: Tense , aspect and narrative structure in oral narratives of students and native speakers of French : A comparative study Turkish

Bu calismanin amaci Fransizca ogrenen Turk ogrencilerin ve Fransizca anadil konusurlarinin sozel anlatilarindaki dilbilgisel zaman, gorunus, kilinis ve soylem ozelliklerini karsilastirmali incelemektir. Kesitsel desenli calismanin orneklemini Cukurova Universitesi Egitim Fakultesi Fransiz Dili Egitimi Anabilim Dalinin birinci, ikinci, ucuncu ve dorduncu siniflarindan 93 ogrenci ile anadili Fransizca olan 16 Fransiz katilimci olusturmaktadir. Katilimcilara Walt Disney yapimi Lion King ve A Christmas Carol adli cizgi sinemalarin Fransizca seslendirmeli bicimleri izletilmis ve bu filmleri sozlu olarak yeniden anlatmalari istenmistir. Veriler kilinis hipotezi ve soylem hipotezi cercevesinde degerlendirilmistir. Bulgular, hem Fransizca ogrencilerinin hem de Fransiz katilimcilarin sozel anlatilarini yuksek oranda prA©sent dilbilgisel zamani ile yapilandirdiklarini gostermistir. Sozel anlatilarin prA©sent agirlikli olmasi, kilinis ve soylem hipotezlerinin passA© composA©/imparfait dilbilgisel zamanlari ile kilinis siniflamasindaki eylem turleri ve anlati yapilari arasindaki etkilesim konusunda one surdukleri savlari bir olcude devre disi birakmis ve Fransizca ogrencilerinin eylem secimlerinde ne soylem yapilarinin ne de eylemlerin ickin anlamsal ozelliklerinin butunuyle etkili oldugunu ortaya koymustur. Bulgular ayrica soylem yapilari ve eylemlerin ickin anlamsal ozelliklerinin etki derecelerinin katilimcilar arasinda farkliliklar sergiledigini gostermistir. Buna gore ucuncu ve dorduncu sinif ogrencilerinin dilbilgisel zaman secimlerinde soz konusu faktorler birinci ve ikinci sinif ogrencilerine oranla daha etkili olmustur.


Introduction
As foreign language teachers working at university, it is not uncommon to hear students in preparatory class or first year who say, when encouraged to get to know other students from upper classes in order to receive help with some language subjects: "But upper classes don't know French (or German, or English) better than us!" Or they convey that these students tell them: "your French is better than ours, we have forgotten French!" Taken from one of the oral narratives for this study, following example (1) illustrates well this situation.These utterances from a senior student seem to verify what is told by students from lower classes: These utterances reflect morphological errors, simple in appearance but serious yet when we think that these are errors by future teachers of French who are not very far from working in the field.We, with our colleagues, are of the opinion that similar errors need to be dwelled on.In addition, there should be no direct relationship between students' grades and the severity of their errors since students with good grades in courses also fall into these errors.Even though these errors were corrected incessantly when these students were in preparatory class and they cannot be able to graduate to first grade without any optimal correction of all these errors, how is it that their linguistic level cannot progress in upper classes?Or, if we say more clearly, do they really have a weaker linguistic level as they approach the end of their studies?Discussions we make on the subject with colleagues from the field not only of French language teaching but also of teaching of other foreign languages seem to require positive answers to these questions.Obviously, other means are clearly needed to be able to perfectly define this problem otherwise what have been told so far would remain hypothetical.All these observations, increasing interest in research on foreign language learners' interlanguages in terms of verbal morphology and lack of cross-sectional studies that focus on verbal morphology of Turkish L1 learners of French as a foreign language led the researcher to carry out this study.Aiming also at contributing to scientifically clarify relationship between progression and linguistic level at undergraduate study as noted before, the present study is theoretically based to two hypotheses: the aspect hypothesis in the context of tense, lexical and grammatical aspect and the discourse hypothesis in the context of tense, aspect and narrative structure.
Providing within itself all the means to refer to time, the concept of temporality has always been an important focus in second language acquisition research.Three concepts are worthy of attention when it comes to acquisition of temporality expressing systems: tense, lexical aspect and grammatical aspect.These concepts have taken part in many descriptive and pedagogical explanations about language.Development of research in acquisition of temporality generally reflects development process of research in second language acquisition as well (Bardovi-Harlig 2000).

Tense, Lexical Aspect and Grammatical Aspect
Tense is a deictic category that allows fixing a process on the time line.In French, Alexandre voit Cécile and Alexandre a vu Cécile are utterances which indicate tense distinctions.Tenses not only convey information about time but also express in what way we envisage the course of the process and its mode of manifestation in time (Maingueneau 1991).This information is called grammatical aspect.
Grammatical aspect is a system of rigorous oppositions instituted by verbal morphology like oppositions of endings (Maingueneau 1991).It provides different ways of looking at situation types (Comrie 1976;Dahl 1985;Smith 1983).In French, there are two forms of past: perfective and imperfective.The passé composé (PC) expresses perfective aspect and situates the utterance in relation to the utterer.The process is completed at the time of the enunciation.The imparfait (IMP), on the other hand, expresses the imperfective aspect and situates the utterance in an indeterminate moment before the moment of enunciation.This indeterminacy is likely to be interpreted as duration, repetition, continuity or state (Maingueneau 1991).Utterances such as Alexandre a dormi (passé compose; simple past) and Alexandre dormait (imparfait; past progressive) show an aspectual contrast although both are in the past tense.
Although grammatical aspect and lexical aspect are two different linguistic categories, it is nevertheless impossible to ignore one when discussing the other.The distribution and interpretation of grammatical aspect are influenced by lexical aspect.
Lexical aspect differs from grammatical aspect.While the latter is essentially a grammatical category like PC and IMP, the former is only lexical which refers to intrinsic temporal features of verbs and predicates.A predicate may assume a process that has an inherent duration such as parler "speak" and dormir "sleep"; a punctuality such as éclater "explode" and remarquer "notice"; both an inherent specific duration and a specific endpoint such as contruire une maison "build a house" and peindre un tableau "paint a picture" or a state such as vouloir "want" and aimer "like/love".In literature, there are two types of aspectual category that have been used in studies dealing with lexical aspect: binary and quaternary categories.
Binary categories of lexical aspect contrast predicates in terms of simple features such as stativity vs dynamicity, durativity vs punctuality or telicity vs atelicity.Static/dynamic opposition distinguishes states (such as sembler"seem" and savoir "know") from other predicates (e.g., jouer "play", lire un livre "read a book", éclater "explode").Durative/punctual opposition distinguishes predicates that can be considered as instantaneous or insegmentables and which have no duration (éclater) from those with duration (chanter "sing", chanter une chanson "sing a song").Telic/atelic opposition distinguishes predicates with a specific endpoint (chanter une chanson) from those who do not (chanter) (Maingueneau 1991).
Most of the early studies on aspect used binary distinctions (Housen 1993(Housen , 1994;;Kaplan 1987;Robison 1990).Binary distinctions have had some undesirable results like grouping of different predicate types (Bardovi-Harlig 2000).For example, durative/punctual opposition distinguishes predicates that can be considered as instantaneous from those with duration, but it also groups stative predicates with dynamic ones in the non-punctual category.Vendler (1967) proposed a quaternary classification, directly classifying verb processes into four aspectual categories such as states, activities, accomplishments and achievements.Each category has distinctive features.
States persist over time.They are not interruptible.They have no beginning, no end and no middle; they assume neither agent nor change.They are not compatible either with punctual indications or with en un + nom temporel as in ( 2) and (3):
Activities are processes with an inherent duration and without temporal closure because they involve a space of time, such as dormir and nager.They do not have a specific ending, as in j'ai nagé toute la matinée (I swam all morning).Activities do not combine with en un + nom temporel (4) but with punctual indications as in (5): 4. *Il a nagé en une heure (He swam in one hour) 5. Il a nagé à midi (Maingueneau 1991, 52).
Achievements, as opposed to accomplishments and activities, are not segmentable, have no duration (apercevoir "see, perceive"), and seize the beginning or end of an action (Mourelatos 1981).They can be considered as reduced to a point (Andersen 1991).They combine with punctual indications as in, Il a aperçu Jean à 8 heures (He saw Jean at 8 o'clock) but little with pendant + syntagme nominal as in * Il a aperçu Jean pendant une minute (He saw John for a minute) (Maingueneau 1991, 52).Arriver "arrive", quitter "leave", remarquer "notice", reconnaitre "recognize" and s'endormir "fall asleep" are examples for achievements.
Accomplishments are processes that have a specific endpoint (like achievements) and an inherent duration (like activities).They are true only when the process has come to an end.Accomplishments are hardly compatible with punctual temporal indications as in *il écrit une page à 8 heures, but associate well with en un + nom temporel as in il écrit une page en un mois (Maingueneau 1991, 52).Construire une maison and peindre un tableau can be given as examples of accomplishments.
Accomplishments and achievements can be grouped as telic predicates, known as "events" (Mourelatos 1981).States and activities can be grouped as atelic predicates.Vendler's categories can be divided into three features: [± punctual], [± telic] and [± dynamic].The [+ punctual] feature distinguishes achievements from all other categories.The [+ telic] feature distinguishes predicates with a specific endpoint (nager 400 mètres "swim 400 meters") from those who do not (jouer), and thus distinguishes achievements and accomplishments from activities and states.The [+ dynamic] feature distinguishes dynamic predicates (e.g., courir, lire un livre, se réveiller "wake up") from static predicates (like sembler and connaitre).Andersen (1989) who used the quaternary classification postulated four stages in the acquisition of perfective past (PC): from achievements to accomplishments, to activities, and finally to states.Imperfective past (IMP) appears later than perfective past and spreads to four stages: from states to activities, to accomplishments, and finally to achievements.Although imperfective past emerges after perfective past, their stages overlap.Many studies have applied the principles established by Andersen regarding the spread of verbal morphology to a variety of other languages (Bardovi-Harlig 2000).

Aspect Hypothesis
The aspect hypothesis is based on a theory of lexical or inherent aspect that differs from grammatical aspect.While the latter is an essentially morphological category such as IMP and PC, the former is only lexical which relates to the temporal structure of verbs and predicates.

513
The aspect hypothesis asserts that at early stages in the acquisition of temporal reference, learners associate lexical aspect with grammatical aspect.They tend to use verbal morphology to mark lexicoaspectual distinctions (stative vs. dynamic events, durative vs. punctual events, telic v. atelic events) rather than temporal distinctions.Andersen (1991) showed that achievement verbs are first marked in perfective past (PC) before they gradually spread to accomplishment, to activity and finally to state verbs.In contrast, verbs that are stative by nature are first marked with imperfective past (IMP) which indicates durativity before they spread to activities, to accomplishments and finally to achievements.Bardovi-Harlig (1998) used oral and written narratives in a study with learners of English to test the aspect hypothesis.The oral data support this hypothesis better than the written data.The oral data indicated a net increase in the simple past from achievements and accomplishments to activities even though achievements and accomplishments were used with similar rates.Cloze tests already used by Bardovi-Harlig, and Reynolds (1995) also showed no difference between these two verbal categories.Robison' study (1993) with Spanish-speaking learners of English also validated the aspect hypothesis.Results obtained from oral interviews showed that the use of simple past was the highest with achievements.The use of simple past with all aspect categories has increased as the language level develops.In another study, Robison (1995) observed that the association of progressive marker with activities is strengthened according to the linguistic level.The use of present tense with states has also been observed in contexts requiring the imperfective past.Salaberry (1999) argued that the effect of lexical aspect at early stages of L2 Spanish should not be as widespread as at advanced stages.Nevertheless, a reanalysis of Salaberry's (1999) data by Bardovi-Harlig (2000) showed that the data truly follow predictions of the aspect hypothesis.On the basis of data obtained from intermediate and advanced Spanish learners by a cloze test, Salaberry (2002) found that the effect of lexical aspect does not seem to be stronger among advanced learners than among intermediate learners.

Discourse Hypothesis
According to Godfrey (as cited in Bardovi-Harlig 2000), the use of temporal morphology could not be completely understood regardless of discourse.Cross-linguistic studies have pointed out that the distinction between foreground (FG) and background (BG) is a universal feature of discourse (Bardovi-Harlig 2000).An early study by Kumpf (as cited in Bardovi-Harlig 2000) found that there is a relationship between the use of verbal morphology in interlanguage and narrative structure.The discourse hypothesis predicts that learners use verbal morphology to distinguish FG from BG in their narratives (Bardovi-Harlig 1994).
According to Hopper (as cited in Bardovi-Harlig 2000), FG implies events belonging to the dynamic structure of discourse.For Dry (as cited in Bardovi-Harlig 2000), FG consists of clauses which succeed one another and move narrative time forward.The most basic narratives of low-level learners consist only of clauses in FG.The information provided in FG clauses must be new rather than given.
BG has different functions that support FG. Events reported in FG proposals are sequential while BG events are often out of sequence with respect to FG. Hopper (as cited in Bardovi-Harlig 2000) claims that BG does not contain main events but serves as a support that details or evaluates events in FG.BG also reports events that precede the events in FG.Dahl and Hopper Hopper (as cited in Bardovi-Harlig 2000) argue that the narrative context influences temporal reference more than the tense itself.Dry (as cited in Bardovi-Harlig 2000) reports that FG clauses are usually formed with perfective past (PC) or historical present while BG clauses include imperfective past (IMP).
Research into learners' narratives has shown that the narrative context influences the distribution of temporo-aspectual morphology.Early studies on the effect of the narrative context were mainly case studies (Flashner 1989;Housen 1994;Treviso 1987;Véronique 1987), while subsequent studies were conducted with more learners.
European Science Foundation (ESF) carried out a project on the acquisition of temporality by immigrant workers in European countries (Bhardwaj, Dietrich & Noyau 1988).The target languages were English, German, Dutch, French and Swedish.The project focused not only on tense and aspect but also on narrative structure.Although the objective of this project was not to test directly lexical aspect, especially one of the results drew back attention to a situation emphasized in studies on the acquisition of French.Although French has an aspectual distinction for past reference, there is no evidence in the data which shows that even advanced learners acquire it.Results also showed that events in learner narratives reported in FG presented an inherent boundary and BG events tended to be rather durative and to appear at the same time as certain events reported in FG.This supports the discourse hypothesis which predicts that learners distinguish FG from BG regardless of verbal morphology.Likewise, Véronique (1987) claimed that the use of verbal morphology in the interlanguage of African workers was not determined by FG and BG.Harley (1992) who conducted a cross-sectional study to examine developmental patterns in the acquisition of L2 French by Canadian English-speaking immersion students in terms of tense and aspect within the framework of properties of French verbal system, found a relationship between tense chosen and lexical aspect.PC was employed particularly with dynamic verbs while IMP was lexically limited to a small number of durative verbs such as aimer, avoir, être, pouvoir, savoir and vouloir.To express temporality, students in this study, like participants in the ESF project, tended to rely on the discursive context or time adverbials rather than on verbal morphology.This suggests an early stage of acquisition that is universal where lexical and discursive means play a more important role than verbal morphology (Fan 2005).Bardovi-Harlig and Bergström (1996) tested the aspect hypothesis.Learners of English and French were asked to retell an 8-minute excerpt from the silent movie Modern Times.Results revealed the effects of lexical aspect on the acquisition patterns in both English and French.For French, IMP appeared later than PC.IMP marking began with states spreading next to other verb classes while PC had an early strong association with achievements, followed respectively by accomplishments, activities and sates.As in Harley (1992), IMP was also found to be lexically restricted to some durative verbs (Fan 2005).Kaplan (1987) conducted an error analysis in a study on patterns of development of PC and IMP acquisition by freshman and sophomore learners of French at university.The results showed that the use of PC has a higher accuracy rate than IMP.Indeed, learners produced PC verbs four times as much as IMP verbs.IMP was often presented by présent (PR) while relatively few PR verbs were used for PC.Kaplan also asserted that, in grammatical time marking, the precedence given to grammatical and not lexical aspectual organization may be considered as a universal feature of language acquisition.If the distinction is not temporal, it must be aspectual.Salaberry (1998) conducted a study to examine the development of aspectual markers in English speaking learners of French and native speakers of French in control group.Results obtained from a written narration of a short film and a cloze test showed that the selection of past tense marking by learners of French was in accord with that by native speakers in terms of the prototypical relations between grammatical aspect and inherent semantic values of verbs.Data from the cloze test indicated that learners had a different use of non-prototypical grammatical aspect in L2 French from that in native speakers.As for the narrative task, the results showed that the use of PC and IMP between telic and atelic verbs was significantly different for both groups.According to Salaberry, extensive use of PC by learners in both tasks can be considered as evidence to see PC as a default past tense marker.The written task revealed over-dependence in non-native speakers on the use of this potential default marker of the past tense.Kihlstedt (2002) examined the acquisition of tense and aspect in advanced learners of French.The results showed that lexical aspect did not influence the choice of PC and IMP by learners more than that by native speakers.Kihlstedt indicated that form precedes function.When past inflexions begin to spread to non-prototypical combinations (telic verbs in IMP or states in PC), they do not necessarily express the same functions as those in the target language.Howard (2004) tested the aspect and discourse hypotheses in a study with advanced English speaking learners of French.The results of this study showed that neither narrative context nor lexical aspect determine alone verbal morphology of the learners.The results also demonstrated that there are interrelations between many other factors such as phonetic quality of verb (regular or irregular verbs), syntactic context, and influence of other markers of temporality, such as adverbs of time.Fan (2005) investigated the processes in the acquisition of tense-aspect morphology by Anglophone learners of French and Chinese.Grammaticality judgement and cloze tests were used as data collection instruments.The results of the data obtained from learners of French as well as those of Chinese supported the prototype hypothesis.The distribution and accuracy of the use of temporo-aspectual morphology in both languages was influenced by intrinsic semantic features of predicates.
Aiming to comparatively examine features of tense, aspect and narrative structure in oral narratives of Turkish L1 learners of French as a foreign language and native speakers of French within the context of aspect and discourses hypotheses, the present study attempts to provide answers to following research questions:  Do oral narratives of Turkish learners and native speakers of French share similarities in terms of lexical/grammatical aspects and the use of verbal morphology?
 To what extent do oral narratives of Turkish learners and native speakers of French resemble each other with respect to narrative structure (foregrounding /backgrounding) and the use of verbal morphology?
 To what extent do lexical aspect and narrative structure influence the use of verbal morphology in oral narratives by Turkish learners and native speakers of French?Hypotheses 1. Learners use telic verbs with PC and atelic verbs with IMP at early stages of second language acquisition.This is called lexical dependence.It loses its influence as learners' level of proficiency increases.If learners use PR verbs independently of telic / atelic properties, this shows that this effect begins to decrease.In other words, the more the learners use verbs in PR the less the effect of lexical dependence is strong.In this case, junior and senior students expected at more advanced language level would use more PR verbs than freshman and sophomore ones.
2. Learners first use PC with mainly achievements at early stages of second language acquisition before spreading PC use respectively to accomplishments, activities and finally to states as they improve their linguistic level.They first use IMP with states before spreading it respectively to activities, accomplishments and finally to achievements.In that case, compared to freshman and sophomore students, junior and senior ones expected at more advanced language level would use, other than achievements, more PC verbs with accomplishments, activities and states and other than states, more IMP verbs with activities, accomplishments and achievements.
3. Within the context of narrative structure, learners at early stages of second language acquisition use PC with telic verbs in FG and IMP with atelic verbs in BG.Learners get rid of the necessity of using PC in FG and IMP in BG as their proficiency level increases.Using verbs in PR means disrupting the lexical dependence above mentioned.In this case, junior and senior students expected at more advanced language level would use, independently of telicity/atelicity, more PR verbs in FG and BG than freshman and sophomore ones.

Method Research Design
The present study is cross-sectional.Cross-sectional design along with longitudinal one has been one of the most commonly used research design since the earliest investigations on verbal morphology in language learners (Bardovi-Harlig 2000).Thus the ESF project in which the acquisition of temporality by immigrant workers was investigated in target languages such as English, German, Dutch, French and Swedish had some cross-sectional studies beside longitudinal ones (Bhardwaj, Dietrich & Noyau 1988).There are several cross-sectional studies in literature that focused on tense, grammatical and lexical aspect and narrative structure (see, for example, Harley 1992;Hasbún 1995;Kaplan 1987;Salaberry 1999;Schlyter 1990).

Participants
Ninety three students in first, second, third and fourth years of French Language Teaching Department at Çukurova University participated in the study.These students formed the experimental group.With an age range of 19-24, they had German or French as selective foreign language course at high school.Out of 93 participative students, 73 had German, 13 had French and 7 had no foreign language course.The control group was formed by 16 native speakers of French, of whom 2 were in their twenties, 12 in thirties and 2 others in forties.French participants were all university graduates.

Data Collection Tools
For investigations on grammatical and lexical aspect, there were mainly two types of data collection instruments: written/oral narratives and cloze tests (Bardovi-Harlig & Reynolds 1995;Bergström, 1995;Collins 1997Collins , 1999)).As data collection tool, the present study used a film retell task as part of oral narrative.As the content is known by the researcher, this instrument allows the researcher to effectively follow narrative elements related by participants and to compare students from the point of view of their success in the transmission of these elements.When looking at movies chosen for film retell tasks in literature, it is seen that these were mostly silent movies (e.g.Bhardwaj, Dietrich & Noyau 1988;Chafe 1980;Dietrich, Klein, & Noyau 1995).The most important reasons for choosing silent movies as data collection instruments were the pellucidity of action scenes and scene changes.The clarity of action scenes provides advantages in correctly encoding verbs produced by participants within the context of aspect hypothesis while the comprehensibility of scene changes provides advantages in properly investigating FG/BG distinction made by participants within the context of discourse hypothesis.
Two French dubbed versions of animated films produced by Walt Disney were used as data collection tools in the present study.These are A Christmas Carol adapted in 1983 from the famous novel by Charles Dickens and the 1994 Lion King movie.Participants had a certain familiarity with these films.Both films have preponderantly more action scenes than conversations.From this aspect, it can be said that they have a structure which resembles that of silent movies.It has been thought that the abundance of action scenes would allow participants to produce as many verbs as needed in order to investigate the aspect hypothesis.In the films, events as well as set and scene changes are clearly distinctive.It was thought that this would provide convenience for participants to distinguish FG from BG within the context of the discourse hypothesis and to facilitate data production.

Data Collection
The data of this study were collected out of course hours.The researcher carried out negotiations with students to determine dates and to prepare a program both for film watching and retelling.In these negotiations, students were kept informed of the study and its process steps.Students of French first watched A Christmas Carol.Before starting to watch the movie, students were informed once more by the researcher about what to do after watching the movie.The movies were separately shown for each class due to time lag.This prevented the researcher from performing film retell tasks immediately after watching the movie.Students carried out film retell tasks approximately one week after watching the film.Even if there was no time limit for film retell tasks, students averagely accomplished the task in one to five minutes.Before viewing the film, students were given a few minutes so that they could ask for words and structures they did not know or they wanted to remember.The same procedure was followed for Lion King.
French participants only watched A Christmas Carol.This has two reasons.Firstly, native speakers of French produced a greater number of utterances in their oral narratives than students.Indeed, the number of utterances produced by students who presented data for both films is considerably lower than that of native speakers.Secondly, the researcher had difficulties in finding native speakers of French and convincing them to participate in the study during data gathering process.At the time of this study, almost all French people said they had time problems and could only present data for only one film.

Data Analysis
Compared to tests that present strict controls such as cloze tests, oral narratives (film retelling task) is a data collection tool that ensures production of free utterances and allows participants to produce their own texts.Oral narratives of participants were recorded with a voice recorder.Recorded data were then saved on personal computer of the researcher to be later transcribed verbatim and coded for analyses.
After being transcribed, verbs in participants' oral narratives were classified and coded according to verbal morphology and lexico-aspectual categories.Tests previously used by Bergström (1995) and Salaberry (1998) to determine lexico-aspectual categories were applied.These are être en train de test (Guenthner, Hoepelmann, & Rohrer 1978) and en X minutes test (Dowty 1979;Guenthner, Hoepelmann, & Rohrer 1978;Nef 1980).The être en train de test is used to distinguish states and achievements (that cannot pass the test) from activities and accomplishments (that pass the test) while the en X minutes test is used to distinguish accomplishments from all other verb types.All verb types (states, activities, accomplishments and achievements) produced by participants were coded in terms of three grammatical tense: PC, IMP and PR.Some researchers who conducted studies in tense-aspect literature (Bardovi-Harlig & Bergström 1996;Bardovi-Harlig & Bofman 1989;Harley & Swain 1978;Kaplan 1987;Trévise 1987) exhibited positive attitude towards incorrect verb types produced by participants and did not leave them out of assessment.This was the case for the present study.Oral narratives were then coded in terms of narrative structure in the light of discussions in literature (Dry 1981(Dry , 1983;;Dowty 1986).Thus, clauses making it possible to convey events which succeed one another and move narrative time forward were coded as FG clauses while others which do not have this narrative function were coded as BG clauses.

Findings
Participants used four grammatical tenses: PR, PC, IMP and plus-que-parfait (PQP).PQP was excluded from analyses because it was used very little, if any, and no participant, except one junior and one senior, had recourse to this grammatical tense.Table 1 shows all verbs and predicates obtained from participants through oral narratives.Total number of predicates produced by participants is 4913.Students of French employed 4258 predicates while native speakers of French used 655.The first notable thing about tense use of participants is the dominance of PR compared to other tense types.Indeed, both students and native speakers of French used 4052 predicates in PR in their oral narratives, which accounts for 82.48% of all predicates in the corpus.
Table 1 indicates that freshman students produced 87.46 % of overall predicates in PR, sophomores 89.78 %, juniors 58.60 % and seniors 55.30 %.This rate was 92.53 % for native speakers of French in control group.Chi-square analysis showed a significant difference among students in PR use ( (6,4251) 2 = 525.15,p = .00).These results revealed that freshman and sophomore students resemble to native speakers in terms of PR choice more than junior and senior ones.This shows that, in contrast to the case asserted in the first hypothesis, junior and senior students were influenced by lexico-aspectual properties of verbs more than freshman and sophomore ones, and therefore, had a lower level of language.
When verbs in PR were examined in the context of lexico-aspectual properties, it is seen that achievements and states came to the fore.Table 2 gives the distribution of PR verbs in terms of lexicoaspectual categories.As can be seen in Table 2, although sophomore students' use of activities (36.23 %) was close to that of achievements (36.77 %), achievements were used more than other verb types in freshmen and sophomores (respectively 36.50 % and 36.77 %).The rate of use for state verbs in PR by junior and freshman students was quite superior to other types of verbs (respectively 48.50 % and 59.50 %).As for native speakers, it is seen that they also preferred states more than other lexico-aspectual classes (46.30%).

519
PC was the second most used tense by students and native speakers of French in the oral film retell task.Total number of PC verbs produced by all participants is 536, which constitutes 10.91 % of all verbs in corpus (see Table 1).When the use of PC verbs is examined among participants, we see that it is senior students that used this tense the most (31.44 %), and sophomore students that used it the least (7.51 %).This rate was 8.50 % in freshman students while it was 21.74 % in junior ones.French participants had quite a little use of PC (3.66 %).Andersen (1991) asserts that learners at early stages of second language acquisition use PC mainly with achievement verbs before they gradually spread it to accomplishment, to activity and finally to state verbs as their proficiency level increases.Likewise, Robison (1993) states that the use of PC in all aspectual classes, that is to say, states, activities, accomplishments and achievements, is only possible as long as one has an advanced linguistic level.On the basis of these theoretical considerations, as stated in the second hypothesis, junior and senior students expected at more advanced language level would use, other than achievements, more PC verbs with accomplishments, activities and states compared to freshman and sophomore ones.Table 3 gives the distribution of PC verbs in terms of lexico-aspectual categories.When examining PC verbs in the context of lexico-aspectual properties, it is seen that achievements were dominant in oral narratives of French language students.It can be said that they used PC verbs with achievements with close rates (71.20 % in sophomores and seniors, 62.30 % in freshmen and 65.60 % in juniors).Compared to native speakers, students used approximately twice as many PC verbs with achievements.Although they used achievements with similar rates, Chi-square analysis revealed a significant difference among students with respect to other verb types (χ (9,511) 2 = 19.12,p = .02).However, it is hard to say that, compared to freshman and sophomore students, junior and senior ones distinctly used more PC verbs with other aspectual classes other than achievements.Thus, sophomore and junior students used PC verbs with states (respectively 10.20 % and 10.10 %), freshman and senior students used PC verbs with activities (respectively 19.50 % and 20.04 %) with close rates and even freshman and sophomore students used more PC verbs with accomplishments than junior and senior ones.In this case, it can be said that what is asserted in the second hypothesis was not fully supported.French participants used three times more PC verbs with states and nearly twice as many with activities than students.Although results indicated that students began to spread PC use to accomplishments, activities and states, this spread was not as equal as in native speakers of French.French participants used no PC verbs with accomplishments.
As regards IMP, students and native speakers of French made use of this tense in oral narratives for 325 verbs which constitutes 6.61 % of all verbs in the corpus (see Table 1).A between-participants examination demonstrates that it is junior students that resorted to this tense the most (19.66 %), and sophomore students that used it the least (2.71 %).When looking at other students, it is seen that seniors used this tense three times more than freshmen (respectively 13.26 % and 4.04 %).This rate was 3.81 % for native speakers in control group.
When looking at IMP verbs in the context of lexico-aspectual properties, it is seen that states were more dominant than verbs of other aspectual classes.Table 4 gives the distribution of IMP verbs in terms of lexico-aspectual categories.States constituted about three fifth of all IMP verbs (respectively 60.00 % and 57.14 %) in sophomore and senior students while this was more than half in freshmen and juniors (respectively 51.20 % and 53.90 %).Activities were the second most used tense with IMP after states in oral narratives of participants.Thus, atelic verbs (states and activities) formed four fifth of IMP verbs in students while this was more than four fifth in native speakers.As noted previously, learners at early stages of second language acquisition use IMP mainly with states before they spread it to activities, to accomplishments and finally to achievements.With reference to this theoretical assumption, as stated in the second hypothesis, junior and senior students expected at more advanced language level would use, other than states, more IMP verbs with activities, accomplishments, and achievements compared to freshman and sophomore ones.However, Chi-square analysis revealed no significant differences among participants in terms of IMP use with activities, accomplishments, and achievements ( χ (9,290) 2 = 8.06, p = .53).
Following the examination of tense-aspect properties of students and native speakers of French in their oral narratives, discourse features were analyzed in what follows.
Table 5 gives the distribution of verbs in participants' oral narratives in terms of discourse structure (foregrounding/backgrounding).As is seen in the table, all participants used more BG verbs than FG verbs.Three fifth of verbs produced by students of French in oral narratives were in BG.This was more than four fifth in French participants (84.70 %).As is shown in the table, all participants used more PR verbs in BG.More than half of PR verbs produced by freshman and sophomore students were in BG (respectively 57.30 % and 55.40 %) while this rate was about two third in junior and senior ones (respectively 62.50 % and 77.20 %).French participants employed three fourth of PR verbs in BG (82.80 %).These data demonstrate that all students, especially junior and senior ones, seem like native speakers regarding high incidence of PR verbs in BG.However, Chi-square analysis revealed significant differences among participants with respect to PR use in FG and BG ( χ (3,3443) 2 = 29.41,p = .00).In the third hypothesis, it was asserted that, independently of telicity/atelicity, junior and senior students would use more PR verbs in FG and BG than freshman and sophomore ones.Although this hypothesis was supported with regard to PR use in BG, it was not upheld in terms of PR use in FG because freshman and sophomore students produced more PR verbs in FG (respectively 42.70 % and 44.60 %) than junior and senior ones (respectively 37.50 % and 22.80 %).
When analyzing FG and BG in the context of PC verbs, it is seen that, in contradiction to PR verbs, participants used more PC verbs in FG than in BG.Table 7 gives the distribution of PC verbs in FG and BG according to participants.
Freshman and sophomore students, as can be seen in the table, used two third of PC verbs (respectively 67.10 % and 57.90 %) in FG while junior and senior students employed more than two third (respectively 70.90 % and 75.00 %).Unlike students, native speakers of French predominantly used PC not in FG but in BG (91.70 %).Although PC use in FG and BG differs among participants, Chi-square analysis indicated that these differences were not statistically significant ( χ (3,510) 2 = 7.02, p = .07).When dealing with oral narratives in the context of narrative structure and IMP, it is observed that all participants dominantly employed IMP in BG.Indeed, as can be seen in Table 8, students used about four fifth of IMP verbs in BG while French participants used more than four fifth.Although Chi-square analysis showed no significant differences among participants in terms of IMP use in FG and BG ( χ (3,295) 2 = .92,p = .82),it can be said that students resemble to native speakers of French regarding predominance of IMP in BG.We present, in what follows, results aiming at understanding whether verbal morphology of participants was affected by lexical aspect and narrative structure.Before going on to present detailed answers to this question with results, let us recall the assertions of two hypotheses, namely the aspect hypothesis and the discourse hypothesis, examined in the study in term of tense choice.
The aspect hypothesis asserts that learners at early stages of L2 are exposed to effects of lexical properties of verbs and it is these properties which determine their use of verbal morphology.According to this hypothesis, these learners use PC mainly with telic verbs and IMP with atelic verbs.For the discourse hypothesis, it is rather narrative structure which is determinative in verbal morphology.Accordingly, learners at early stages of L2 use PC with telic verbs in FG and IMP with atelic verbs in BG.Both hypotheses also assume that the effects of lexical aspect and narrative structure in determining the use of verbal morphology begin to lessen as learners' proficiency level increases.
When looking at results within the context of the aspect and discourse hypotheses, the first notable thing is again that participants structured their oral narratives considerably with PR.Table 9 gives the distribution of PR verbs and PC/IMP verbs in both FG and BG.As is seen in the table, 82.44 % of verbs were used in PR while 17.56 % of verbs in in PC and IMP.
The fact that participants structured their oral narratives largely with PR may be seen as an indication that neither lexical aspect nor narrative structure is effective anymore in the use of verbal morphology of participants.

Discussion, Conclusion and Implementation
This study aims to comparatively study features of grammatical and lexical aspect and narrative structure in oral narratives of Turkish L1 learners of French as a foreign language and native speakers of French and to investigate whether intrinsic semantic features of verbs and narrative structure influence verbal morphology of participants.

Do oral narratives of Turkish learners and native speakers of French share similarities in terms of lexical/grammatical aspects and the use of verbal morphology?
When comparing participants with regards to temporo-aspectual properties, it is seen that both students and native speakers of French formed their oral narratives substantially with PR.Having as equivalents both the present tense (geniş zaman) and the present continuous tense (şimdiki zaman) in Turkish, PR is one of the basic elements of French tense system.PR, as in Turkish again, may express past and future events.PR conveys a past event in the following example (7): 7. Hier, je vais chez lui, sa mère veut que j'entre (Maingueneau, 1991, 65) (Yesterday I go to his house, his mother wants me to enter) Temporal information is generally conveyed by adverbs of time in such utterances.But PR is also used to express past events without using any adverbs of time when the context is known to both speakers and listeners.In the present study, participants also used PR to retell events in movies that they watched and that should be told with past tenses (PC/IMP).They employed PR for these events without any adverbs of time or having only recourse to some adverbs such as un jour (one day) or un matin (one morning) in order just to show the course of time in narratives.Examples (8) and ( 9 When reviewing studies in tense-aspect literature, it is noticed that some researchers included PR verbs in their works while others excluded them in order to focus only on past tenses.In a study with controlled tasks, Hasbún (1995) told student participants to commence their narratives with habia una vez (once upon a time) which allows beginning narration with past tenses.Likewise, Bergström (1995) afforded a written prompt to students of French participating in her study in an effort to incite them to structure their narratives with past tenses.The works of Robison (1995) and Bada and Genç (2007) may be cited as examples of studies where tasks were not strictly controlled.During data collection, these researchers gave no instructions to the participants in order not to guide them with respect to tenses they should use in their oral narratives.Few PR verbs, if any, are chosen in the case of studies with strictly controlled tasks where participants are instructed to use of PC and IMP in their narratives at the time of data collection.The use of PR verbs increases when these are not studies with controlled or strictly controlled tasks.Consequently, the presence of numerous PR verbs in the present study can be explained by the non-use of design with controlled or strictly controlled tasks during data collection.
Freshman and sophomore students resembled more to native speakers of French in terms of PR choice density while higher lexical dependence was observed in junior and senior ones.Even if the fact that junior and senior students approach native speakers with regard to the rate of use for states in PR seems, at first glance, to contradict what has just been said, a closer examination would reveal that the abundance of states in PR in native speakers arises from the fact that they have frequently used, in oral narratives, presentational cleft constructions beginning with c'est (the c'est-cleft) which are very common.Indeed, the great majority of states produced by native speakers are constituted by the verb être in presentational cleft constructions.The abundance of states in PR in junior and senior students is not due to these constructions.The use of the verb être in this context is considered non-lexical and excluded from analyzes by some researchers (e.g.Bardovi-Harlig 1998).In the present study, all occurrences of the verb être, including that in presentational cleft constructions, were included in analyses in order to obtain natural environment of utterances produced and to completely compare students and native speakers of French.An example (10) of the verb être in presentational cleft constructions is given below: 10.C'est pas affreux c'est plutôt euh ….c'est plus affreux que sympathique je trouve, c'est assez … assez triste comme endroit -la maison d'Ebenezer Scrooge, Picsou en français (FP 1) ‡ "It's not ugly it's rather uh ... it's more awful than friendly, I think, it's pretty ... pretty sad as place -Ebenezer's house" PR has been considered as default tense by some researchers in tense-aspect literature.Kaplan (1987) noted that, before IMP, French learners used PC instead of IMP.It was also found that learners of English (Bardovi-Harlig & Reynolds 1995;Robison 1995) and Danish (Housen 1993(Housen , 1994) ) used PR with states.Some researchers indicated that PR was used to distinguish states from non-states (dynamic verbs).In this sense, Kaplan (1987) stated that learners of French as a foreign language used PC to report events (activities, accomplishments and achievements) and PR to report non-events (states).Results of the present study revealed that students of French did not use PR to distinguish non-events from events.
PC was most associated with achievements and employed with similar rates by students of French.French participants interestingly had no recourse to accomplishments in PC.Rarity in accomplishments was also a conspicuous point in Fan's (2005) study but the researcher did not deal with probable causes of this fewness.We think that the scarcity of accomplishments in the present study can be explained by the fact that, in movies constituting a basis for oral narratives, there were few situations which required this type of verb.
IMP was most associated with states and employed with close rates by students of French.French participants used states in IMP with very high rates.The reason is that native speakers used presentational cleft constructions with c'était (IMP) a lot in oral narratives.‡ FP1 (French Participant 1)

Research Question 2
To what extent do oral narratives of Turkish learners and native speakers of French resemble each other with respect to narrative structure (foregrounding/backgrounding) and the use of verbal morphology?
When comparing participants regarding verbal morphology and narrative structure, it is seen that both students and native speakers of French composed their oral narratives with high incidence of PR in FG and BG and all participants produced more predicates in BG.PC was most chosen in FG within the context of narrative structure and used with similar rates by students of French.

Research Question 3
To what extent do lexical aspect and narrative structure influence the use of verbal morphology in oral narratives by Turkish learners and native speakers of French?
When analyzing participants in terms of temporo-aspectual properties, it is found that both students and native speakers of French formed their oral narratives with high frequency of PR.Fan (2005) indicated that perfective/imperfective distinction in French disappears when verbs are used in PR.In other words, although PR can express past events, it is not possible with PR to track the distinction between perfective past (PC) and imperfective past (IMP) in French.That is, there is an inversely proportional relationship between the density of choosing PR by student participants in their oral narratives and being under the influence of intrinsic semantic properties of verbs in tense choosing.We think that the disappearance of this distinction means that lexical aspect has no more influence on the use of verbal morphology.For Maingueneau (1991), PR seems to be aspectually not very restrictive.This aspectually unrestrictive character of PR also helps understanding that the influence of lexical aspect was lost.
PC was most associated with achievements and produced with similar rates by students of French.Although this gives the impression that lexical aspect continues to influence verbal morphology, the fact that PC spread to other types of verbs, especially to activities and states, shows that this effect remains only with a low level.This spread is more balanced in French participants, which means that they are not exposed to the effect of lexical aspect.
IMP was most associated with states and used with close rates by students of French.French participants used more states in IMP.Even though students of French who used the most states in IMP made think that the effect of lexical aspect continues, the fact that IMP spread to other verbs types, especially to achievements, reveals again that this effect goes on only with a low level.
We think that the fact that students of French used PR in FG and BG indicates that they are anymore under the influence of lexical aspect and they also begin to get rid of the dependence of narrative structure to determine verbal morphology.This interpretation is also in line with the results obtained from immigrant workers in the project of ESF because it was discovered, in this project, that the participants could distinguish FG from BG regardless of verbal morphology (Bhardwaj, Dietrich & Noyau 1988).Likewise, Véronique (1987) found that the use of verbal morphology in the interlanguage of African workers was determined independently of discursive grounding.
Students of French used the most PC in FG and they produced PC in FG with similar rates.This gives the impression that narrative structure influences verbal morphology within the assertions of the discourse hypothesis.However, we believe that the abundance of PC in FG is not caused by narrative structure but rather by the attitude of students of French who would like to minimize errors of language use and to be careful by avoiding descriptive expositions and by putting utterances one after the other that provide the progression in their oral narratives as if they make collages.French participants used PC in BG with high rates.It was observed that French participants did not hasten to finalize their narratives, care about descriptive expositions instead of simply employing utterances that move narrative time forward and, in this context, had frequently recourse to above mentioned presentational cleft constructions with c'était.
Results of this study showed that students of French become more exposed to the effects of intrinsic semantic properties of verbs and narrative structure towards the end of their university studies.This indicates that as students move on to upper classes, their linguistic level decreases in inverse proportion.The fact that, from the first year onwards, the number of courses which directly deal with French decreases, and there are very few courses targeting French as such, if any, especially in the fourth year, could be given as one of the reasons for which the language level of upper class students weakens.It should also be noted that, outside of the courses, students are deprived of environments and opportunities to practice French.Although they have one year of preparatory class, it must be taken into consideration that students who choose to pursue their studies in French language teaching begin with almost zero knowledge of French.Restructuring the undergraduate program so that there are courses in each year that are directly relevant to French would contribute to the resolution of such problems.
Another reason for the decrease of language level in upper class students, in particular junior and senior ones, is that they feel more intensely a concern for the professional future and consequently lose their enthusiasm and desire about foreign language learning and teaching.Indeed, almost no French teachers have been nominated in recent years.This creates ambiguities in students' professional future.In order to support linguistic diversity, increasing the number of French language courses, especially in secondary schools, the number of French teachers in teacher assignments would enhance the enthusiasm and willingness of future French teachers in upper class.
This study was carried out with students enrolling in first, second, third and fourth years of French Language Teaching Department, excluding preparatory class students from the corpus with the thought that they were not at sufficient level of language during data collection.In order to properly investigate and evaluate the fact that the linguistic level of students decreases as they progress to upper years, future studies which would aim to test the aspect and discourse hypotheses with Turkish-speaking students of French should also involve students towards the end of their preparatory year.
Data of the present study were obtained by film retell tasks.Future works that would use both oral and written film retelling as data collection tools would provide a more complementary view of the lexical and discursive properties of verbs in oral narratives of students.
At the time of data collection, tasks with strict controls were not used.This allowed participants to form their oral narratives more freely with regards to verbal morphology.In order to ideally evaluate the assertions of aspect and discourse hypotheses with respect to language level and lexical / discursive effects, it would be appropriate to design a study of mixed type, in the future, which would also involve tasks with strict controls.

Table 1 .
Distribution of Verbs in PR, PC and IMP by Participants.

Table 2 .
Distribution of PR according toLexico-Aspectual Classes and Participants.

Table 3 .
Distribution of PC according toLexico-Aspectual Classes and Participants.

Table 4 .
Distribution of IMP according toLexico-Aspectual Classes and Participants.

Table 5 .
Distribution of Foreground and Background Verbs according to Participants.When considering narrative structure in terms of tense, it is seen that PR use was also dominant in grounding.Participants used PR verbs in order to distinguish FG events from BG ones.An example about this (6) is given below.Table6gives the distribution of PR verbs in FG and BG according to participants.

Table 6 .
Distribution of PR Verbs and Grounding according to Participants.

Table 7 .
Distribution of PC Verbs and Grounding according to Participants.

Table 8 .
Distribution of IMP Verbs and Grounding according to Participants.

Table 9 .
Distribution of PR and PC/IMP Verbs in Foreground and Background.