is also mentioned by grammarians and remarqueurs, we coded each token based on the temporal distance between the time of speech (t0) and the time of the future event (t1). Although Poplack and Turpin (1999: 150) found “no systematic association of form with fine degree of temporal proximity” and “collapsed the data into proximal, including events occurring within the day, […] and distal contexts”, we initially opted for a finer degree to allow a comparison with French varieties in which the temporal distance constraint is operative. Thus, we distinguished between imminent events likely to occur within the minute (12) and future events likely to occur within the hour (13), the day (14), the week (15), the year (16), or farther into the future (17).Footnote 9 Events that began in the past but that perdure into the future were coded as continual, and those that could happen at any point into the future were coded as indeterminate.
(12) Là ça va pas être très discret hein… (Sophie F., 172)
‘Now it's not going to be very discreet…’
(13) On en parlera peut-être après si on parle e-d'la commune. (Paul B., 149)
‘Maybe we'll talk about it later if we talk about the commune.’
(14) Dès qu'vous allez partir, j'vais partir au jardin. (Norbert T., 733)
‘As soon as you leave, I'm going to go to the garden.’
(15) Madame, elle va aller chez toi bientôt. (Monique F., 261)
‘The lady is going to go to your house soon.’
(16) Moi j'pense même que ça va être cette année. (Fabienne A., 53)
‘I even think it's going to be this year’
(17) Le mouvement va s'inverser dans les années qui viennent peut-être. (Annick M., 474)
‘The movement is maybe going to go in the opposite direction in years to come.’
If Vimeu French patterns like Martinique French or Acadian French, we might expect that either proximal contexts favour the PF or distal contexts favour the IF (or both). Conversely, if Vimeu French patterns along the same lines as Laurentian French or other varieties of Continental French, we would expect this factor group to not constrain variant choice at all.
4.2.2 Sentential polarity
We include sentential polarity given its strong effect in both Laurentian and Continental French. We coded each token based on whether the token occurred in an affirmative clause (18) or in a negative clause (19).
(18) À Abbeville, vous trouverez un… une petite ville bourgeoise. (René D., 631)
‘In Abbeville, you'll find a small bougeois town.’
(19) L'expression, j’vais pas la trouver. (Jean-Pierre D., 732)
‘The expression, I'm not going to find it.’
If Vimeu French patterns like Laurentian and Continental French, we would expect the IF to be favoured in negative contexts.
Despite a documented change toward the disappearance of the negative particle ne in Continental French (Ashby Reference Ashby2001), which has virtually reached completion in North American French (Martineau and Mougeon Reference Martineau and Mougeon2003, Poplack and St-Amand Reference Poplack, St-Amand, Martineau, Mougeon, Nadasdi and Tremblay2009), the use of ne is still somewhat productive in Continental French varieties. Thus, following Roberts (Reference Roberts2012, Reference Roberts2014), we further distinguished among negative contexts those in which the negative particle was used (20) and those in which it was omitted (21).
(20) Or vous n'aurez pas la majorité pour changer l'fonctionnement. (Fabienne A., 204)
‘Yet, you won't have the majority [needed] to change the way it works.’
(21) Vous aurez peut-être pas l'temps que j'vous montre les photos. (Françoise D., 189)
‘Maybe you won't have time for me to show you the pictures.’
4.2.3 Adverbial specification
The role of temporal adverbials on constraining FTR has shown mixed results across varieties of French. For example, it played no role in Montreal French (Blondeau Reference Blondeau, Papen and Chevalier2006), Ontario French (Grimm and Nadasdi Reference Grimm and Nadasdi2011), and in PEI and NL Acadian French (King and Nadasdi Reference King and Nadasdi2003), but it did play a role in Ottawa-Hull French (Poplack and Turpin Reference Poplack and Turpin1999) and in Baie Sainte-Marie Acadian French (Comeau Reference Comeau2015). In the two varieties where it does constrain variant choice, the PF is favoured in contexts without temporal adverbials, as in (22).Footnote 10 Therefore we coded each token based on whether there is no temporal adverbial (22), a specific temporal adverbial (23), or a non-specific temporal adverbial (24).
(22) D'ailleurs […] il t'en parlera aussi d’Chutt le Hutteux. (Jean-Marc A., 145)
‘In fact he'll tell you about Chutt le Hutteux too.’
(23) Là cette année c'est les Picards qui vont parler en breton. (Michel G., 693)
‘This year, it's the Picard [actors] who are going to speak Breton.’
(24) J'te donnerai ses coordonnées tout à l'heure. (Joël T., 651)
‘I'll give you his contact information later.’
If Vimeu French patterns like other varieties in which the adverbial constraint is operative, we might expect that the absence of temporal adverbials to favour the PF.
4.2.4 Subject type or vouvoiement
We also wanted to consider whether the IF had taken on a formal stylistic role in the variety under investigation. Poplack and Turpin (Reference Poplack and Turpin1999) consider whether the IF co-occurs with other markers of formal style, such as the use of the 2s pronoun vous ‘you’; they find that the IF is indeed favoured with vouvoiement, which suggests its role as a marker of formal speech. The purported stylistic role of the IF was further confirmed by Wagner and Sankoff (Reference Wagner and Sankoff2011), who found that speakers with a higher socioeconomic status favour the IF. Here, we replicate Poplack and Turpin's (Reference Poplack and Turpin1999) approach by coding each token for presence of formal vous (25) versus other subjects (26).
(25) Ah ben vous allez voir le lin fleuri. (Denis F., 852)
‘Well you're going to see the flax in bloom.’
(26) J'vais voir un petit peu si ça colle au niveau d'mes horaires. (Sophie F., 557)
‘I'm going to see if it fits into my schedule.’
If the IF does have a formal stylistic role, we would expect the IF to co-occur with a marker of formality, such as the 2s vous.
5. Results
After excluding tokens that did not fit within the variable context, we were left with 267 tokens of the FTR variable. The distribution of the variants is shown in Table 3: the PF makes up more than 60 per cent of all tokens in our Vimeu French corpus, a rate both comparable to Roberts’ (Reference Roberts2012) 58.8 per cent in 1980s Continental French and identical to Comeau's (Reference Comeau2015) 62 per cent in 1990s Nova Scotia Acadian French.
Overall distribution of FTR variants in Vimeu French
However, we want to compare the underlying constraints operating in Vimeu French to the ones reported for other French varieties, including the Continental French data examined by Roberts (Reference Roberts2012). Therefore, we submitted our data to multivariate analysis using GoldVarb X (Sankoff, Tagliamonte, and Smith Reference Sankoff, Tagliamonte and Smith2005).
The results of this multivariate analysis are shown in Tables 4 and 5. The results of the effect of social factors show that education level, not bilingualism, sex or age, condition variant choice in that speakers with no baccalauréat favour the PF while speakers with a baccalauréat or a university degree favour the IF.Footnote 11
Social factors affecting the periphrastic future in Vimeu French
Input: .63; Log likelihood = −171.533
Linguistic factors affecting the periphrastic future in Vimeu French
Input: 0.64; Log likelihood = −172.711
The fact that no other social factors were retained as significant in the model suggests that the FTR variable is stable in Vimeu French and that contact with Picard has no discernible effect based on our measure of influence due to contact. The fact that we find a linear correlation between education level and use of the IF suggests that the IF may have a stylistic value in this variety of French, similar to what has been reported in other varieties of French.
As for the effect of linguistic factors on variant choice (Table 5), the only factor group selected as significant was temporal distance, a result similar to both Acadian and Martinique varieties. While we initially coded for a fine breakdown of temporal distance (see section 4.2.1), we ultimately had to collapse factors within this group into a binary (i.e., proximate vs. distal) factor group, as in Poplack and Turpin Reference Poplack and Turpin1999 and Roberts Reference Roberts2014, due to the distribution of the data. In Vimeu French, the PF is favoured in proximal contexts, while the IF is favoured in distal contexts, thus confirming a temporal distance effect. The effect of polarity, although in the expected direction in terms of percentages – negative contexts disfavouring the PF – was not selected as statistically significant. Negative sentential polarity was collapsed into a single factor due to the categorical use of the IF in the presence of negative ne (N = 5).Footnote 12 The other linguistic factor, adverbial specification, exerted no statistically significant effect on variant choice, in keeping with some studies of FTR in French (King and Nadasdi Reference King and Nadasdi2003, Grimm and Nadasdi Reference Grimm and Nadasdi2011).
As we were unable to include two potential socio-stylistic factors – negative ne and vouvoiement – into the multivariate analysis, we opted to compare data from speakers without the baccalauréat, which favour the PF, with data from speakers with a higher level of education. This allowed us to conduct statistical tests to measure the potential effect of these two socio-stylistic factors and further investigate the role of sentential polarity.
Table 6 shows that the use of one FTR variant over another is not correlated with the use of formal 2s vous in Vimeu French, but the categorical co-occurrence of the IF and negative ne suggests that this variant may carry a mild stylistic value for the group with a higher level of education. As shown in example (20), repeated in (27), the use of several formal variants – the connector or ‘yet, but’ (instead of mais), along with vouvoiement and negative ne – may offer a favouring environment for the IF.
(27) Or vous n’aurez pas la majorité pour changer l'fonctionnement.
‘Yet, you [2s] won't have the majority [needed] to change the way it works.’
Socio-stylistic factors and FTR in Vimeu French
In short, the effect of sentential polarity remains insignificant in the no baccalauréat group, who produce no negative token with the formal ne particle, and even in the groups with a higher level of education, polarity emerges as marginally significant (p = 0.039) only when tokens with negative ne (N = 5) are considered.Footnote 13
6. Temporal distance in Vimeu and beyond
Now that sentential polarity has been shown to have no effect on FTR in Vimeu French, let us focus on the results for temporal distance, the only statistically significant linguistic constraint. Beyond the temporal distance effect uncovered by the distinction between proximal and distal contexts, the original finer breakdown reveals a finding that is unfortunately masked by the broader binary coding necessary for the multivariate analysis: the high frequency of the periphrastic future in the most immediate contexts. In Table 7, we present the rates of the periphrastic future based on our initial fine-grained breakdown of temporal distance. These results show that the PF is used at an extremely high rate (93.3%) with events anticipated to occur within a minute, which suggests that the PF is marking imminence, and appears to confirm early grammarians’ descriptions: the periphrastic future does mark le futur proche.Footnote 14
Temporal distance and the periphrastic future in Vimeu French
The fact that the PF occurs so frequently in imminent contexts in Vimeu French may be related to a feature of the French language more generally, since the same pattern was found in the Nova Scotia variety of Acadian French. In the other varieties which display a strong temporal distance effect, it is impossible to establish whether or not the PF is as highly favoured in imminent contexts (i.e., within the minute) due to the fact that the authors did not adopt as fine a temporal breakdown.
Given that both studies which distinguished imminent from other proximal contexts found such a strong correlation with aller ‘to go’ periphrasis, we extend the comparison across varieties where temporal distance was reported to be the most significant factor group constraining FTR variant choice. In addition to our Vimeu results, we include the data for the Acadian varieties spoken in Baie Sainte-Marie (Comeau Reference Comeau2015) and in Prince Edward Island and Newfoundland (King and Nadasdi Reference King and Nadasdi2003), as well as French data from Martinique (Roberts Reference Roberts2014). Table 8 shows the token counts for both PF and IF in each of these varieties broken down by the temporal breaks.
Temporal distance and the expression of the future across French varieties
Chi-square tests were performed on all four sets of results to determine which temporal breaks were statistically significant (indicated by bold lines).Footnote 18 Once we compare Vimeu with Baie Sainte-Marie, we note that both varieties share a statistically significant break between the within the minute contexts and the following temporal periods.Footnote 19 Once we extend the comparison to the communities that did not distinguish the within-the-minute contexts, we note that the statistically significant breaks occur much later. In the PEI and NL varieties, the PF is marking a more general proximity while in Martinique it is the IF which is favoured in distal contexts. The fact that both Baie Sainte-Marie and Vimeu French strongly favour the PF in the most immediate context suggests that this variant marks imminence rather than general proximity in at least some contemporary varieties of French.
How might we account for these differences between the varieties that show a temporal distance effect? We could entertain the possibility that an examination of the within-the-minute tokens both in NL and PEI and in Martinique might have revealed a cohesive pattern (i.e., an imminence effect) across all four varieties, but clear differences across varieties would still remain with respect to the later temporal periods (e.g., longer than a year in Martinique, unlike other varieties). An initial imminence function of the periphrastic construction is argued for in the literature (Flydal Reference Flydal1943). According to Fleischman (Reference Fleischman1982: 84), there is a “conceptual bridge linking imminence and futurity” diachronically, and this bridge is argued to be related to the notion of proximity (i.e., proximal future or le futur proche). As the use of the PF as a future marker increases, its association with imminence or proximity may weaken and eventually fade out. We argue that differences found across varieties may well be indicative of different stages of grammaticalization of the PF. For instance, the PF has fully grammaticalized into a general futurity marker in some varieties (i.e., Laurentian varieties, which show little to no temporal distance effects), appears to be in an intermediate stage in others, clearly functioning as a marker of general proximity (i.e., PEI and NL varieties), and in a third group, retains functions associated with the early stages of grammaticalization, marking imminence (i.e. Vimeu French and Baie Sainte-Marie). Although a detailed discussion of grammaticalization with respect to FTR is beyond the scope of this article, the parallels between the conservative nature of Baie Sainte-Marie French (Comeau Reference Comeau2015) and the recent emergence and apparent rise of the PF in Vimeu Picard (Auger and Villeneuve Reference Auger and Villeneuve2015) may, by extension, suggest that Vimeu French is, too, in the early stages of grammaticalization of the PF.Footnote 20 Further variationist research on FTR in French and in Picard will hopefully shed more light on this question.
7 Conclusion
Our analysis of FTR in Vimeu French yielded somewhat unexpected results. Our first two research questions focused on the comparison between our Vimeu data and Roberts’ previous variationist study of FTR in Continental French. The distribution of variants and the effect of educational level found in Vimeu French, with the PF used around 60 per cent and preferred by speakers with a lower level of education, are strikingly similar to those obtained in Roberts’ (Reference Roberts2012) analysis of Continental French from the 1980s. Contrary to our hypothesis about the potential effect of bilingualism and contact with Picard at the community level, Vimeu French does not display lower rates of the PF than other Continental French varieties. As for the effect of negative ne and of vouvoiement, our limited results appear to suggest that if the IF variant carries any formal value, this stylistic effect was found only in speakers with higher levels of education.
On the one hand, we anticipated that contact with a language in which the IF is the preferred variant (i.e., Picard) may have had an effect on the FTR system in French. This hypothesis was not borne out. The absence of a bilingualism effect, along with a distribution of IF and PF almost identical to that described by Roberts (Reference Roberts2012), may be due to a significant rise in PF in Picard since the 1960s, as Auger and Villeneuve (Reference Auger and Villeneuve2015)’s results appear to suggest. The fact that the FTR variable may not operate above the level of consciousness in Picard and that the traditional Picard variant (i.e., the IF) is the prestige variant in French may have favoured similarities in variant rates between the two languages over the years.Footnote 21
With regard to the linguistic conditioning, we anticipated that much like Roberts’ (Reference Roberts2012) results for Continental French, sentential polarity would be found to constrain variant choice in Vimeu French. This hypothesis was also not verified. The fact that our results differ from Roberts’ may be an artefact of the data (i.e., different types of sociolinguistic corpora reflecting a range of registers) or they may be evidence of grammatical variation across Continental French regional varieties. At this point, we cannot offer a definitive account of this difference.
Aside from the lack of a polarity effect, another major finding is that Vimeu French patterns alongside a number of disparate varieties of French (i.e., the Acadian varieties spoken in Baie Sainte-Marie, NS and in PEI and NL as well as Martinique French). It is fairly clear that no sociohistorical or genetic filiation between the communities can account for similar patterning (aside from the similar finding within Acadian varieties).Footnote 22 We must therefore conclude that the temporal distance effect that emerges in at least some varieties of French may be (and have been) more widespread than previously thought in the French language as a whole, suggesting that early grammarians and remarqueurs may have identified an actual constraint on the FTR variable present in the language spoken in Continental France at the time of their descriptions.
While we initially set out to investigate the potential effects of Picard on French by filling a gap in the literature on FTR in Continental France, our results add to the broader discussion of this variable. The linguistic and social constraints operating on the variable in Vimeu thus contribute some pieces of the puzzle to understanding the mechanism behind FTR in French. While our study does offer some answers, it also raises new questions pertaining to the variable: what is the nature of FTR in contemporary Picard? How have the distribution of variants and linguistic conditioning changed over the last century, a period of language shift from regional languages to French? Did Vimeu French and Vimeu Picard influence one another with regard to the FTR variable and to morphosyntax more generally? Is the imminence constraint more widespread than it appears and may it be obscured by differences in coding procedures? We hope that future work on FTR may help answer some of these questions.
