Editor

JLLT edited by Thomas Tinnefeld
Journal of Linguistics and Language Teaching
Volume 3 (2012) Issue 2




Teaching the imparfait and the passé composé in French:

New Findings on Sequence

Valerie Wust (Raleigh (NC), USA) / 
Barbie Book Brown (Knightdale (NC), USA)

Abstract
Research on instructed second language (L2) learners of French consistently shows that the preterit emerges before the imperfect and remains the primary past tense marker until the advanced stages of acquisition. This study investigated whether instructional order (preterit-imperfect vs. imperfect-preterit) impacts upon past tense-aspect acquisition by high school learners of French. It was hypothesized that a sequence accounting for the increasing formal complexity of the two primary past tenses in French (e.g., imperfect before preterit) would positively affect mastery of the formation and distribution of both tenses, resulting in increased performance on a written cloze task. This hypothesis could be verified. The findings offer support for conducting further empirical investigations of grammar sequencing as a means of improving the effectiveness of instruction.
Key words: French as a second language, past tense, aspect, instructional sequence

Résumé
Les recherches sur l’acquisition du français par les apprenants en milieu scolaire montrent que le passé composé est traditionnellement enseigné avant l’imparfait et reste le marqueur du passé principal jusqu’aux stades avancés de l’acquisition. Dans la présente étude, la question est examinée de savoir si la séquence d’instruction (passé composé-imparfait vs. imparfait-passé composé) a un impact sur l’acquisition des temps du passé et de l’aspect par les apprenants de français inscrits à l’école secondaire. Nous partons de l’hypothèse qu’une séquence d’instruction tenant compte d’une complexité formelle croissante des deux temps du passé principaux en français (ex. l’imparfait avant le passé composé) a un impact positif sur la maîtrise de la formation et de la distribution des deux temps. La méthode choisie est un test de closure écrit. Dans la présente étude, cette hypothèse a pu être vérifiée. Pourtant, il importe de constater qu’avant que l’on puisse transférer les résultats de cette étude à l’enseignement pratique du français langue étrangère, d’autres études empiriques du même type devront être effectuées.
Mots clés: Français langue étrangère, temps du passé, aspect, séquence d’instruction


1   Introduction

Studies of past tense-aspect acquisition by Anglophone, instructed second language (L2) learners of French consistently generate two principle findings:

1)     the passé composé becomes productive before the imparfait; and
2)     form distribution varies across verb types.

The passé composé appears in conjunction with dynamic predicates (e.g. tomber, nager un kilomètre, manger), while the imparfait is used with a restricted group of statives, namely être, avoir, aimer, vouloir and pouvoir (Ayoun 2001, 2004, Bergström 1997, Harley 1992, Kaplan 1987, Salaberry 1998). These findings align with certain predictions of the aspect hypothesis, wherein the semantic properties of verbal predicates constrain the acquisition of perfective and imperfective markings (Andersen 2002). The aforementioned researchers, then, have focused on documenting a developmental sequence whereby imperfective usage spreads from stative to non-stative verbs, and perfective usage spreads from punctual to non-punctual verbs. However, instructional-, input- and form-related factors which could potentially contribute not only to the ‘early’ emergence of the perfective (in contrast to the imperfective), but also to the sustained use of the perfective as the preferred marker of past tense in French have been largely ignored to date.  

As relates to potential instructional influences, Kaplan (1987) argues that the past tense sequencing in mass-marketed L2 French pedagogical materials (passé composé before imparfait) contributes to the emergence of the former tense before the latter. Ayoun (2005) actually poses the question of whether the passé composé, as the preferred past tense form in the production of L2 French learners, is a product of instruction, given that the majority of textbooks and curricula present the passé composé before the imparfait (Ayoun 2005: 248). As for a potential role for input, Kaplan (1987) and Swain (1990) document an ‘input bias’ in classroom settings that could conceivably account for this well-documented developmental sequence of French past tense. More precisely, Kaplan (1987) reports that 84% of all past tense forms present in a sample of discourse from an L2 French classroom are in the passé composé. Swain’s (1990) analysis of classroom talk in grade six French immersion classes shows that 15% of all teacher-generated verbs are in the past, two-thirds of which are in the passé composé and one-third of which are in the imparfait. Acknowledging the importance of input, Ayoun & Salaberry concede that “…input biases may be all there is to the significant use of perfective forms in beginning and subsequent stages of acquisition’ (Ayoun & Salaberry 2005: 265). As for distribution-related factors, in her earlier work, Ayoun (2004) asserts that her data show that the imparfait is more difficult than the passé composé for English-speaking, university-level learners of French due its increased aspectual complexity (e.g., imperfective, iterative and durative values), as compared to the single aspectual value of the passé composé (perfective). An alternative, and opposing, form-related argument presented in Kaplan (1987) is that the passé composé (a compound tense with requisite auxiliary verb selection between avoir / être and many irregular past participles) is acquired first, despite being more grammatically-complex than the imparfait, which represents the morphologically most regular tense of French.

The extant research on the acquisition of past tense-aspect by Anglophone learners of French clearly shows that this is a problematic area in development. Researchers frequently cite the negative effects of L1 influence on French tense-aspect acquisition for these learners (e.g. Ayoun 2005 Izquierdo & Collins 2008, Kaplan 1987, Macrory & Stone 2000). Pedagogues, for their part, propose a litany of ‘best practices’ for teaching the passé composé and imparfait although no consensus has been reached to date (e.g. Blackburn 1995, Blyth 2005, Connor 1992, Gezundhajt 2000, Martin-Lau 2001, Moore 1981, Trèscases 1979). Taking the research agenda of the last thirty years in a new direction, the current study asks how the success of past tense-aspect instruction in L2 French, as measured by form- and tense-selection accuracy rates on a written cloze task, is affected by pedagogical sequence (passé composé-imparfait vs. imparfait- passé composé). As they are being taught the less formally complex imparfait  before the more complex passé composé, it is hypothesized that students in the imparfait- passé composé condition will more accurately identify appropriate usage contexts (e.g. tense selection) and form targetlike past tense exemplars (e.g. tense formation) on a contextualized production task.



2   Literature Review

2.1 Target Structures
In the current study, the target forms are the passé composé (preterit) and the imparfait (imperfect), which are the two primary tenses used by French speakers to communicate ideas in the past. Acquisition of these two tenses embodies three dimensions: morphosyntax (form), semantics (meaning) and pragmatics (use):



Passé composé (preterit)


Imparfait (imperfect)

Form

subject + auxiliary verb (present indicative) + past participle (lexical verb)
e.g. J’ai, tu as, il a, nous avons, vous avez, ils ont parlé.

subject + root of lexical verb (first person plural, present indicative) + inflectional endings, e.g. Je parlais, tu parlais, il parlait, nous parlions, vous parliez, ils parlaient


Meaning

- what one did

- what one was doing
- what one used to / would
  do
- what one thought / knew
(or any other stative verb)
for a certain period of time


Use

- completeness of past event, e.g., Pendant le semestre, j’ai beaucoup étudié

- enduring / ongoing actions
or state in the past,
   e.g. Vous vous parliez       au   téléphone.  
- a state or an action with  a certain duration in the
past, e.g. Nous ne
savions pas qu’il mourait.
- habitual action or state in
   the past, e.g. Le week-
   end, j’allais au cinéma.
   











        


















    Tab.1:Distribution (form, meaning, use) of French passé composé and imparfait

Table 1 clearly shows that the French passé composé and imparfait differ both in terms of formal and functional complexity, as can be seen in the form and use columns. For our purposes, formal complexity is determined by the number of transformations required from an underlying base form to the target form (e.g. from the infinitive of the lexical verb, and in the case of the preterit, the auxiliary verb). Functional complexity is operationalized as the degree to which a structure has multiple form-function mappings or the overall transparency of its meaning.  

In the current study, the passé composé is classified as having greater formal complexity than the imparfait, because it requires more transformations of its underlying base form to arrive at a well-formed exemplar. Here are the pedagogical rules provided to our learners:

1) As a function of the primary verb, determine which auxiliary verb to use (être for reflexive verbs and sixteen stative and motion verbs vs. avoir for all other verbs);  
2) Conjugate the chosen auxiliary verb in the present indicative, according to the person and number of the subject;
3) Determine if the primary verb has a regular past participle or not;
4a) If the verb is regular and belongs to the –er, -ir or –re class, drop the infinitive ending,
4b) and add the appropriate past participle ending: -é for –er verbs, -i for –ir verbs or –u for –re verbs;
5) If the verb has an irregular past participle, use the memorized form;
6) For être-verbs, determine the gender and number of the subject of the sentence (masculine or feminine; singular or plural);
7) Add the appropriate agreement ending(s) to the past participle (masculine singular -, feminine singular –e, masculine plural –s, feminine plural –es).

In comparison, the structure of the imparfait is relatively simple. To arrive at a well-formed exemplar, students were instructed to:

1) take the stem of the first person plural form of the present indicative of the verb;
2) drop the –ons ending;
3) add the endings of the imparfait in the corresponding person and number:


Singular

Plural


I -ais

I –ions


II -ais

II –iez


III -ait

III -aient

Tab. 2: French Imperfect Inflectional Endings

Based on the number of transformations from the base form to the target form for these two past tenses, this formal complexity analysis shows the potential interest in presenting the imparfait first in the L2 French classroom.

There is, however, also the question of the functional complexity of the passé composé and imparfait in French. As can be seen in Table 1 above, the imparfait is more functionally complex than the passé composé as it encompasses three difference semantic aspectual values (imperfective, iterative and durative). In contrast, the passé composé only embodies the perfective aspect. On the basis of the functional complexity of the two primary past tenses, it could be argued that we should continue to present the passé composé at an earlier point in the instructional sequence than the imparfait.

As we will later describe the methodology used in the study, we will show how our design will allow us to examine the differential effects of presenting both the more formally complex past tense (the passé composé) and the more functionally complex past tense (the imparfait) first in the instructional sequence.


3   Past Tense-Aspect Acquisition by Anglophone  
     Learners of French in Instructed  Settings

As the literature on the acquisition of past tense-aspect by L2 learners of French has been reviewed extensively by Ayoun (2001, 2004, 2005), we will focus our review specifically on trends which can be identified in select studies conducted on instructed learners in the North-American, and to a lesser extent, European contexts. Research on instructed past tense-aspect acquisition by English-speaking students in Canada (e.g. Harley 1992, Harley & Swain 1978, Izquierdo & Collins 2008, Kenemer 1982;, Knaus & Nadasdi 2001), in the United States (e.g. Ayoun 2001, 2004, 2005, Bardovi-Harlig & Bergström 1996, Bergström 1997, Kaplan 1987, Kenemer 1982, Salaberry 1998) and in Europe (e.g. Harris 1988, Howard 2001, Macrory & Stone 2000) has employed a variety of tasks, such as written clozes, oral interviews as well as oral and / or written narrations. Despite the differences in learner-related factors (e.g. child vs. adult learners), program-related factors (e.g. immersion programs vs. L2 programs) and task-related factors (e.g. oral vs. written mediums, highly structured tasks vs. open-ended tasks), the results are strikingly similar. The passé composé is the first past tense to become productive for beginning-level language learners (e.g. Ayoun 2004, Harley 1992, Kaplan 1987, Bergström 1997). These results align with the default past tense hypothesis, whereby in the early stages of acquisition learners mark tense distinctions, as opposed to aspect distinctions, and in doing so use the perfective form as their one and only past tense marker (Salaberry 1999, 2003). In early stages, the passé composé is used with telic verbs, such as se lever or écrire une lettre (e.g. Bardovi-Harlig & Bergström 1996, Harley 1992). The imparfait, which does not become productive until later in the course of acquisition, is used exclusively with the high-frequency statives avoir and être before its usage is extended to the domain of modals (e.g. vouloir, pouvoir) and other statives like détester, penser, habiter (e.g. Harley 1992, Harley & Swain 1978). As acquisition unfolds, English-speaking learners of French extend the scope in which the imparfait is used, adding atelic verbs, i.e. activities such as manger, courir and nager (e.g. Bardovi-Harlig & Bergström 1996, Harley 1992, Harley & Swain 1978). As proficiency increases, learners also use specific verbs in non-prototypical contexts: the passé composé with atelic predicates (e.g. balader dans le parc) and the imparfait with telic predicates (e.g. fermer la fenêtre), as in Salaberry (1998). Advanced English-speaking learners of French, in comparison to native speakers, have been shown to over-use the passé composé and to prefer prototypical uses of the passé composé and imparfait despite the fact that any verb can in either tense depend on the speaker’s perspective (e.g. Harley 1992, Salaberry 1998).

While the passé composé may be the first past tense which is used productively by English speakers learning French, form-related difficulties persist, even with what could be considered advanced-level learners. As the following examples clearly show, emergence of the passé composé and its mastery are two different things. Researchers note, in particular, that the presence of a dual auxiliary system and the necessity of variable agreement marking on past participles are problematic for beginning- to advanced-level L2 learners of French. For example, learners delete auxiliary verbs, creating simplified preterit forms, as in *Je ø parlé for J’ai parlé’ (e.g. Harris 1988, Kenemer 1982, Macrory & Stone 2000). The auxiliary verb avoir is overused, as in *J’ai parti instead of Je suis parti(e) (e.g. Harley 1992, Harris 1988, Kenemer 1982, Knaus & Nadasdi 2001, Macrory & Stone 2000). L2 learners also fail to mark subject agreement on past participles with être-verbs, as in *Elles sont tombé - masculin singular for Elles sont tombées, feminine plural (Kenemer 1982).

Not surprisingly, given its relatively decreased formal complexity, research on Anglophone learners’ of French production of the imparfait does not focus on form-related learner errors. ‘Distribution’ problems have been mentioned, such as present indicative forms being used in imperfect contexts (Kaplan 1987). More commonly, learners overuse the passé composé, also applying it to imperfect contexts (e.g. Ayoun 2004, Izquierdo & Collins 2008). Problems relating to the functional complexity of the imperfect, which encompasses three different semantic aspectual values, are documented in Ayoun (2004). In this study, the imparfait is primarily used by students in second- to fourth-semester university level courses in its durative function, with minimal tokens of its imperfective- and iterative-functions.

There are clearly two major ‘learnability’ issues identified in the instructed L2 French acquisition literature which are important for the current study:

·     the high-level of formal complexity of the passé composé, which has been shown to result in common interlanguage (IL) forms in learner production and
·   the functionally-complex imparfait, for which asymmetries in learner production of the three aspectual values have been documented.


4   The Study

The general aim of our research is to investigate the role of sequence in promoting the development of past tense-aspect in an instructed L2 learning context. Specifically, we examine possible differential roles for explicit, form-focused instruction as a function of whether a more formally complex past tense structure (e.g. passé composé) or a less formally complex one (e.g. imparfait) is presented first in the instructional sequence.

4.1 Participants
One hundred and three high school students enrolled in French II, aged 14-18 years, participated in the current study. The students attended two different high schools located in the same school board in the state of North Carolina. French II, which is typically taken during the second semester of instruction, is a block course, such that students receive instruction in the form of five ninety-minute classes per week over a five-month period. The first half of the course includes a review and expansion of grammatical structures such as the present indicative tense, object pronouns, reflexives and basic vocabulary related to food, clothing, homes, shopping and travel. During the latter part of the course, students are introduced to new structures, including the passé composé and imparfait, as well as vocabulary pertaining to childhood activities, health and feelings.
All of the students were enrolled in one of four intact classes taught by two female, National Board-certified educators, who were pursuing Master’s degrees in the same program. One instructor taught two classes of students using a traditional sequence of instruction in which the passé composé was presented before the imparfait, henceforth PRET-First. A second instructor taught two classes using a non-traditional sequence in which the imparfait was introduced before the passé composé, which will henceforth be referred to as IMP-First. Each high school was purposely selected due to similar profiles as determined by the socioeconomic status, race and achievement levels of the student body.

4.2 Materials
Over the course of four chapters from the county-mandated French II textbook, Allez, viens! French Level 2 which present the passé composé and the imparfait, the two experimental groups received instructional treatments that were identical in terms of content The treatment took place during the final two months of the instructional year and varied only in the sequencing of Chapters 5, 6, 7 and 9. The fifty PRET-First students were presented with a traditional learning sequence: treatment of the passé composé over the course of two chapters (5 and 6), followed by instruction on the imparfait for one chapter (7) and instruction on choosing between the two tenses in the final chapter (9). The fifty-three IMP-First students, who were enrolled at a second high school, received instruction first on the imparfait (Chapter 7), followed by the passé composé for two chapters (5 and 6) and instruction on the correct use of the two tenses in the final chapter (9). 
The instructors met before the onset of instruction on the past tense and prepared and agreed upon all of the content to which their students would be exposed. All students received identical grammatical explanation in the classroom, with deliberate focus on the two target tenses. In the presentation of the two tenses, both inductive and deductive strategies were used, with the goal of helping the students to understand and apply concrete rules for both forming and using the tenses in a systematic manner. The students were asked to complete the same mechanical, meaningful and communicative activities throughout the semester, beginning with discrete sentences and moving to longer stretches of contextualized discourse. Opportunities for practice consisted of speaking, reading, writing and listening tasks. Students also completed the same assignments outside of class, primarily from their workbook. In addition, the instructors prepared a timeline and decided how many instructional hours would be spent on each chapter. The two instructors regularly communicated with each other via e-mail so as to ensure that they were providing the same instruction and practice opportunities as originally agreed upon. To this end, the input and output activities in the two conditions were identical, with variations in the course instructor and the order of presentation of the two past tenses under examination.  

Following the four-unit instructional treatment, the learners were given two assessments measuring their ability to both conjugate the passé composé and the imparfait (e.g. form accuracy) and their ability to correctly select which tense would be most appropriate to use within the context provided (e.g. appropriate form distribution on a written cloze and a sentence-preference task). Learners began with a written cloze passage containing 27 target forms (Appendix A). The story, about a group of girls who get a flat tire on the way to their friend’s house for dinner, was developed by experienced L2 French instructors using verbs and other vocabulary from the North Carolina French II Standard Course of Study in order to facilitate overall comprehension of the text. A bilingual glossary of potentially difficult lexical items from the cloze was provided for students in an attempt to eliminate any potential interactions due to variations in individual vocabulary breadth. Native speakers checked the text for appropriateness. The 27 predicates of the cloze task targeted the imparfait (15 tokens) and the passé composé (12 tokens), and were distributed across four lexical aspectual categories (states, activities, accomplishments and achievements). Given the well-documented variability in IL (interlanguage) forms produced by L2 learners of French (e.g. Kenemer 1982, Macrory & Stone 2000), learners were required to indicate their tense selection by writing PC (= passé composé) or I (= imparfait) in the space provided (to indicate preterit and imperfect, respectively) as they completed the cloze. This served as a measure of tense-distribution knowledge. Tense-distribution information was supplemented by the learners’ actual conjugation of the targeted lexical verbs, which served as a measure of targetlike formation.  

As the goal of this study was to examine the impact of instructional sequence on past tense acquisition, and not to test the aspect hypothesis, no attempt was made to balance the number of tokens across either tenses or aspectual categories. Preference was given to creating a natural-sounding text, with target sentences that mirrored the demands of the language to which the students were regularly exposed in their classes. Obligatory contexts were determined as a function of the responses of four native speakers of French, who served as a control group for this pilot study.

4.3 Procedure
All instruction and assessment occurred in the students’ classrooms, during regularly-scheduled class times. The treatments were administered by the students’ regular instructor, while a single researcher administered the two assessments (written cloze; sentence-preference task) on the same day. Students were given adequate time (never exceeding 45 minutes) to complete the two assessments, which were supplied in a numbered envelope to ensure anonymity. Only the results of the written cloze will be examined in the current article. Following from Ayoun (2005), the participants were instructed to read the short story, containing a series of blanks followed by the infinitive form of lexical verbs, in its entirety before beginning the production component of the task. They were encouraged to use the bilingual glossary at the end of the text for lexical items that were marked with an asterisk to enhance their comprehension. Students were told that their task was two-fold:

1)    to indicate their tense selection for each targeted verb by writing PC or I inside the square bracket provided (to indicate preterit and imperfect, respectively); and
2)    to conjugate the lexical verb in the chosen past tense, making all necessary agreements in the case of the preterit.

Students were instructed that once they had completed the cloze task, they were to return it to the envelope containing their assessment materials and not permitted to return to it once they had begun the sentence-preference task.

4.4 Coding and Analysis
The cloze data was coded in a binary fashion (1 = correct, 0 = incorrect) for both tense distribution and tense formation. As was the case in Izquierdo and Collins (2008), the cloze data was coded twice by the same rater to ensure coding reliability. For imperfect tokens, the stem and the ending were coded separately, for a maximum of two points. To this end, one point was still awarded in the case of errors in agreement between the imperfective inflection and the head noun phrase (e.g. *ses amies était en route (3rd person singular) instead of of ses amies étaient en route (3rd person plural). Full points were awarded in instances in which appropriate diacritics were absent in the imparfait, such as the third person singular form etait (instead of était) written without an accent aigu. However, in the preterit, the absence of an accent aigu (chante instead of chanté) in the past participles of -er verbs such as chanter [to sing] was considered incorrect as it made the form graphically indistinguishable from the first- and third-person singular verbs in both the present indicative and subjunctive moods. For passé composé tokens, auxiliary verb selection and conjugation, past participle formation and agreement markings on the past participle were all coded individually for each token, for a maximum of four points. For example, two points were deducted in cases in which the learner used an inappropriate auxiliary in a perfective construction (e.g. *elles ont arrivées chez Michèle (feminine plural + incorrect auxiliary choice) instead of the accurate form elles sont arrivées chez Michèle (feminine plural + correct auxiliary choice). Similarly, one point was deducted in cases where the past participle was ill-formed (e.g. *elles ont voit une station-service (feminine plural + incorrect past participle)   instead of the accurate form elles ont vu une station-service  (feminine plural + correct past participle). A point was also deducted in cases in which correct agreement was not marked on the past participle for être-verbs (e.g., *ses amies …. se sont dépêché (feminine plural + incorrect agreement on past participle) instead of the accurate form ses amis … se sont dépêchées (correct agreement on past participle). This fine-grained coding scheme allowed for more detailed insight into students’ IL representations of past tense forms.


5   Results

5.1 Tense Selection
The data were first analyzed to see whether students were able to correctly identify contexts on the cloze in which the passé composé (PC) or imparfait (I) should be used. Students indicated their tense selection directly on the cloze-task sheet inside the square brackets preceding each targeted lexical verb. Native-speaker responses indicated that fifteen imperfect tokens and twelve preterit tokens should be used (for a maximum of twenty-seven points). In the fifteen imparfait contexts, students in the IMP-First group had a mean score of 10.06, and students in the PRET-First group had a mean score of 8.44. In the twelve passé composé contexts, IMP-First students had a mean score of 9.53, compared to those in the PRET-First group who had a mean score of 6.90. Students in both conditions were more accurate in the identification of appropriate contexts for imperfect, as opposed to preterit, usage.

A one-way ANOVA test revealed that in imparfait contexts, the difference in means between the PRET-First and IMP-First groups was statistically significant (F (1, 101) = 7.080, p < .01). In passé composé contexts, the difference in means was statistically significant to an even greater degree (F (1, 101) = 24.072, p < .001). We see that the IMP-First group consistently and significantly outperformed the PRET-First group on tense selection on the cloze task. This result simply means that the IMP-First learners were better at distinguishing between contexts necessitating the use of the passé composé and the imparfait after two months of instruction on past tense-aspect. The fact that the difference in means between the two groups reached greater significance in passé composé -contexts, as opposed to imparfait -contexts, is particularly striking precisely because IMP-First learners were exposed to the passé composé later in the instructional sequence than learners from the PRET-First condition. Given the comparable levels of students in the two conditions prior to the treatment as determined by a one-way ANOVA test (F (1, 91) = .001, p = .974), these results could indicate that students in the IMP-First group outperformed their peers in the PRET-First group on tense selection despite being presented with a more functionally-complex past tense (the imperfect: three functions) before a functionally ‘simpler’ one (the perfective: one function).

5.2 Tense Formation
A second analysis of the data allowed us to determine whether the students were able to produce targetlike forms of both the passé composé and the imparfait on a written cloze task. For accurate tense formation in imperfect contexts, students in the IMP-First group had a mean score of 16.64 (out of 30) and students in the PRET-First group had a mean score of 13.88. In passé composé contexts, IMP-First students had a mean score of 29.19 (out of 48) and PRET-First students had a mean score of 13.38.
In Table 3, a component analysis of tense formation on the written cloze is provided, including differences between numeric means for correct formation of the stem and ending for the imparfait contexts and correct selection and formation of the auxiliary verb, correct formation of the past participle and correct agreement markings on the past participle for the passé composé contexts. In all instances, and as group, students in the IMP-First group outperformed those in the PRET-first group:



Condition

M

SD

SE


Score for correct stem (imperfect) /15

IMP

8.74

2.33

.32




PRET

8.14

2.67

.38


Score for correct ending (imperfect) /15


IMP

7.91

3.36

.46


PRET

5.74

3.33

.47


Score for correct aux. selection (preterit) /12

IMP

7.00

4.12

.57




PRET

2.58

2.99

.42


Score for correct aux. conjugation (preterit) /12

IMP

6.21

4.27

.59




PRET

1.62

2.26

.32


Score for correct past participle (preterit) /12

IMP

8.79

2.35

.32




PRET

5.08

3.64

.51


Score for correct agreement (preterit) /12

IMP

7.19

2.59

.36




PRET

4.10

2.77

.39

Tab. 3: Component analysis of tense formation on the written cloze (Descriptive statistics)

A one-way ANOVA test revealed that in imparfait contexts, the difference in means for correct tense formation between the PRET-First and IMP-First groups was statistically significant (F (1, 101) = 7.531, p < .01). In passé composé contexts, the difference in means for correct tense formation reached even greater statistical significance (F (1, 101) = 53.849, p < .001). Once again, the results show that the greatest impact of instructional sequence on the formation of the past tense rests with the use of the passé composé, a more formally complex tense requiring multiple transformations from its base form (the infinitive of the lexical verb). The difference in numerical mean scores for the formation of the passé composé is about five times larger than the difference in mean scores for the formation of the imparfait, in favor of students in the IMP-First group. A difference of this magnitude could offer support for our hypothesis whereby learners who are introduced to a ‘simple’ past tense before a ‘complex’ one (from a formal complexity standpoint) would generate more accurate tokens of both tenses on a production task, such as a cloze, for reasons of developmental readiness.

In summary, these results provide tentative support for our initial hypothesis that the sequence of presentation of the imparfait and passé composé has a significant impact on learners’ abilities to both correctly identify appropriate usage contexts (e.g. tense selection) and to form targetlike past tense exemplars (e.g. tense formation) on a contextualized production task such as the written cloze.


6   Discussion

The general question addressed in this study is how the success of past tense-aspect instruction in L2 French, as measured by form- and tense selection-accuracy rates in written production, is affected by pedagogical sequence. Or in even more general terms, is one past tense more teachable than another as a function of its complexity? Researchers working with instructed L2 learners of French have made suggestions to this effect without actually testing them empirically (e.g. Ayoun 2004, 2005, Kaplan 1987). In this study, we hypothesized that a sequence accounting for the increasing formal complexity of the two primary past tenses in French (e.g. imparfait before passé composé) would positively affect L2 French learners’ mastery of the formation and distribution of both tenses, resulting in increased performance on a written cloze task. This hypothesis could be verified.

In many ways, it is difficult to compare our results with those of the extant research on the acquisition of past tense in L2 French precisely because ours is the first study to examine the impact of presentation order on the acquisition of the passé composé and the imparfait. Comparisons to Kaplan’s (1987) study, in which beginning-level university students were taught the imparfait before the preterit, are less than ideal as her data are oral, as opposed to written, and were collected using questions and cues for eliciting past tense utterances. Kaplan’s study design is also characterized by a number of weaknesses, which she herself acknowledges, such as the impossibility of distinguishing between certain homophonous passé composé / imparfait forms in oral data and exaggerated passé composé accuracy rates dues to chunking phenomenon. For this reason, we compared learner performance in our two conditions instead, using the form and distribution error rate comparison suggested in Kaplan (1987). The students in the IMP-First group in our study make fewer distribution-related errors in both past tense contexts than do their counterparts in the PRET-First condition (21% vs. 42.5% for preterit contexts; 33% v. 44% for imparfait contexts, respectively). As for the correct formation of the two past tenses, IMP-First students commit fewer errors than those in the PRET-First group (39% v. 72% for PRET contexts; 45% v. 54% for IMP contexts, respectively). It is noteworthy that learners in the IMP-First condition not only perform better on the formation of the more formally complex perfective tense, they also have a better mastery of the distribution of the more semantically complex imperfective tense, which encompasses imperfect, durative and iterative meanings. The rates of form- and distribution-related errors show that learners in both conditions still need significant exposure to and practice of the passé composé and imparfait so as to achieve native-like mastery. This is not surprising as they are only in their second semester of language study, particularly given the abundant research in L2 French detailing the prolonged developmental trajectory of past tense acquisition. What is more important, however, is that our research suggests that past tense instruction may be more beneficial, in terms of accurate form production and distribution on a written cloze, when beginning level learners are presented with the imparfait before the passé composé.

Clearly, the results of this study must be interpreted with caution given certain limitations of the design.

First, acting under the advisement of a statistician, we compared the percent means of the percent final course grades in French I of students in both the PRET-First and IMP-First groups, using a one-way ANOVA test, which indicated no significant differences prior to the administration of the past tense treatment in the current study. For future studies, the use of a standard pre-test measure as a baseline allowing for comparisons of the mastery of important syntactic structures (e.g. regular and irregular present indicative verbs) by students in both conditions is recommended so as to better determine potential differences prior to administering the instructional treatment.

Second, and for practical reasons, the study was designed such that each instructor and her two intact classes were assigned to the same treatment (e.g. instructor 1 = 2 PRET-First groups; instructor 2 = 2 IMP-First groups). In future empirical investigations on the role of grammar sequencing in improving the effectiveness of instruction, it would be desirable to control for teacher effect as a possible confounding variable by having each instructor deliver both treatments to two intact sections of the same course. While the current study controlled for teacher input and student opportunities for practice, it cannot be stated unequivocally that the superior performance of IMP-First students is solely due to instructional sequence rather than teacher factors, such as personality or rapport with students.


7   Conclusion
Previous research on the acquisition of past tense and aspect by instructed learners of L2 French shows early emergence of the preterit as compared to the passé composé (e.g. Ayoun 2001, Bergstrøm 1997, Harley 1992). The current study is, to the best of our knowledge, the first one to use empirical data to challenge the assumption that the so-called ‘early’ emergence of the passé composé is not, in fact, due to instructional sequence (possibly in combination with another factor like form frequency in classroom input). Our findings show that despite a delayed introduction to the passé composé, IMP-First students correctly mark perfective and imperfective contexts with the passé composé and imparfait, respectively, more frequently than do PRET-First students. Moreover, students in the IMP-First group are more accurate in the targetlike formation of these two tenses.

If these preliminary findings for increased accuracy on the identification of appropriate contexts for passé composé and imparfait usage and targetlike formation of these two tenses as a function of order of presentation can be supported by future research, they will have profound implications for L2 teaching and syllabus design. By challenging the traditional passé composé -before-imparfait sequence which abounds in L2 classrooms and mass-marketed pedagogical materials, teachers will be able to capitalize on what may be more developmentally appropriate orders, potentially increasing the rate of acquisition. A sequencing approach to the acquisition of past tense-aspect by instructed learners suggests an alternative answer to the longstanding debate between researchers and teachers, alike, as to the most effective way to help students learn the preterit / imparfait nuance. To be clear, we are not arguing that the way in which the past tenses are taught (e.g. explicit vs. implicit, deductive vs. inductive) has no influence whatsoever on acquisition as this is beyond the scope of our current research. Rather, taking a similar stance to Pienemann (1989), we maintain that the influence of teaching is restricted to the learning of items for which learners are developmentally ready. Teaching can only promote acquisition by presenting the learner with that which is learnable at a given point in time. In other words, grammatical forms characterized as ‘developmental’ (such as the passé composé and imparfait in French) could be positioned in a syllabus according to their learnability as a function of their formal and / or functional complexity.

The fact that learners in the IMP First condition in our study significantly outperformed their peers in the PRET-First condition on both the formation and the distribution of the two primary French past tenses constitutes an argument for conducting further studies on the impact of sequencing in instructed L2 acquisition. This is not a call for a return to the morpheme studies which went out of fashion in the early 1980s (e.g. Dulay & Burt 1973, Larsen-Freeman 1975, Pica 1983), but rather a chance to explore whether empirical investigations of grammar sequencing might provide solutions for enhancing instructional outcomes for the acquisition of difficult aspects of L2s, such as verbal and pronominal systems. As a first step in this direction, our results point to the potential efficiency of teaching less formally complex target forms (passé composé) before their more complex counterparts (imparfait), even in cases in which the less complex form has multiple form-meaning mappings.




Appendix: The Written Cloze Task

Instructions
Throughout this task, you will read a short story, which contains a series of blanks followed by a verb in the infinitive form. For each verb, you are to decide whether it should be in the passé composé or the imparfait and then conjugate the verb appropriately. Before each conjugation, confirm the tense you chose by writing a [PC] to indicate the passé composé or an [I] to indicate the imparfait. If you hesitate between forms, choose the one that makes the most sense out of the two. It is recommended that you read the entire passage carefully as the context will help you choose the appropriate verb form. A vocabulary list is also provided at the end of the story for the words and phrases that are followed by a star.
Il [ ] ….. (être) une fois, une fille qui [ ] ….. (faire) la tête*, parce qu’elle [ ] ….. (vouloir) sortir avec ses amies au parc. Elle [ ] ….. (être) triste parce qu’il [ ] ….. (pleuvoir) et qu’elle [ ] ….. (devoir) finir ses devoirs pour le lycée. Tout à coup, elle [ ] ….. (avoir) une très bonne idée. Elle [ ] ….. (téléphoner) à ses amies et elle [ ] ….. (inviter) ses amies chez elle pour dîner et regarder la télévision. Ses amies [ ] ….. (accepter) l’invitation et [ ] ….. (se dépêcher)* de partir de chez elles. Pendant* ce temps, Michèle [ ] ….. (ranger) le salon et elle [ ] ….. (nettoyer) la cuisine. Pendant que* ses amies [ ] ….. (être) en route*, leur voiture [ ] ….. (tomber) * en panne. A 7h30 du soir, il [ ] ….. (faire) noir et elles [ ] ….. (avoir) assez peur. Elles [ ] ….. (pleurer)* au moment où elles [ ] ….. (voir) une station-service. A la station-service, elles [ ] ….. (acheter) des bonbons et un mécanicien [ ] ….. (réparer) la voiture. Enfin, quand elles [ ] ….. (arriver) chez Michèle, elle [ ] ….. (faire) la cuisine. Puis, Michèle [ ] ….. (demander) pourquoi elles [ ] ….. (être) en retard. Elles [ ] ….. (expliquer) la situation pendant qu’elles [ ] ….. (manger) le dîner.

Vocabulary List (in alphabetical order)
en route                  on the way
faire la tête              to sulk
pendant                   during
pendant que            while
pleurer                    to cry
se dépêcher            to hurry
tomber en panne      to break down


References

Anderson, Roger. (2002). The Dimensions of Pastness. In Rafael Salaberry & Yasuhiro Shirai (eds.). The L2 Acquisition of Tense-Aspect Morphology, pp. 79-105. Amsterdam: Benjamins.

Ayoun, Delila. (2001). The Role of Negative and Positive Feedback in the Second Language Acquisition of the passé composé and imparfait. Modern Language Journal, vol. 85, pp. 226-243.

Ayoun, Delila. (2004). The Effectiveness of Written Recasts in the Second Language Acquisition of Aspectual Distinctions in French: A Follow-up Study. Modern Language Journal, vol. 88, pp. 31-51.

Ayoun, Delila. (2005). The Acquisition of Tense and Aspect from a Universal Grammar Perspective. In Delila Ayoun & Rafael Salaberry (eds.). Tense and Aspect in Romance Languages, pp. 79-128. Amsterdam: Benjamins.

Ayoun, Delila., & Salaberry, Rafael. (eds.). (2005). Tense and Aspect in Romance Languages. Amsterdam: Benjamins.

Bergstrøm, Anna. (1997). L’influence des distinctions aspectuelles sur l’acquisition des temps en français langue étrangère. [The Influence of Aspectual Distinction of Tense Acquisition in French as a foreign language]. AILE, vol. 9, pp. 51-82.

Blackburn, James. (1995). Teaching Aspect in Beginning French Courses. Dialog on Language Instruction, vol. 11 (1-2), pp. 33-47.

Blyth, Carl. (2005). From Empirical Findings to the Teaching of Aspectual Distinctions. In Delila  Ayoun & Rafael Salaberry (eds.). Tense and Aspect in Romance Languages, pp. 211-252. Amsterdam: Benjamins.

Connor, Meryl. (1992). A Processing Strategy Using Visual Representation to Convey the passé composé/imparfait Distinction in French. International Review of Applied Linguistics in Language Teaching, vol. 30 (4), pp. 321-329.

K-12 Curriculum and Instruction: NC Standard Course of Study. World Languages 2004.High School Level II. Department of Public Education. Public Schools of North Carolina. Retrieved from: http://www.dpi.state.nc.us/curriculum/worldlanguages/scos/2004/32level2.

Dulay, Heidi, & Burt, Marina. (1973). Should We Teach Children Syntax? Language Learning, vol. 23, pp. 245-258.

DeMado, John., & Rongieras d’Usseau, Emmanu. (2000). Allez viens! French 2. New York, NY: Holt, Rinehart and Winston.

Gezundhajt, Henriette. (2000). Pour un enseignement aspectuo-modal et illustré des temps du passé. [Teaching Aspect and Mood in Past Tense]. Canadian Journal of Applied Linguistics, vol. 3 (1-2), pp. 63-78.

Harley, Birgit. (1992). Patterns of Second Language Development in French Immersion. French Language Studies, vol. 2 (2), pp. 159-183.

Harley, Birgit., & Swain, Merrill. (1978). An Analysis of the Verb System Used by Young Learners of French. Interlanguage Studies Bulletin, vol. 3, pp. 35-79.

Harris, Vee. (1998). Natural Language Learning and Learning a Foreign Language in the Classroom. British Journal of Language Teaching, vol. 26 (1), pp. 26-30.

Howard, Martin. (2001). The Effect of Study Abroad on the L2 Learners’ Structural Skills. EUROSLA Yearbook, vol. 2, pp. 87-113.

Izquierdo, Jesus., & Collins, Laura. (2008). The Facilitative Role of L1 Influence in Tense-Aspect Marking: A Comparison of Hispanophone and Anglophone Learners of French. Modern Language Journal, vol. 92, pp. 350-368.

Kaplan, Marsha. (1987). Developmental Patterns of Past Tense Acquisition Among  Foreign Language Learners of French. In Bill VanPatten, Trisha Dvorak, & James  Lee (eds.). Foreign Language Learning: A Research Perspective, pp. 52-60. New York: Newbury House.

Kenemer, Vera. (1982). Le français populaire and French as a Second Language: A Comparative Study of Language Simplification. Québec: CIRB.

Knaus, Valere., & Nadasdi, Terry. (2001). Etre ou ne pas être in Immersion. Canadian Modern Language Review, vol. 58 (2), pp. 287-306.

Larsen-Freeman, Diane. (1975). The Acquisition of Grammatical Morphemes by Adult ESL Students. TESOL Quarterly, vol. 9, pp. 409-420.

Macrory, Gee, & Stone, Valerie. (2000). Pupil Progress in the Acquisition of the Perfect Tense in French: The Relationship between Knowledge and Use. Language Teaching Research, vol. 4 (1), pp. 55-82.

Martin-Lau, Philippe. (2001). Bambi Rules: A Visual Approach to imparfait – passé  composé Usage in French. Francophonie, vol. 23 (Spring), pp. 20-23.

Moore, Victoria. (1981). L’anglophone et l’apprentissage du passé en français. (Anglophones and the Learning of Past Tense in French]. Le Français dans le Monde, vol. 21 (160), pp. 39-44.

Pica, Teresa. (1983). Adult Acquisition of English as a Second Language under Different Conditions of Exposure. Language Learning, vol. 33, pp. 465-497.

Pienemann, Manfred. (1989). Is Language Teachable? Psycholinguistic Experiments and Hypotheses. Applied Linguistics, vol. 10, pp. 217-244.

Salaberry, Rafael. (1998). The Development of Aspectual Distinctions in L2 French Classroom Learning. Canadian Modern Language Review, vol. 54 (4), pp. 508-543.

Salaberry, Rafael. (1999). The Development of Past Tense Verbal Morphology in Classroom L2 Spanish. Applied Linguistics, vol. 20, 151-178.

Salaberry, Rafael. (2003). Tense and Aspect in Verbal Morphology. Hispania, vol. 86, pp. 559-573.

Swain, Merrill. (1990). Manipulating and Complementing Content Teaching to Maximize Learning.  In Eric Kellerman, Robert Phillipson, Larry  Selinker, Michael Sharwood Smith, & Merrill Swain (eds.). Foreign/Second Language Pedagogy Research: A Commemorative Volume for Claus Færch, pp. 235-250. Clevedon, UK: MultilingualPress.

Trèscases, Pierre. (1979). Une pédagogie de l’imparfait et du passé composé. [A Pedagogy of the imparfait and the passé composé]. Le Français dans le Monde, vol. 19(148), pp. 60-65.




Authors:

Valerie Wust, Ph.D.
Assistant Professor of French Applied Linguistics
Department of Foreign Languages and Literatures
North Carolina State University
Raleigh
USA
E-mail: vawust@ncsu.edu

Barbie Book Brown
French Teacher
Department of World Languages
Knightdale High School
Knightdale
USA
E-mail: bbrown1@wcpss.net