2 PAST TEMPORAL REFERENCE IN THE HISTORY OF FRENCH

The French imperfect descends from the Classical Latin imperfect in –ebam (>eie>-oie>ais, etc.) while the French simple past descends from the Classical Latin preterite. The periphrastic present perfect was a Low Latin innovation which has been argued to be in competition with the simple past since the Old French period (Fournier, Reference Fournier1998: 395). There is, however, some debate as to when exactly it became a real competitor of the preterite. For instance, Judge and Healey (Reference Judge and Healey1983, cited by Labeau, Reference Labeau2004) allow that only “occasional examples” of true variation existed during the Old French period. In general, the passé simple and the imparfait are said to differ in terms of whether or not the event or state entirely precedes the moment of speech (PS) and whether or not the event itself has internal temporal structure (I). As for the passé simple and the passé composé, the distinction is said to turn on whether the event is viewed as a relatively distant event completed in the past (PS) or as having some present relevance (PC).

The passé simple and the passé composé have long been suggested to differ in terms of the degree of recentness required for their use. For example, the famous ‘24 hour’ rule codified by Estienne (Reference Estienne and Smith1569) seeks to quantify this degree of remoteness by prescribing that events taking place within 24 hours of the moment of speaking be rendered in the passé composé and events taking place more than 24 hours prior to the moment of speech be rendered in the passé simple. Other early grammarians and remarqueurs (e.g. Maupas, Reference Maupas1618; Chifflet, Reference Chifflet1659) place considerable emphasis on the co-occurrence of certain adverbial expressions, in particular hier, with the passé simple. Such expressions situate the event at particular points in time, and, it appears, go hand in hand with robust use of this tense (a correlation lost in the 18th century, according to Caron and Liu's 1999 study of personal letters from that period).

In French, as in many other Romance languages (Squartini and Bertinetto, Reference Squartini, Bertinetto and Dahl2000), the decline of the passé simple has coincided with a relaxation in the degree of recentness required for use of the passé composé. Ultimately, the passé composé won out in the spoken language: the passé simple was in decline by the second half of 17th century, highly restricted by the 18th century, and near extinction by the 19th century (Fournier, Reference Fournier1998: 399). For instance, as Meillet (Reference Meillet1948: 150) notes it is almost entirely absent from the French data for the Atlas linguistique de la France (Gilliéron and Edmont, Reference Gilliéron and Edmont1901–1910), which records the usage of older rural male speakers from the late 19th/early 20th century. Meillet also remarks that in areas for which the passé simple is still attested, in the ALF the passé composé is also found and indeed dominates.Footnote ²

In a similar vein, Carruthers' (Reference Carruthers2006) examination of tense marking in narrative clauses found in the Aubrac collection recorded in the 1960s by Marie-Louise Tenèze found only one of 31 folktales involved use of the passé simple. This finding is particularly striking given that Tenèze's informants were French/Occitan bilinguals and Occitan itself retains a preterite. However, Chaurand (Reference Chaurand, Antoine and Cerquiglini2000: 646) reports that the passé simple was still common in some western and southern varieties in the second half of the 20th century and Jones (Reference Jones2001: 22–23) notes that it is still in use in spoken Jersey French.Footnote ³ Note, though, that while these latter sources attest to the persistence of such usage, they do not place it in the larger context of the varieties' system of past temporal reference.

Decline of the passé simple is typically seen as part of a larger shift from synthetic to analytic variants in the morphosyntax of French (Blanche-Benveniste and Jeanjean, Reference Blanche-Benveniste and Jeanjean1986; Lodge, Reference Lodge2004; King, Martineau and Mougeon, Reference King, Martineau and Mougeon2011). This shift would also include the rise of the periphrastic (versus inflected) future and of first person plural on + verb (versus je + verb + -ons and nous + verb + -ons).

Based on his analysis of written usage, the twentieth-century linguist Émile Benveniste (Reference Benveniste1959/1966; see as well Millon, Reference Millon1936; Imbs, Reference Imbs1960, and Weinrich, Reference Weinrich1973) has famously argued that the essential difference between use of the passé simple and the passé composé involves the distinction between histoire and discours:

Since the passé simple presents an event as an unanalyzable whole, complete at a particular moment in time, it lends itself to the narration of a series of events. Note, though, that Benveniste's distinction between histoire and discours is rather different from more current use of these terms, at least in sociolinguistics and discourse analysis, as mention of the absence of speaker intervention in histoire in the above quote indicates. Thus, for Benveniste a text in which the speaker him or herself intervenes, for example, by use of a first or second person singular pronoun, would not strictly count as histoire, even though it would involve the temporal sequencing of clauses mirroring the original order of events. In addition, since the passé simple is today largely confined to formal, written French and even in literary texts and journalistic writing is undergoing reduction in usage (Herzog, Reference Herzog1981; Labeau, Reference Labeau2004, Reference Labeau2009), there are clearly exceptions to the histoire-discours distinction, as Carruthers (Reference Carruthers2006) notes.

As for the imparfait, it has typically been associated with habituality, simultaneity and durativity, i.e. it introduces a condition on the temporal referent of discourse. The decline of the passé simple may have also involved a repartitioning of the functions of the imparfait as well as of the passé composé. For instance, Leroux's (Reference Leroux2005) analysis of two generations of Quebec French speakers' spoken language use (one born between 1893 and 1918 and the other born between 1957 and 1965, who did not use the passé simple) found change in the conditioning of the passé composé/imparfait alternation from one generation to another.

3 THE BAIE SAINTE-MARIE SPEECH COMMUNITY

Acadian French refers to (marginalised) varieties of French spoken principally in the Canadian provinces of New Brunswick, Newfoundland and Labrador, Nova Scotia, and Prince Edward Island. It differs from its more well-known neighbour, Quebec French, in part due to the different European origins of the colonists, with the majority of Acadian settlers coming from the provinces of the centre-ouest of France in the 17th century, whereas Quebec colonists were more diversified, with substantial numbers of settlers from north of the Loire Valley. Another reason for differences between the two varieties, perhaps the most important one, is the relative isolation of Acadians from contact with other francophones and from the normative influences of a French education system, isolation which lasted into the 20th century for some varieties and which continues to the present day for others (Flikeid, Reference Flikeid, Mougeon and Beniak1994). For example, Acadian varieties preserve, to varying degrees, archaic first person plural pronominal use (e.g. je parlons ‘we speak’) and third person plural verbal morphology (e.g. ils parlont ‘they speak’), both of which were in decline in colloquial speech in France by the end of the 18th century (King, Martineau and Mougeon, Reference King, Martineau and Mougeon2011). In addition, while it is in decline elsewhere (Poplack and Dion, Reference Poplack and Dion2009), the inflected future remains in robust use in conservative Acadian varieties (King and Nadasdi, Reference King and Nadasdi2003).

Not surprisingly, then, the passé simple survived in varieties of Acadian French well into the 20th century. Haden (Reference Haden1948) recorded its usage in New Brunswick varieties in the 1940s, although Flikeid and Péronnet (Reference Flikeid and Péronnet1989) note its almost complete disappearance from these same varieties by the 1980s. Flikeid and Péronnet did, however, find it to still be in frequent usage in all five Acadian regions of the province of Nova Scotia in the 1980s. The data in 1–3 illustrate use of the passé simple, passé composé, and imparfait, in a sociolinguistic interview corpus recorded in 1990 in the village of Grosses Coques, a village of approximately 350 inhabitants found in the relatively isolated Baie Sainte-Marie region.Footnote ⁴

1. 1 Je lui parlis un élan, là.

   1-pers to-him speak-ps a while there

   ‘I spoke to him for a while.’ (Hector, GC-25)

2. 2 J'ai point parlé trop longtemps.

   1-pers have neg spoke too long

   ‘I didn't speak too long.’ (Hector, GC-25)

3. 3 Elle se parlait toute seule, steady.

   she refl speak-imp all alone all-the-time

   ‘She used to talk to herself, all the time.’ (Hector, GC-25)

As the example in 1 illustrates, the <a> and <i> conjugations of the passé simple have fallen together (the <u> conjugation remains intact, e.g. il but “he drank”), a simplification which has its origins in European French. Fouché (Reference Fouché1967: 259–261) gives the origins of <a> - <i> variation in the 13th and 14th centuries while Dauzat (Reference Dauzat1927) suggests that the reduction to two conjugations dates from the 16th century in colloquial French in the west, in Picardie, and in l'Île-de-France. Lodge (Reference Lodge2004: 134) notes that most of the 16th-century grammarians discuss this variation, expressing a preference for the <a> variant. The reduction to two conjugations is represented in 17th-century colloquial French texts such as Les Mazarinades. Since Acadian French has its origins in the centre-ouest, with immigration to the New World taking place in the 17th century, a two-conjugation passé simple in this variety is not surprising.

Baie Sainte-Marie (also known as the Municipality of Clare) is one of five predominantly Acadian regions in Nova Scotia indicated in grey on Figure 2. While speakers of Acadian French are in a minority position within the province as a whole (those reporting French as a mother tongue made up 4% of the province's population in the 1996 Canadian Census), Baie Sainte-Marie is itself predominantly French-speaking, with 73% of inhabitants reporting French as a mother tongue (Statistics Canada, 1996). Grosses Coques is located near the northern border of the region.

Figure 1.

The Atlantic Provinces

Figure 2.

Acadian Regions of Nova Scotia

Thirty-one Grosses Coques consultants were interviewed in 1990 by community-resident interviewers with a goal of obtaining naturalistic speech. Further, the interview protocol had as a primary focus the recording of narratives, which led to the recounting of many narratives of personal experience and community narratives. As would be expected, the interviews also contain a substantial amount of conversational data as well, focussing on consultants’ life histories and day-to-day life in the community. A total of 30 interviews were recorded in both individual and in group settings, with a total of approximately 260,000 words. Data for 14 consultants, representative of younger and older age groups and of both sexes, were retained for the purposes of this study. Table 1 gives a speaker breakdown.

Table 1.

The Grosses Coques Consultants

4 THE VARIABLE CONTEXT

While in the early years of variationist research a major source of debate was whether or not variants of morphosyntactic variables needed to be semantically equivalent (cf. Labov, Reference Labov1978; Lavandera, Reference Lavandera1978), this criterion has been replaced in recent years by the broader criterion that variants should share the same grammatical function(s). Schwenter and Torres Cacoullos (Reference Schwenter and Torres Cacoullos2008: 10–11) make a strong argument for the latter criterion. In language change, they argue, particular variants may retain earlier functions as they grammaticalize:

With this in mind, here we initially define one variable involved in the expression of past temporal reference and go on to refine the analysis.

The variable context involves use of verbal forms in either the imparfait, passé simple, or passé composé which have past temporal reference. We leave aside other compound tenses, such as the plus-que-parfait and the passé surcomposé, and largely ignore use of the historic present as the latter occurs infrequently in the Acadian data. The first step in data extraction from the Grosses Coques corpus involved distinguishing instances of the passé simple from the completely homophonous imparfait du subjonctif. This overlap had been noted by Flikeid and Péronnet (Reference Flikeid and Péronnet1989) for Acadian varieties; it was also found in European vernaculars (Dauzat, Reference Dauzat1947). For the purposes of their 1989 study, Flikeid and Péronnet chose to provide proportions of the passé simple/imparait du subjonctif combined in each of the Acadian regions of Nova Scotia because of the perceived difficulty in distinguishing the two. Thus theirs was a measure of preservation of inflectional verbal morphology rather than use of a particular form of the past tense. For the present study, however, we first determined the set of subjunctive-selecting contexts for this variety on the basis of usage with other tense/aspect combinations (see in this regard Comeau, Reference Comeau2011). For instance, there is categorical use of the subjunctive rather than the indicative mood following the main clause verb falloir ‘to be necessary’. Thus tokens such as Il fallait qu'il venit (‘It was necessary that he came.’) were excluded from the present study. In the case of variable subjunctive-indicative usage (as with the verb ne pas croire ‘to not think’), these tokens, too, were excluded since we could not be certain that they actually involved use of the passé simple.

In addition, tokens were excluded for the following reasons:

• – if the imparfait verb morphology was used in a conditional construction (e.g. si j'allais là . . . ‘if I went there’) since this does not constitute a token with actual past temporal reference;

• – if a verb did not display variable tense/aspect marking (être ‘to be’, il y (en) avoir ‘there was/were’, and savoir ‘to know’ almost always occurred in the imparfait);

• – if singular forms were homophonous with another tense (e.g. dire ‘to say’, finir ‘to finish’, grandir ‘to grow up/tall’, cuire ‘to cook’, assir ‘to sit’, etc.). For these verbs, the passé simple is usually homophonous with the present tense (e.g. il dit might be either a (historic) present or a preterite);

• – if a verb was conjugated in other past tense forms (e.g. plus-que-parfait, passé surcomposé);

• – if the token occurred in the reported speech of others;

• – if the token involved a false start or hesitation.

Once all such data were excluded, a total of 2,998 tokens were retained for quantitative analysis.

Since the verbal morphology of conservative varieties of Acadian French differs considerably from that of Standard French as well as that of better known French vernaculars, we present in Table 2 paradigms for the three past temporal reference variants under study.

Table 2.

Acadian French Verbal Paradigms (Conservative Varieties)

In the singular, verb forms are usually homophonous, although here we follow orthographic convention in using different spellings. In the plural, linguistic features typically associated with Acadian French include the use of pronoun je in combination with the –ons inflectional ending to mark first person plural, as shown in the imparfait and passé composé here (indeed je + verb + –ons is found across all tense/aspect/mood combinations with the exception of the passé simple and the imparfait du subjonctif). According to Zink (Reference Zink1997: 335–6), je + verb + –ons usage is already attested in Middle French but we know that it fell out of favour by the 16th century with the higher classes since several grammarians condemn its use (including Meigret, Reference Meigret1550; Ramus, Reference Ramus1572; and Estienne, Reference Évang1578, all cited by Lodge, Reference Lodge2004: 135). Likewise, use of third person ils in combination with the –ont inflectional ending is attested in the late medieval period but was more geographically limited than je + verb + –ons, found only in the west/south-west (Pope, Reference Pope1952: 89, also cited by Lodge, Reference Lodge2004: 89). From the 18th century on, these vernacular verb forms survive only in isolated rural varieties of French.

However, the –ons/–ont inflectional endingsFootnote ⁷ are missing from both the passé simple and the imparfait du subjonctif in Acadian varieties. There is one verb form in the singular [i] for the <i> conjugation verbs and [y] for the <u> conjugation verbs. In the plural, the corresponding forms are [ir] and [yr], spelt –irent and –urent respectively, which coincide with the Standard French (and Acadian) forms of the third person plural of the passé simple. In conservative varieties of Acadian French (including this Baie Sainte-Marie variety as well as other Nova Scotia varieties reported on by Flikeid and Péronnet) this regularisation of plural forms is not found in the présent du subjonctif nor in any other tense/aspect/mood. While these particular plural forms have not been documented in the well-known sources for 17th-century European French (e.g. they are not found in the speech of Molière's peasant characters or in Les Mazarinades), one does find a number of patterns of regularization in the plural for the passé simple in the Atlas Linguistique de la France and in 20th-century dialect descriptions for several varieties.Footnote ⁸ Most relevant for the present study is Svenson's (Reference Svenson1959) description of Marais-Vendéen (the traditional variety of part of the centre-ouest), for which he reports regularized plural endings containing [r] (e.g. donner: 1 pl [duɲiraŋ]; 2 pl [duɲirϵɪ]; 3 pl [duɲirαt]). Since these endings display person agreement, we might expect to find corresponding forms such as [dunirõ] in Acadian French varieties, given the fact that ancestors of present-day Acadians were of centre-ouest stock. However, to our knowledge such forms are unattested in Acadian French past or present. Thus the relationship between 20th-century centre-ouest and Acadian variants of the passé simple – with the latter having the status of derived forms or independent devolopments – is unknown.

5 POTENTIAL CONDITIONING FACTORS

A review of the literature revealed a number of potential conditioning factors.Footnote ⁹ For instance, the data were coded for inherent lexical aspect: for the present study, stative verbs were contrasted with non-statives (the Vendlerian categories accomplishment, achievement, and activity of verbs were thus not distinguished). The prediction is that stative verbs will be most often associated with imperfective marking, given that their basic meanings are not associated with temporal boundaries. For a discussion with respect to French, see Imbs (Reference Imbs1960). This is shown in 4a with rester ‘to stay’ in the imparfait; the non-stative verb sauter ‘to jump’ in 4b occurs in the passé simple. Of course, nonstative use of basically stative verbs is possible, as in 4c where the event is presented as a completed whole, with rester in the passé composé.

1. 4

   1. a Moi, je restais su Hector.Footnote ¹⁰ [stative verb]

      me 1-pers stay-imp at-the-house-of Hector

      ‘Me, I was staying at Hector's.’ (Amélie, GC-27)

   2. b Je sautis dans l'eau. [activity verb]

      1-pers jump-ps in the water

      ‘I jumped in the water.’ (Joanne, GC-10)

   3. c Il a resté jusqu'à tant que Joanne ait arrivé. [stative verb]Footnote ¹¹

      he has stayed until that Joanne has-subj arrived

      ‘He stayed until Joanne arrived.’ (Marie, GC-6)

With respect to sentential aspect, punctual, durative, and habitual contexts were distinguished. Punctuals would be expected to favour the passé composé and the passé simple and to disfavour the imparfait. Duratives and habituals would be expected to favour the imparfait. For discussion of these distinctions in French, see Maupas (Reference Maupas1618) and Grevisse (Reference Grevisse1985) to name but two works. These distinctions are illustrated in 5.

1. 5

   1. a Et puis là Joseph a arrivé. [punctual]

      and then there Joseph has arrived

      ‘And then Joseph arrived.’ (Evelyn, GC-27)

   2. b Je chasions [t∫eɪsjõ] les boeufs à Aldric.Footnote ¹² [habitual]

      1-pers chase-imp the bulls of Aldric

      ‘We used to chase Aldric's bulls.’ (Joanne, GC-10)

   3. c Le chat draggait [draga] le morceau de steak partout dans la place. [durative]

      the cat drag-imp the piece of steak everywhere in the floor

      ‘The cat was dragging the piece of steak all over the floor.’ (Carole, GC-6)

Also taken into account was whether or not there is an expression of simultaneity, anteriority, or sequentiality with the verb of the preceding clause. Simultaneity is largely agreed to be associated with the imparfait, and anteriority and sequentiality with both the passé simple and the passé composé.

1. 6

   1. a Il s'en allait. Il drivait [draiva] un, un, un ten-wheeler.Footnote ¹³ [simultaneous] he pron go-imp he drive-imp a a a ten-wheeler

      ‘He was leaving. He was driving a, a, a ten-wheeler.’ (Amélie, GC-27)

   2. b Marie itou a arrêté après qu'elle a été amener sa car à la garage. [anterior]

      Marie also has stopped after that she has been bring-inf her car to the garage

      ‘Marie also stopped by after she went to bring her car to the garage.’

      (Carole, GC6)

   3. c Il se levit. Il divit [dαivi] la tête là première dans la place.Footnote ¹⁴ [sequential]

      he refl get-up-ps he dive-ps the head there first in the floor

      ‘He got up. He dove headfirst onto the floor.’

      (Hector, GC-25)

In a variety in which the passé simple and the passé composé are both used productively, we would expect the former to be associated with more remote and the latter with more recent past. We were successful in coding for Estienne's 24 hour rule, i.e. whether or not the event described had taken place within 24 hours of the time of speaking, one way of distinguishing temporal distance, for all tokens.Footnote ¹⁵

1. 7

   1. a J’avons été nettoyer su ma mère aujourd'hui. [<24 hours]

      1-pers have been clean-inf at-the-house-of my mother today

      ‘We went to clean at my mother's today.’ (Marie, GC-10)

   2. b Bruce Wagner arrêtit icitte l'autre semaine. [>24 hours]

      Bruce Wagner stop-ps here the other week.

      ‘Bruce Wagner stopped here the other week.’ (Hector, GC-13)

The decision to code for discourse mode was prompted in part by descriptions of the passé simple as the narrative tense (much discussed by 20th century grammarians, as noted above, following a long tradition dating back to the 17th-century grammarians).

An earlier study of the Baie Sainte-Marie variety based on a small data set (reported on in Gesner, Reference Gesner1979a; Reference Gesner1979b) found that all 43 tokens for this tense occurred in narratives. It is also the case that even the most preliminary inspection of the Grosses Coques corpus reveals a clear narrative/non-narrative contrast. The framework used for narrative analysis is that of William Labov, given its prominence in sociolinguistic research on narrative and the fact that it lends itself fairly easily to quantitative analysis. All clauses were thus coded as narrative versus non-narrative, and, following Labov and Waletzky (Reference Labov, Waletzky and Helm1967) and Labov (Reference Labov1997), each narrative clause was identified by structural type, i.e. abstract, orientation, complicating action, coda, or evaluation. In the Labovian tradition, a narrative is defined structurally as a sequence of clauses which follows ‘the basic rule of narrative sequencing [. . .] which allow the listener to infer the reported temporal order to past events from the temporal sequence of clauses in the report of these events’ (Labov, Reference Labov, Lehman and Malkiel1982: 225). Thus we may identify narrative clauses as those whose temporal order may not be switched without changing the chronological order of events. While coding for discourse mode is found in a number of studies of past temporal reference in English, such as Elsness (Reference Elsness1997) and Lawrence (Reference Lawrence2000), we are unaware of prior studies for French which quantitatively investigate such effects. The data in 8 a and b come from a narrative of personal experience which will be presented in its entirety below.

1. 8

   1. a Bien je nous greyirent. Je nous levirent. Je nous habillirent. Je nous en furent back aux Grosses Coques. [narrative: complicating action]

      well 1-pers Refl get-ready-ps 1-pers Refl get-up-ps 1-pers Refl get-dressed-ps 1-pers Refl part go-ps back to Grosses Coques

      ‘Well we got ready. We got up. We got dressed. We went back to Grosses Coques.’ (Joanne, GC-10)

   2. b Je croyais qu'il était dans la camp à cause qu'il avait mouri là. [narrative: orientation]

      1-pers believe-imp that he be-imp in the camp because that he have-imp died there

      ‘I believed that he was in the camp because he had died there.’ (Joanne, GC-10)

   3. c Tu contais ça l'autre jour. [conversation]

      you tell-imp that the other day

      ‘You were saying that the other day.’ (Carole, GC-10)

Several other independent variables have been suggested primarily by the variationist literature on past temporal reference (e.g. Leroux, Reference Leroux2005 for Laurentian French, Schwenter and Torres Cacoullos, Reference Schwenter and Torres Cacoullos2008 for Castillian Spanish, and Van Herk, Reference Van Herk2004 for Early Modern English) which have themselves been largely derived from the functionalist literature on language change. These include whether or not there is “extra” specification of time or duration, by means of adverbial phrases (e.g. souvent ‘often’), prepositional phrases (e.g. en mai ‘in May’) or determiner phrases (e.g. ce matin ‘this morning’), as in 9.Footnote ¹⁶ As Comrie (Reference Comrie1985: 78–9, see also Bybee et al., Reference Bybee, Perkins and Pagliuca1994: 54–5, 61) notes, time-point adverbials generally favour the preterite. Below we consider separately the effect of particular adverbs, such as hier, suggested by the literature on French.

1. 9

   1. a J'ai un présent et je l'ai eu aujourd'hui. [temporal specification]

      1-pers have one present and 1-pers it have had today

      ‘I have one present and I got it today.’ (Zabeth, GC-12)

   2. b Ç’a duré presque la journée. [durative specification]

      it has lasted almost the day

      ‘It lasted almost all day.’ (Edwin, GC-24)

   3. c Robert a perdu trente piastres. [no such specification]

      Robert has lost thirty dollars

      ‘Robert lost thirty dollars.’ (Martin, GC-19)

The data were also coded for person and number of the subject, as in 10. Bybee et al. (Reference Bybee, Perkins and Pagliuca1994: 54–5, 60) state that crosslinguistically plural subjects, especially generic ones, favour the present perfect in the early stages of grammaticalization. All possible person/number distinctions were coded for in the present study.

1. 10

   1. a Et je callirent [kalir] docteur d'Eon et il venit. [1 P PL; 3 P SG]

      and 1-pers call-ps doctor d'Eon and he came-ps

      ‘And we called Dr. d'Eon and he came.’ (Zabeth, GC-12)

   2. b Il mouillait tout ce qu’il pouvait. [3 SG impersonal]

      it rain-imp everything that it can-imp

      ‘It rained a lot.’ (Marie, GC-6)

   3. c Le wire noir allait en bas. [3 SG inanimate]

      the black wire go-imp in below

      ‘The black wire went in the bottom.’ (Evelyn, GC-27)

Also taken into consideration was whether or not the tense of the clause in question matches that of the preceding clause, i.e. whether or not there is a priming effect, as in 11. See Poplack and Tagliamonte (Reference Poplack, Tagliamonte, Baker and Syea1996) for an illustration of the importance of this factor in past temporal reference marking in English, along with Szmrecsanyi (Reference Szmrecsanyi2005) for general discussion.

1. 11

   1. a So je rentrirent dans le liquor store et il braquit à garocher ce fusil là partout. [parallel]

      so 1-pers enter-ps in the liquor store and he start-ps to shoot-inf this gun there everywhere

      ‘So we went in the liquor store and he started to shoot this gun everywhere.’ (Richard, GC-29)

   2. b Bien elle a callé [kale]. Elle braillait. [non-parallel]

      well she has called she cry-imp

      ‘Well she called. She was crying.’ (Marie, GC-6)

Lastly, we coded for the usual speaker social factors, age and sex. Speakers 35-years-old and under were categorized in the younger cohort while speakers 45-years-old and over were categorized in the older cohort. Education was not included since it is strongly correlated with speaker age, as shown in Table 1.

7 COMPARISON WITH EARLIER STAGES OF THE LANGUAGE

As was outlined above, the sociolinguistic data which form the basis of this article come from a variety of French transplanted in the New World in the 17th century, a variety which has had little contact with normative French. In what follows we attempt to compare our results with those for earlier representations of (spoken) French. Caution is needed in interpreting data for earlier stages of the language, as researchers such as Ayres-Bennett (Reference Ayres-Bennett2004) and Lodge (Reference Lodge2004) discuss in detail. For instance, grammatical commentary is aimed at upper class speakers, or, perhaps more correctly, those aspiring to the upper class: it is most useful when it is proscribing particular usage where we may infer that the latter is actually representative of vernacular speech and thus to be avoided. Exactly whose vernacular is targeted is not always clear, although we may assume it to be that of urban, particularly Parisian, speakers, whose usage may not actually correspond to regional vernaculars, such as that of the centre-ouest, the main source area for Acadian emigration. Similarly, sources aimed to be representative of the spoken language, such as theatrical representations of lower class speech, either focus on urban speech or do not distinguish explicitly among regional varieties.

In contrast to 17th-century grammatical commentary, cited in section 2 and supported by recent corpus-based studies such as Galet (Reference Galet1974, plays and personal letters) and Évang (Reference Estienne1984, model dialogues), there is no association in the present-day Baie Sainte-Marie findings between use of the passé simple and use of particular adverbial expressions such as hier. On the other hand, there is a clear association with degree of remoteness (measured by the 24 hour rule) and with narration, with the latter exhibiting the stronger effect, in line with the earlier commentary. The overriding factor in the Acadian variety studied here is discourse mode, independent of whether the speaker is telling a traditional folktale or a recounting a personal experience narrative from a few days earlier.

In the works of 17th-century playwrights like Molière, known for his representations of lower class speech, use of the passé composé and other past tenses such as the plus-que-parfait far outnumber instances of the passé simple. Footnote ²² The relative lack of passé simple usage overall in such theatrical data may be linked to the fact that plays typically have much more conversational discourse than narrative discourse. Thus the diachronic data most similar in discourse type to our synchronic data is likely to be the narratives contained in theatrical representations. In Table 7 we present a quantitative analysis of past tense usage by narrative clause type in Molière's comedy Les Fourberies de Scapin, a play for which there are a sufficiently large number of narrative clauses to reveal patterns in usage. The data were coded using Labovian narrative conventions. Thus we do not include reports, which do involve the temporal sequencing of events but do not involve evaluation of the relevance of the account to the present discourse (Labov and Waletzky, Reference Labov, Waletzky and Helm1967).

Table 7.

Analysis of Past Tense Usage in Molière's Les Fourberies de Scapin

CA = complicating action clause; Other = other narrative clause

While it should be noted that the characters involved are members of the nobility or the bourgeoisie, there are clear points of reference with Acadian French usage described above.

In the analysis of the play data, the same exclusions (e.g. the ambiguous third person singular of dire) were made as for the Grosses Coques data. The most striking finding for the 92 narrative clauses retained for analysis is the near-complete absence of the passé composé from the first three narratives: it occurs only once, in one of a total of 15 complicating action clauses, with the remainder in the passé simple.Footnote ²³ The imparfait is found in 22 of 33 ‘other narrative clauses’ in these three narratives, typically in orientation clauses. But what of the two remaining narratives, which include a sizable number of passé composé tokens? Here we see the effect of temporal distance. The events recounted in the fourth narrative, which we have entitled ‘The Kidnapping 1’, are alleged by the narrator, Scapin, to have occurred within two hours of the moment of speech: in line with the 24 hour rule, all 15 complicating action clauses are rendered in the passé composé. As for the final narrative, ‘The Kidnapping 2’, the narrator Zerbinette first recounts events at some remove from the moment of speech, employing the passé simple: ‘En arrivant dans cette ville, un jeune homme me vit (CA), et conçut (CA) pour moi de l'amour. . .’ She later turns to events deemed to have taken place on that very day, and switches to the passé composé: ‘Il est allé trouver (CA) ce chien d'avare, ah, ah ah; et lui a dit (CA) qu'en se promenant sur le port avec son fils, hi, hi, ils avaient vu une galère turque où on les avait invités d'entrer.’ In recounting this more recent sequence of events, Zerbinette also makes use of the présent historique: ‘Il veut envoyer (CA) la justice en mer après la galère du Turc. Ah, ah, ah. Il sollicite (CA) son valet de s'aller offrir à tenir la place de son fils, jusqu'à ce qu'il ait amassé l'argent qu'il n'a pas envie de donner.’ The full narrative (actually a complex sequence of clauses with a narrative embedded within the main narrative), coded for complicating action clauses, is presented in the appendix.

Thus we see usage of past tense forms in Molière's play thoroughly in keeping with descriptions of 17th-century usage as noted above, and with usage in the present-day Grosses Coques corpus. Note that along with similar uses of the passé simple, the imparfait is entirely absent from complicating action clauses in both data sources.

While the 17th-century data are limited to representations of the spoken language from one play (obviously by one author), they are reflective of a period before which the passé composé had made inroads on the passé simple in narrative contexts and in which the latter was associated with temporal distance from the moment of speech. The play was first performed in 1671, but may reflect somewhat earlier usage, if Fournier and other commentators are correct in their claim that the passé simple was in decline by the second half of the 17th century. Most importantly, the data strongly resemble those for the isolated variety of North American French spoken more than three centuries later, examined in this paper.