This thematic issue of the Canadian Journal of Linguistics explores the expression of future temporal reference (FTR), that is, the different ways of expressing that an event will occur after the moment of speech, in French. Despite numerous studies on this topic dating as far back as the 1980s (Deshaies and Laforge Reference Deshaies and Laforge1981, Emirkanian and Sankoff Reference Emirkanian, Sankoff, Lemieux and Cedergren1985, Poplack and Turpin Reference Poplack and Turpin1999, King and Nadasdi Reference King and Nadasdi2003, Wagner and Sankoff Reference Wagner and Sankoff2011, among others), a number of hotly debated points continue to incite discussion. To shed light on some of these issues, we assembled articles that deal with FTR from a range of perspectives.
In most contemporary varieties of spoken French, three main forms (or variants) are used to express that an event will take place in the future: the inflected future (or synthetic or morphological future, henceforth IF) as in (1); the periphrastic future (or analytic future, composed of the semi-auxiliary aller ‘to go’ followed by the infinitive, henceforth PF), as in (2); and the futurate present (or present-for-future, i.e., use of the present indicative morphological form with an expressly future temporal reference, henceforth P), as in (3).
(1) Je quitterai la semaine qui vient. ‘I will leave next week.’
(2) Je vais quitter la semaine qui vient. ‘I’m going to leave next week.’
(3) Je quitte la semaine qui vient. ‘I’m leaving next week.’
At first glance, FTR may seem a somewhat mundane feature of French, especially considering that, on the surface, the same three variants (i.e., the IF, PF, and P) are consistently found across varieties. However, once we examine the underlying system of linguistic and social constraints that govern the alternation among these three forms, we begin to observe divergence across varieties. This raises new questions, not only regarding the constraints operating on these forms, but also with respect to dialectal differences across varieties of French and the history of the French language more generally.
1 The development of FTR variants in French
While the three FTR variants persist in most spoken varieties of French today, each form has its unique history. The IF is derived from the Vulgar Latin periphrastic construction involving a lexical verb in its infinitival form (e.g., cantare ‘to sing’, dicere ‘to say’) followed by the auxiliary habeo, as shown in example (4), from Lucretius (circa 99–55 BCE).
(4) item in multis hoc rebus dicere habemus (Fleischman Reference Fleischman1982: 52)
‘similarly we have this to say about many things’
Its presence in Late Latin accounts for why it gave rise to a synthetic future form across Romance languages. According to Fleischman (Reference Fleischman1982: 70–71), the development of the infinitive + habeo construction into the Romance IF forms can be interpreted as a reanalysis of the verbal phrase into a single verbal form, as shown in (5).
(5) Latin cantare habeo (Fleischman Reference Fleischman1982: 72)
French chanter – ai
Spanish cantar – é
Italian canter – ò
In fact, the IF is attested in some of the earliest Old French sources, as shown in this example from Les Serments de Strasbourg from 842 CE (Gasté Reference Gasté1888: 12).
(6) Et ab Ludher nul plaid nunquam
prindrai qui meon vol cist meon fradre
Karle in damno sit
‘And I will never hold knowingly any counsel with Luther, which may be harmful to my brother Charles’ (our translation)
With regard to the PF, this relative newcomer – its appearance as a future marker dates to the Middle French period, from the fourteenth to the seventeenth century (Wilmet Reference Wilmet1970, Champion Reference Champion1978) – initially denoted spatial movement, but its co-occurrence with temporal adverbials allowed it to be situated in the future, as in (7).
(7) Je vais lui parler tout à l'heure. (Molière, Monsieur de Pourceaugnac, Act 2, Scene 1)
‘I'm going to talk to him shortly.’
According to Fleischman (Reference Fleischman1982: 84), the future meaning eventually spread from the adverbial to the verbal phrase: it grammaticalized as a future marker and then came to be used without a temporal adverbial, as in (8).
(8) C'est sûr qu'il va vous parler picard, lui. (Vimeu French; cf. Villeneuve Reference Villeneuve2011)
‘It's certain that he is going to talk to you in Picard.’
Today, spoken varieties of French make use of the PF as a marker of FTR, as do many Romance languages, such as Spanish (9) and Portuguese (10), despite the fact that this variant emerged in the respective Romance languages only after they had developed from Proto-Romance.
(9) ¿ Donde vamos a dar con un equipo así ? (Colombian Spanish; cf. Orozco Reference Orozco and Eddington2005: 63)
‘Where are we going to end up with a team like that?
(10) Ano que vem eu vou tirar a Cleide dessa escola
‘Next year I'm going to take Cleide out of this school.’
(Brazilian Portuguese; cf. Poplack and Malvar Reference Poplack and Malvar2007: 123)
As for the third main variant, the P, it predates Old French; it was possible to use a present tense verb with future temporal reference as early as in Classical Latin, as shown in example (11), from Cicero (Lebreton Reference Lebreton1901: 192–193).
(11) ita relinqui nihil praeter ignem (Cicero, De natura Deorum)
‘so nothing will remain but fire’
The presence of P in Latin explains why we find this variant used, at variable rates, today across Romance languages.
2 Sociolinguistic studies of FTR in French
2.1 The variable context
Through decades of work on FTR in French, the variable context has undergone substantial revisions since the early 1980s. In the first few studies (Deshaies and Laforge Reference Deshaies and Laforge1981, Emirkanian and Sankoff Reference Emirkanian, Sankoff, Lemieux and Cedergren1985, etc.), the variable was defined primarily based on form rather than function, so that it included tokens of the so-called ‘false futures’ (i.e., tokens of the variants which do not express a future eventuality). In her analysis of FTR in Montreal French, Zimmer (Reference Zimmer1994: 219) first raised this methodological issue: “si l'on veut examiner la variation entre le FS [futur simple] et le FP [futur périphrastique], on est obligé de ne retenir que les formes du FP ayant un aspect temporel, et d’écarter les formes ayant un aspect modal”.Footnote 1 Her proposal led subsequent studies (e.g., Poplack and Turpin Reference Poplack and Turpin1999, King and Nadasdi Reference King and Nadasdi2003, Blondeau Reference Blondeau2006) to exclude from the variable context all tokens with a non-temporal function, such as those denoting habitual aspect, as in (12).
(12) je vas m'acheter des fois le Devoir des fois le Soleil (Zimmer Reference Zimmer1994: 218)
‘I sometimes go to buy Le Devoir, sometimes Le Soleil.’
While the FTR variable in French involves variation between three forms, most studies focus on the alternation between only two of these forms, the PF and the IF. To our knowledge, only a handful of sociolinguistic studies retain the P in their analyses of FTR: Poplack and Turpin (Reference Poplack and Turpin1999), Poplack and Dion (Reference Poplack and Dion2009), Gudmestad et al. (Reference Gudmestad, Edmonds, Donaldson and Carmichael2015), and Grimm (Reference Grimm2015). The decision to exclude the P has been justified on the basis that “the [futurate] present is very much a marginal form in future contexts, accounting for less than 10% of tokens in studies that have reported on its frequency” (Wagner and Sankoff Reference Wagner and Sankoff2011: 278). Another reason cited is “its almost categorical co-occurrence with future adverbials” (Roberts Reference Roberts2012: 97). However, the choice to exclude the P may not be completely misguided. As Poplack and Dion (Reference Poplack and Dion2009: 572) point out, based on their diachronic analysis of Laurentian French, “the minority variant P [i.e., futurate present] remains unchanged, in terms of both rate and conditioning” hovering around 7–9 per cent in Laurentian French since the nineteenth century. Therefore, it appears, at least for some Laurentian French varieties, that if there is ongoing linguistic change in how French expresses FTR, it primarily affects the respective proportion of the IF and the PF.
2.2 Community-wide and individual language change in FTR
Aside from the methodological issues, some have questioned whether FTR is undergoing change. Despite centuries of variability between the three main forms, most studies of contemporary Laurentian varieties report high rates of the periphrastic variant, as high as 91% for some Ontario French speakers (Grimm Reference Grimm2015: 261). In contrast, Acadian varieties spoken in Canada's Atlantic Provinces, especially those in minority French contexts, report much lower rates of the PF, ranging from 41.4 per cent in Prince Edward Island to around 60 per cent in Newfoundland and Nova Scotia (King and Nadasdi Reference King and Nadasdi2003, Comeau Reference Comeau2015). As for non-Canadian varieties, rates of the PF in spontaneous speech range from around 60 per cent in Continental French (Roberts Reference Roberts2012, Villeneuve and Comeau, this issue) to 72.3 per cent in Martinique French (Roberts, this issue). These rates have led some to argue, as was proposed by some grammarians, that the PF is gradually ousting the IF.
Alongside the gradual rise of the PF, Wagner and Sankoff (Reference Wagner and Sankoff2011) argue for individual retrograde change (an age-grading effect) towards a greater use of the IF by older speakers and by those from higher socioeconomic groups. In fact, the use of the inflected form has been associated with speakers from higher socioeconomic classes and with more formal registers in several French communities. For instance, Roberts (Reference Roberts2012) found that all speakers of Continental French except those with a university degree favour the PF. The effect of socio-economic class on FTR across other varieties of French is far less established. While several contributors in this issue take socio-stylistic constraints into consideration, the main focus of this thematic issue is placed on linguistic constraints, two of which have occupied a sizeable place in the FTR literature: temporal distance and sentential polarity.
3 Language-internal constraints on FTR: A temporal distance-polarity divide?
Aside from the issue of language change, one longstanding debate about the underlying grammatical system operating on FTR is the nature of linguistic constraints governing the use of each variant. While a number of constraints have been considered in the extant literature, the results of sociolinguistic studies have shown major differences across varieties of French. Since the early 1980s, linguists working on Laurentian French have commented on the sentential polarity constraint, namely that negative contexts strongly favour the IF variant. First observed by Seutin (Reference Seutin1975) for the Iles-aux-Coudres Québec French variety, this constraint has since been confirmed by several studies of Laurentian French: in Québec City (Deshaies and Laforge Reference Deshaies and Laforge1981), Montréal (Emirkanian and Sankoff Reference Emirkanian, Sankoff, Lemieux and Cedergren1985, Blondeau Reference Blondeau2006, Wagner and Sankoff Reference Wagner and Sankoff2011), Ottawa-Hull (Poplack and Turpin Reference Poplack and Turpin1999, Poplack and Dion Reference Poplack and Dion2009), and Ontario (Grimm Reference Grimm2010, Grimm and Nadasdi Reference Grimm and Nadasdi2011, Grimm Reference Grimm2015). The discovery of the polarity constraint was somewhat of a surprise since, for centuries, it had not been commented upon by grammarians (cf. Poplack and Dion Reference Poplack and Dion2009: 574). The predictive strength of polarity challenged the relevance of the predictor of FTR variant choice most commonly advanced by grammarians: temporal distance, namely that events occurring in the near (or proximal) future favour the PF. However, despite the preponderance of the polarity constraint observed in Laurentian French, it soon became clear that this constraint was not shared across all varieties of contemporary French.
King and Nadasdi's (Reference King and Nadasdi2003) study of the Acadian French spoken in Prince Edward Island and Newfoundland provided a sharp contrast to the consistency of Laurentian results: the polarity constraint was clearly absent from these Acadian varieties, where temporal distance was the strongest predictor of variant choice. This discovery highlighted a Laurentian–Acadian split in Canadian French with respect to the underlying constraints conditioning the FTR variable. Comeau's recent (Reference Comeau2015) study of another Acadian French variety, that spoken in the Baie Sainte-Marie region of Nova Scotia, has further confirmed these results: while Laurentian varieties are governed by a constraint absent from grammarians’ commentaries, the Acadian system closely mirrors what has long been described in French grammars (i.e., the temporal distance constraint is operative).
Once we extend the comparison beyond Canada, the picture becomes even blurrier. Roberts’ (Reference Roberts2012) study of Continental French reported that the polarity constraint is operative in some of the varieties of French spoken in France, thus suggesting a link between the Laurentian varieties and Continental French. In line with this point of debate, a number of articles in this issue provide additional evidence bearing on this seeming dichotomy. Roberts’ analysis of FTR in Martinique French adds a Caribbean variety to the picture while Villeneuve and Comeau's investigation of the French spoken in the Vimeu region contributes recent data from Northern France; both contributions uncover a temporal distance effect, thus further supporting grammarians’ claims. Blondeau and Labeau's article, which deals with a different genre (i.e., prepared oral speech rather than sociolinguistic interviews), also fails to uncover a polarity effect in their Continental and Québec data, finding a weak temporal distance effect in the latter but not in the former.
Clearly, the contrast in how varieties pattern with respect to these two constraints raises several questions about FTR in French: Why are so many varieties of French aligning themselves with either the polarity or temporal distance constraints? What can account for the emergence of the polarity constraint? Is the temporal distance constraint a conservative feature of French? How do non-Canadian varieties of French behave with respect to this constraint dichotomy? While this thematic issue does not offer definitive answers to these questions, the empirical studies of FTR in less commonly studied varieties (Martinique, Picardie), genres (weather forecasts) or speakers (learners of French as a second language) presented here contribute to our general understanding of the mechanisms behind FTR across French more generally.
Aside from the temporal distance–polarity dichotomy, other weaker constraints have been proposed in the literature, often with conflicting results. For instance, the presence of temporal adverbials (or lexical temporal indicators) has been argued to condition the choice between FTR variants. Poplack and Turpin (Reference Poplack and Turpin1999) report that for Ottawa-Hull French, the presence of specific temporal adverbials (e.g., demain ‘tomorrow’, ce soir ‘tonight’) favours the use of P while non-specific adverbials (e.g., plus tard ‘later’, un jour ‘one day’) favour the IF; a lack of adverbial specification is associated with use of the third variant, the PF. However, the effect of this constraint is not constant even within Canadian French varieties: it is operative in some Laurentian (Poplack and Dion Reference Poplack and Dion2009) and Acadian varieties (Comeau Reference Comeau2015), but inoperative in others (e.g., Blondeau Reference Blondeau2006, King and Nadasdi Reference King and Nadasdi2003). There also appears to be a lack of effect in Continental French (Roberts Reference Roberts2012, Gudmestad and Edmonds, this issue, Villeneuve and Comeau, this issue). Other constraints examined in the literature include the speaker's certainty regarding the future event (King and Nadasdi Reference King and Nadasdi2003), whether the future event is contingent on another event (Wagner and Sankoff Reference Wagner and Sankoff2011), and the grammatical person and number of the subject (Grimm Reference Grimm2010, Roberts Reference Roberts2012, Wagner and Sankoff Reference Wagner and Sankoff2011, inter alia). While studies reveal different results with these other constraints, their effect is consistently weaker than that of either sentential polarity or temporal distance.
4 The contributions to this thematic issue
The articles in this thematic issue address a number of questions about FTR in French. They provide current empirical studies in various settings, thus contributing to our knowledge of how this variable operates in French. As the reader will note, there are already several studies of FTR in Laurentian and Acadian varieties based on traditional sociolinguistic corpora; the main focus of contributions to this issue moves beyond strictly Canadian varieties to provide analyses of new communities, genres, and speakers. For instance, Blondeau and Labeau provide a comparative analysis of Continental French and Quebec French within the context of televised weather forecasts, a genre particularly suitable for the FTR variable. While their article involves varieties which have already been previously investigated (i.e. Continental and Laurentian French), it sheds light on how FTR operates in prepared oral speech, a genre distinct from the typical sociolinguistic interview where the goal is to elicit the vernacular. Their findings indicate that in both data sets (France and Quebec) the IF is the most frequent variant, a striking contrast to previous studies conducted on French vernaculars. Furthermore, the effect of polarity, clearly documented in Laurentian French, is notably absent in the data, including the Quebec weather reports, thus suggesting that linguistic constraints vary according to linguistic genre.
This issue also presents analyses of FTR in previously unexplored French communities. For the variety of French spoken in Martinique, Roberts' paper advances the analysis of the FTR variable on a number of points. Firstly, it sheds light on the issue raised in section 3 regarding the role of linguistic constraints and the apparent divide among varieties of French. In addition, his analysis makes use of the current tools made available in the variationist paradigm (the Rbrul software package) in order to take into account the potential effect of variation at the individual speaker and lexical verb levels (i.e., by taking into account random effects in the mixed-effects model). With regard to the main constraints operating on the variable, Roberts finds that once the random effects of speaker and lexical verb are taken into account, only one linguistic constraint influences variant choice: temporal distance. While this finding is similar to previous work on Acadian French (King and Nadasdi Reference King and Nadasdi2003, Comeau Reference Comeau2015), the particular effect of this constraint is different: distal contexts favour the IF while the PF is the default future marker (and not associated with proximal contexts). Villeneuve and Comeau's article also extends the study of FTR from commonly studied varieties to a rural variety of Continental French in contact: that of Vimeu, an area of northern France where Picard–French bilingualism persists. They note a distribution of variants and an effect of education similar to those documented by Roberts (this issue) for Martinique, whereby speakers with a baccalauréat or a university degree favour the inflected form. These authors also find that temporal distance is the only linguistic predictor of FTR variant choice, but unlike in Martinique, the PF is strongly correlated with imminent contexts. They argue that this correlation may reflect an early stage of the grammaticalization process of the PF into a general marker of futurity, as it appears to also be the case in some Acadian varieties in Canada (Comeau Reference Comeau2015). A correlation between the IF and other markers of formality (e.g. presence of negative ne) emerges from their analysis, especially when considering the role of education. Finally, although the Picard substrate appears to have lacked the periphrastic form until recently (Auger and Villeneuve Reference Auger and Villeneuve2015), Picard–French bilinguals behave like their French monolingual counterparts with respect to FTR.
Bilingualism also relates to Gudmestad and Edmonds' contribution, which brings a second-language acquisition (SLA) perspective by investigating FTR in learners of French as a second language (L2) through the use of a written-contextualized task rather than the typical sociolinguistic interview. While this elicitation task does not test for the effect of sentential polarity, it considers other linguistic factors such as temporal distance, (un)certainty markers and temporal adverbials (or ‘lexical temporal indicator’), as well as two social factors: type of learning context (home or abroad) and proficiency level (low or high). Native speakers from Southwestern France serve as a control group, offering a glimpse into potential regional variation in Continental French. They find that the control group uses the PF most frequently but that no consistent distribution emerges from the L2 groups. The presence of a temporal adverbial (or lexical temporal indicator) favours the P for high-proficiency study abroad learners, as is documented in some native speakers of Laurentian varieties (Poplack and Turpin Reference Poplack and Turpin1999, Nadasdi, Mougeon, and Rehner Reference Nadasdi, Mougeon and Rehner2003), but this effect is absent from the native speaker control group. All groups were subject to a temporal proximity effect, with proximal contexts favouring the PF; high proficiency also showed a stronger preference for P in proximal contexts. As for (un)certainty, they argue that differences between the low- and high-proficiency groups indicate a developmental progression. Low proficiency L2 learners display similar distributions and linguistic conditioning whether they studied at home or abroad, and while there are some differences between the two high proficiency groups, none of the L2 groups' behaviour is identical to that of native speakers.
While we don't presume to resolve all outstanding questions pertaining to FTR in French within this thematic issue, we hope that these contributions further our knowledge of how it operates in French. In addition to being important for understanding one dimension of the French temporal system, the study of FTR allows us to explore questions relating to dialectal differences, the grammaticalization process, social and stylistic stratification, and, ultimately, language change.
