A Narrative Perspective on The Development of The Tense System in Second Language Acquisition

SSLA, 17,263-291. Printed in the United States of America.A NARRATIVE PERSPECTIVEON THE DEVELOPMENTOF THE TENSE/ASPECTSYSTEM IN SECONDLANGUAGE ACQUISITIONKathleen Bardovi-HarligIndiana UniversityA number of studies on the acquisition of tense and aspect by learnersof a second language point to the hypothesis that narrative structureinfluences the distribution of tense/aspect forms in interlanguage. How-ever, the studies have reported conflicting profiles of tense/aspect use.This study suggests that much of the variation that has been previouslyreported stems from the level of proficiency of the learners. This cross-sectional study examines 37 written and oral narrative pairs producedin a film retell task by adult learners of English as a second language.The analysis approaches the texts from two perspectives, from theperspective of acquisition, taking narrative structure (specificallygrounding) as an environment for acquisition of tense/aspect, and fromthe perspective of the narrative itself, characterizing the foregroundand background by the tense/aspect forms used. The study finds adevelopmental pattern in the distribution of tense/aspect morphologywith respect to narrative structure. These results permit the assimilationof earlier findings into a developmental sequence in the acquisition ofthe tense/aspect system.This study investigates the development of tense and aspect in the narratives oflearners of English as a second language. Recent research has demonstrated theimportance of universal properties of narrative structure in explaining patterns oftense/aspect use among native and nonnative speakers alike. Second language stud-ies of the development of tense and aspect have, however, reported somewhat1 thank the members of the Discourse Group at Indiana University for valuable comments on earlier versionsof this paper. This research was funded by grants from the National Science Foundation, Grant DBS-19616,and Indiana University, Research and the University Graduate School.c 1995 Cambridge University Press 0272-2631/95 $7.50 + .10 2632 Kathleen Bardovi-Harligcontradictory results with regard to learners' use of tense/aspect in relation to theforeground and background elements of narrative structure. It is the aim of thepresent study to clarify the relationship between second language proficiency andthe influence of narrative structure on tense/aspect distribution.Previous work in second language acquisition that links tense/aspect to narrativestructure suggests that a relationship exists between the use of verbal morphologyand the foreground (the actual story line) and the background (the supportivematerial) of the narrative. However, these studies have reported somewhat differentdistributions of tense across grounding. Some of the variation may be attributable tothe fact that the earliest studies were case studies. Kumpf (1984) found that aJapanese learner of English used the base form of the verb to express completedaction in the foreground. Past forms in the foreground were limited to a few irregularverbs which were used only rarely. In the background many morphologicallymarked verbs occurred. Stative verbs showed tense while active verbs were markedfor habitual and continuous aspect, but were marked irregularly for tense. In con-trast, Flashner (19) found that three Russian learners of English marked fore-ground actions and left background portions unmarked. The foreground verbsoccurred predominantly in simple past with the background verbs being in predomi-nantly base forms.Veronique (1987) and Bardovi-Harlig (1992b) investigated larger groups of learn-ers. In a study of seven untutored learners of French, Veronique's findings suggestthat the distribution of verbal morphology (and lack of it) across background andforeground differs according to level, but these findings also show variation withinlevels across individuals and within individuals across texts. Intermediate learners ingeneral showed a predominant use of past markers in the foreground with either theverb stem (a morphologically unmarked form) or past (or both) in the background.One advanced learner showed the intermediate pattern of past in the foregroundand the verb stem in the background, but the other advanced learner showed amuch more targetlike system using the compound past in the foreground and theimperfect in the background of one text. In a study of 16 intermediate learners ofEnglish as a second language, Bardovi-Harlig (1992b) found that 12 learners showeda greater use of past in the foreground than in the background. Nine learners showedboth a robustly greater use of past in the foreground and greater use of nonpast inthe background (over 50% greater), and three additional learners showed 50%greater use of past in the foreground than in the background with no appreciabledifference in nonpast.Level of proficiency clearly emerges as a likely factor in the distribution of tenserelative to grounding, especially when one takes into account that very low levellearners show no systematic use of tense (Schumann, 1987) and that advancedlearners must eventually use past in both foreground and background to reach atargetlike use of tense in English narratives. The exact role of level of proficiencyhas been difficult to identify, however, because of the relatively small number oflearners whose narratives have been investigated and because of the different crite-ria that have been used to determine levels. Veronique grouped his learners accord-ing to the conversational criteria of turntaking, leading the conversation, holding thefloor, and average length of response. Bardovi-Harlig's intermediate learners wereA Narrative Perspective on Tense/Aspect in SLA 265identified by instructional level in an intensive English program. Both subject popula-tions showed broad within-level variation with respect to appropriate use of tense.This study attempts to address the question of the relation of acquisition to theinfluence of narrative structure through a cross-sectional study in which learners aregrouped according to the development of their tense/aspect systems. Groupinglearners according to appropriate use of tense eliminates less relevant variables suchas conversational skills (Veronique, 1987) and other linguistic and academic skillsthat figure in instructional placement (Bardovi-Harlig, 1992b) and thus facilitates thecomparison of learners on the single relevant variable of development of tense. In alongitudinal study of the acquisition of the tense/aspect system, the rate of appro-priate use of simple past tense proved to be a predictor of the emergence of otherforms (Bardovi-Harlig, 1994b), indicating that this is a promising approach to theinterpretation of level.NARRATIVE STRUCTUREIn linguistic studies (as opposed to literary studies), a narrative is considered to be atext in which \"the speaker relates a series of real or fictive events in the order inwhich they took place\" (Dahl, 1984, p. 116). Narrative discourse is comprised of twoparts: the foreground and the background. The foreground relates events belongingto the skeletal structure of the discourse (Hopper, 1979) and consists of clauses thatmove time forward (Dry, 1981, 1983). The temporal point of reference of any oneevent in the foreground is understood as following that of the event preceding it. Soimportant is the concept of sequentiality that foreground clauses may be defined bythe interpretation of their order: \"If a change in the order of the two clauses resultsin a change in the interpretation of what actually happened, then those two clausesare narrative [i.e., foreground] clauses\" (Schiffrin, 1981, p. 47; see also Labov, 1972;Labov&Waletzky, 1967).1Examples (1) and (2) show a sequence of events reported as foreground innarratives by learners of English as a second language. In the examples from thewritten narratives the spelling is that of the learners. The numbers in the squarebrackets were added and represent the order of the foregrounded events.1. Ora1, LI JapaneseForeground[1]Then she, stole the bread.[2]And the- she ran away,[3]and she... hit the ChaplinI:Mm-hm.BackgroundS: ... and uh, different woman sawthe, she stole the, uh, bread...I: Mm-hm.[4][5][6][7]... and uh, and cried,and chased her,and the employer caught her,and, but Chaplin said,\"I did it, you know, I... stole.. .the breads.\"266 Kathleen Bardovi-Harlig2. Written, LI ArabicForeground Background[ 1 ] she stol abread[2] and they cutch her[3] she me/charliethe nice man who was trying to helpe her[4] when he said \"she didn't sleel thebread I did that\"The central characteristics of the foreground can be summarized by what Rein-hart calls temporal criteria (1984, p. 801):Temporal Criteria1. \"Narrativity,\" or temporal continuity:Only narrative units, i.e. textual units whose order matches the order ofthe events they report, can serve as foreground.2. Punctuality:Units reporting punctual events can serve more easily as foreground thanunits reporting durative, repetitive, or habitual events.3. Completeness:A report of a completed event can serve more easily as foreground thana report of an ongoing event.In addition to chronological order (Reinhart's \"narrativity\"), Dry (1983) claimsthat a second textual criterion for evaluating time movement is information value;the information communicated in the foregrounded clause must be new rather thangiven. In Example (3), clause (c) is foregrounded, whereas in Example (4), (d) isnot.3. (a) John gave Mary an apple, (b) and she sat down to take a bite.(c) She took the bite deliberately, savoring the taste, (p. 34)4. (a) John gave Mary an apple (b) She sat down (c) and took a bite.(d) She took the bite deliberately, savoring the taste, (p. 34)In Example (3) she took the bite deliberately presents ordered, new informationand is foregrounded, whereas in Example (4) the clause elaborates informationalready presented in (c).In contrast to the foreground, the background elaborates or evaluates the eventsin the foreground (Hopper, 1979). Although events reported in foreground clausesare understood to be sequential, background events are often out of sequence withrespect to the foreground and to other background events. The background doesnot itself narrate main events but provides supportive material that elaborates on orevaluates the events in the foreground. For example, a background clause maycontribute to the interpretation of an event by revealing a prior event (locatedbefore the narrated event on the time line), making a prediction about the outcomeof an event (located after the event on the time line), or evaluating an actionreported in the foreground (not located on the time line). Background may provideorientation (scene setting) (Example 5), evaluation (Example 5), or explanation/identification (Example 6).A Narrative Perspective on Tense/Aspect in SLA5. Oral, LI SpanishForegroundOrientation267BackgroundAnd then, in another situation, inanother place, there is a old ladythat is really hungry is starving-She is starving. She is starving andso a- a business of bread, loaf ofbread, [..][...] to support it.ForegroundEvaluation6. Written, LI Arabic11 ] And she decided to hope totake out the the loaf of bread.And although this was a wrongthing.ForegroundBackgroundThis movie talk about the man andthe women. One day the womenwas too hungre,OrientationAction[ 1 ] so she stole a big piece ofbread [2] but when she wasranning [3] she hit the manwho is Charlie Chaplin,Although the manwho is a Baker[4] he said to the policmanthat there is women who stolehis bread, [5] so the policman goafter the women.IdentificationIdentificationAn event that precedes the foreground is reported in the background in Example(1), and a simultaneous event is reported in the background in Example (2). Incontrast to the foreground, the background has many individual functions that to-gether serve the greater function of supporting the foreground.Crosslinguistic investigations suggest that the distinction between backgroundand foreground may be a universal of narrative discourse (Hopper, 1979; see alsoLongacre, 1981). Hopper observes that competent (native) users of a language\"mark out a main route through the narrative and divert in some way those parts ofthe narrative which are not strictly relevant to this route\" (1979, p. 239). One suchmarking may be the use of tense and aspect, and other markers include word orderand voice (Hopper, 1979). Hopper observes that \"one typically finds an aspectmarker specialized for foregrounding, or one specialized for backgrounding, or bothfunctions indicated\" (1979, p. 239). Tense markers may occur with aspectual mark-ers. Because of its function, the background may exhibit a variety of tenses such aspluperfect, remote past, future-perfect, and future which do not occur in the fore-ground. In the foreground, Hopper observes, successive events may be marked inthe preterite or simple past (1979, p. 239). Dahl observes that in some languagesverbs in the foreground may carry no marking, concluding that \"it is always possibleto use the least marked indicative form in a narrative [i.e., foreground] past context\"(1984, p. 117).In her examination of foreground and background in contemporary English liter-ary narratives, Dry (1981, 1983) reports that foreground clauses are usually in the268 Kathleen Bardovi-Harligsimple past or historical present. As in the foreground, the simple past is also themost prevalent verb form in the background clauses. However, English does not relyprimarily on tense or aspect markers to distinguish foreground from background.Instead, characteristics tend to cluster as shown by Hopper and Thompson (1980),Reinhart (1984), and Fleischman (1985).METHODTo examine the distribution of tense/aspect morphology in interlanguage, the pres-ent study addressed three questions:1. Does past emerge differentially in the foreground and the background?2. What tense/aspect forms characterize foreground and background?3. How do the dominant forms of foreground and background change as the tense/aspectsystem is acquired?A cross-sectional sample of oral and written narratives from 37 learners of Englishas a second language form the corpus for this study.The narratives were elicited by means of a film retell task. Many studies of firstand second language narratives have elicited narratives through the retelling ofsilent films (e.g., the well-known Pear Stories [Chafe, 1980], the European ScienceFoundation investigation [Klein & Perdue, 1992; Tomlin, 1984]), performed stories(Bardovi-Harlig, 1992b), and picture stories (Bamberg, 1987; Bamberg & Marchman,1990). As Chafe (1980) observed, many studies have found it \"useful to collectexamples of different people talking about the same thing ... in order to see whatsimilarities and differences emerged between different verbalizations of what was,at least to a large extent, the same knowledge\" (p. xii). Although topically similarpersonal narratives such as those collected by Labov and Waletzky (1967) andTannen (19) address in part the advantage of having different people talk aboutthe same thing, personal narratives still differ in knowledge and in experience. Inaddition, for the investigation of narrative from the perspective of tense/aspect,personal narratives also lack the kind of external reference that retell tasks canprovide. Retell tasks provide narratives in which the sequence of events is known tothe researcher independently of the narratives themselves. Thus, comparabilityacross speakers and independent verifiability of the story were key factors in thedecision to use elicited narratives in this study.2MaterialsThe particular film selected for this study was an 8-minute excerpt from the silentfilm Modern Times.3 A silent film was selected to avoid complications of listeningcomprehension (see Bardovi-Harlig, 1994a, for further discussion). The particu-lar excerpt was chosen because there were a series of discrete, easily identifiableaction sequences as well as some simultaneous action (ideal for examining theencoding of tense/aspect morphology) and changes of scene (ideal for examiningbackgrounding).A Narrative Perspective on Tense/Aspect in SLA 269In the film excerpt, Charlie Chaplin befriends a young woman who has stolen aloaf of bread and at the same time tries to ensure his return to jail where he haspreviously found food and shelter. The segment has four distinct foreground epi-sodes (which may be further divided): (on the street) the stealing of the bread bythe girl and the events leading to her arrest, (in the cafeteria) Chaplin's eating of twotrays of food without paying and the events leading to his arrest, (in the policewagon) the meeting of Chaplin and the young woman, and (on the street) theensuing escape. The excerpt ends in an imagined sequence that depicts Chaplin andthe woman in a blissful domestic scene; at the conclusion of their daydream, Chaplinresolves to obtain the house of their dreams. The excerpt has 10 titles; the first onesets the scene, \"Alone and hungry,\" and nine report the speech of the characters.4ProcedureThe learners watched the 8-minute excerpt, which was played twice. Learners weretold that they would be asked to tell the story after they had seen the film. Afterviewing was completed, participants met individually with an interviewer to recordthe story. All participants were tested orally within 30 minutes of viewing the film.Following the oral narratives, learners produced written narratives during theircomposition classes that met later in the day. Learners were given the class period,or approximately 45-50 minutes, to produce their written narratives.Selection of ParticipantsAll learners were enrolled in the Intensive English Program (IEP), Center for EnglishLanguage Training at Indiana University. One listening and speaking class from eachof the six instructional levels of the IEP was selected to participate in the study. Oraland written samples were collected from 51 learners.A subset of 37 pairs (74 narratives in all) was selected on the basis of usablelanguage samples and rate of appropriate use of past tense. The first step excludedall nonnarrative texts such as movie reviews and answers to questions. The nextstep grouped learners by overall rate of past tense use. Each narrative was coded foruse of a past tense form in past-time contexts, which included simple past, pastprogressive, and, more rarely, pluperfect. Rates of past use were calculated for verbtypes rather than tokens; that is, each verb form was counted only once per sampleregardless of frequency. In this way frequency rates were not inflated by multipleoccurrences of common verbs such as was and went (see Bardovi-Harlig, 1992c,1994b, for discussion).The learners in this study are compared on the basis of their appropriate use ofpast morphology rather than on the basis of their proficiency level as determined bytheir placement in the instructional program because learners show a considerablerange of appropriate use of tense/aspect when considered individually (Bardovi-Harlig, 1994b).5 Moreover, describing interlanguage in terms of the rate of use ofverbal morphology facilitates comparison of learners across studies (see Robison,1990; Schumann, 1987).270 Kathleen Bardovi-HarligTable 1. Subject distribution% AppropriateUse of Past102030405060708090TotalWrittenn21233497637Subjects1-234-56--1112-1516-2425-3132-37n36665137OralSubjects1,3,86,9,12,15,16,282,5,7,21,30,334,10,11,13,20,2923-25,27,3117,19,32,34,35,3714,18,22,2636Learners were grouped according to the percentage of appropriate use of past bydivision of 10% (10-19%, 20-29%, 30-39%, etc.). Subjects were ranked separatelyfor written and oral texts. Fifteen learners scored between 10% and 69% appropriateuse of past on the written narratives and all were included in the study. Theselearners were placed in four groups numbering between three and six members. Athigher rates of appropriate use, there were more learners to choose from, andsubjects who showed 70-100% rates of appropriate use were therefore selectedon the basis of LI background. Three groups of higher proficiency learners wereconstituted, each numbering between six and nine subjects.6 This procedure resultedin the selection of 37 narrative pairs, as shown in Table 1. The learners representedfive LI backgrounds: Arabic (14), Korean (10), Japanese (6), Spanish (6), and Man-darin (1), distributed evenly across groups.Two idiosyncrasies in these groupings merit further explanation. In the writtennarrative, only five learners exhibited 13-33% appropriate use. These learners weregrouped together as W10-30. In contrast, the one learner who showed over 80%appropriate use in the oral narrative was not included with O70 because he showed.5% appropriate use whereas that group showed a range of only 70.0-75.0%(Table 1). Complete subject profiles can be found in the Appendix for the writtenand oral narratives.ANALYSISAs outlined previously, the first analysis of the narratives scored each verb in apast-time context as \"past\" or \"nonpast.\" The number of distinct verb forms werescored for the type score and the number of total verbs used (tokens) was alsoscored. Direct speech was excluded from both calculations. These calculations wereused to identify the participants and to place the learners into groups as outlined inthe preceding section (Table 1).The narratives were then coded for grounding. Following Dry (1981, 1983) andA Narrative Perspective on Tense/Aspect in SLA 271Dowty's (1986) discussion of sequencing, clauses that moved the narrative timeforward were identified as foreground clauses (see the Narrative Structure section,earlier). Determination of grounding was made independently of verbal morphol-ogy. Quoted speech was not counted as a foregrounded action itself, although theverb that introduces it, if sequenced, was treated as foreground (Dry, 1981, 1983;Labov, 1972; Schiffrin, 1981 ).7 Grounding analysis was performed on all 74 texts bythe researcher and a second experienced coder. Interrater reliability on the writtennarratives was 98.1% (agreement on 1,446/1,474 coding decisions) and on the oralnarratives 98.4% (agreement on 1,6/1,927 coding decisions). Disagreements wereresolved by discussion.Next, all verbs were coded for verbal morphology (simple past, past progressive,pluperfect, present, base, 0 + progressive, present perfect, and miscellaneousforms). Some special cases should be noted. Uninterpretable forms such as tooks andis stole were coded as \"other.\" Propositions that require verbs but lacked them werecoded as \"no verb\" (i.e., she hungry) and were counted as either foreground orbackground as appropriate. Verbs that have the same form for past and base wereexcluded from the sample (e.g., hit, put, let; cf. Bardovi-Harlig, 1992b, 1992c; Silva-Corvalan, 1983). In the written narratives, misspelled verbs such as cot/caut forcaught or regularized past verbs such as telled were counted as past as long as theinnovation did not result in an extant verb. It should be noted here that, in contrastto the analysis used to group the learners, this analysis is a token analysis. Althoughtype analysis controls for multiple uses of a single form in a corpus, it does notrespect the integrity of the text and thus cannot be used to analyze the structure ofnarratives.The oral narratives presented additional problems for analysis. One such problemwas the repetition of verbs. When a learner repeated a verb exactly, it was countedonly once so as not to inflate the number of propositions in the narrative (Example 7).7. Then, the police, policeman, take him, take him to... [LI Korean]In cases where a verb form was not an exact repetition, as in Examples (8) and (9),the ratio of forms was calculated. In Example (8), for instance, past and base eachscored 0.5; the total number of verbs was 1.0.8. while, uh, Charlie Chaplin go to, went to restaurant [LI Arabic]9. He said, he's say... [LI Arabic]This gives priority to neither the first form that the learner produced nor the cor-rected form (which is not always \"better\"), but factors in all forms produced by thelearners without inflating the numbers of propositions.In addition, phonological considerations in the oral narratives necessitated theexclusion of regular verbs in certain environments. Tokens of nonsyllabic past tensefollowed by homorganic stops were excluded from the sample (e.g., walked to/down, pulled down/to), as were verbs that ended in consonant clusters followed byinterdental fricatives (e.g., walked through/the, entered the; Bayley, 1994; Wolfram,1985).8272 Kathleen Bardovi-HarligIn addition, because the oral narratives were interactive, there were interviewerrequests for clarification and elaboration. Elaborations were generally requested atthe end of the narrative sequence. Answers to interviewers' requests for elaborationwere not considered to be part of the main narrative and were not coded foranalysis.RESULTSThis section is divided into three parts. The first part compares the use of tense inthe foreground and background. The second and third parts examine tense usewithin the foreground and background separately. It begins with a general descrip-tion of the narratives.The narratives varied in length (15-77 verbs in written, 11-125 verbs in oral)and in degree of detail. Foreground verbs outnumbered background verbs in bothwritten narratives (928 to 456) and oral narratives (1,002 to 2). This is consistentwith Tomlin's (1984) findings for on-line narration. As a group the learners showedlower use of past tense in the oral narratives. Consequently, the oral narrativesprovide a fuller view of tense marking at lower levels of tense use, whereas thewritten narratives provide a fuller view of tense marking at the higher levels of tenseuse (see the Selection of Participants section, earlier).Tense Marking in Foreground and BackgroundTables 2 and 3 show the distribution of verbal morphology with respect to groundingby group. The superordinate categories of Past, Nonpast, No Verb, and \"Other\" aregiven in boldface. The categories of Past and Nonpast are further divided intospecific tense/aspect forms: Past (simple past, past progressive, pluperfect) andNonpast (base, present, present progressive, progressive [e.g., 0-eating, 0-walking],and [present] perfect).9Reading Tables 2 and 3 across the row for \"Past\" we see that the simple pastemerges first in foreground. The simple past scores are higher in foreground than inbackground in the lowest three groups (the combined W10-30, W40, and W50) inthe written narratives, and the simple past is again used relatively more often inforeground than background by groups W70, W80, and W90 (Figure 1). Similarly, inthe oral narratives from group O20 on, the use of simple past is greater in foregroundthan background in past time contexts (Figure 2).10In group W60 the use of simple past in the background catches up to use in theforeground at 59.1 and 59.0%, respectively (Table 2). By the time that learners areable to mark 60% of past-time contexts appropriately for tense, there is a qualitativedifference in the forms that are used in lieu of the simple past. These are discussedin the Background section.If we understand the foreground and the background of a narrative as environ-A Narrative Perspective on Tense/Aspect in SLA27310090-8070-60-Q_ForegroundBackground\"5to50-40-30-20-10-0-10-30 40 50 60Group7090Figure 1. The distribution of simple past by grounding (written narratives).ments in which tense/aspect morphology emerges, we see that the use of simplepast emerges first and more strongly in foreground. As Figures 1 and 2 show, theuse of past in the foreground exceeds its use in the background in both written andoral conditions.ForegroundComparing the foreground and background for the emergence of tense/aspect ap-proaches the text with an acquisitional question \"where does past (first) appear?\"and views grounding as an environment for acquisition. A second view, althoughstill acquisitional, approaches the text from the perspective of the narrative andasks \"what (verb) forms characterize the foreground and background?\"—this is thequestion that LI studies have asked—and at what point?—this is an acquisitionalquestion. The use of past in the foreground exceeds the use of the past in thebackground for the first three groups in the written narratives (read across the\"Foreground\" and \"Background\" columns in Table 2). However, a comparison ofverbal morphology used in the foreground (read down the \"Foreground\" columns inTable 2) shows that in the first group, WlO-30, the use of base exceeds the use ofsimple past (50.9% base to 30.4% simple past). Thus, in WlO-30 past cannot be saidto characterize the foreground. This begins to change in W40, and in W50 past is thedominant form with 57.1% use compared to the base with 34.7% use. The gapcontinues to widen in the higher groups (Figure 3).In the oral narratives, the use of simple past also pulls ahead the use of nonpastin Group O40 (Table 3) at 49.4% simple past to 39.5% base. As Figure 4 shows, baseck7694.6)a.B202.94.716.25.9.2690= Ne77(r01925o.2..21.0..6F74.03991ck5343.20111.07)aB28.9.4.1.87513.113.2.36.980= Ne(r57826638o.3..F06.87.0.1.5981ck80880887...99)a85.49.20.920.9..230B6513.211.170= Ne(r734737o..485F86.2.6..477116.0.032ck91141.)a46.8..43..p594u B59.65211.1124o60-=Gr-Ne(r10763o...65F19.2.121.1.165339puorgck0000555...5..3)a0 yB00.05.02.5.02710.41714b s50= eNev(r110770000io..taF97.2.71.1..2855334.2.9rran nck00318433.....e)a 3B22.34636.2.7712224rittw40= Ne884026...n(ro88.1.21..348 iF4444ygoolhck7291..739)p0.1.8..r05a11039.876723.238.5o3 -=Bm0-- 1Ne44196f(roo.778..2 F00.80.2.2.1.11335511noutiebviriesvstiestsrcDigeesoef rrergt.ocPvie 2tt Pbsertr lnnessbeepP erftrntpaseeegererV bltupimaslasesesoesaTasSPPnoBrrrrVelaPPPP ohttPNNOoT2k22844c...)aB913 1669.0315.1580=Ne0(r0.0o..00.F088002210ck785...97..2.6)a47486.011.6812315.175 B70=Ne(r713o...39F27.671.0875.1169ck569.05.22186)aB76.7551.315.317.2.2.1.3.2960= Ne36(r0o..9952.1.F995.2.0.81.10.66622188ck6098.15.5)aB74.3.56.5.72.42.4.74421850= Ner4.91.16(o61.3.08.1.60.60.62.F5531puorGck2.12.499416)a..138.8B32.4.9.434423.142.130= Ne(r355o.49.26F09.0.59.2.51.41.470.4.1432puck2238.8....33oar6)gB40.43.85713.86.1.39335227 30= Ne by(r118so..1..38eF760.4.0448440.7149vitarrack4.784000n 6)a91.5.B2..117.986357.53.6.6.726alr20= o Ve268(Jro.0.2.8 inF067.0.73.31.32.3306.152ygoolhck303p)ar3B8.8.3.7..3.08.3138123om10 = Ne761f(r33..1991.o4.4.587.2.2.75 oF7417.13ionebutviriesvsstietsrcDisgeeofe rrsgtPveriP.ocesbt ts trert 3lPnnebeep erfsltpaeeernrssgeserV tmsupseeoeabsialrrrrVelaTaSPPnaoBPPPP hottPNNOoTlOto276Kathleen Bardovi-Harlig100Background20 3040 50 60Group70 80Figure 2. The distribution of simple past by grounding (oral narratives).100o1b10-30 40 50 60Group80 90Figure 3. The distribution of past and base in foreground (written narratives).A Narrative Perspective on Tense/Aspect in SLA27710 20 3040 50 60Group70 80Figure 4. The distribution of past and base in foreground (oral narratives).forms retreat in the foreground after Group 40, although somewhat more graduallythan in the written narratives (cf. Figures 3 and 4).This series of stages can be illustrated by examining individual learners. Forexample, Learner 4 (WlO-30, LI Spanish) shows 34.5% use of past in the foregroundand 11.1% use in the background. Although his use of past is three times morefrequent in the foreground, and we can say that past emerges first in the foreground,past cannot be said to characterize the foreground where 55.2% of the forms arenonpast. In contrast, Learner 9 (W40, LI Arabic) shows 66.7% use of past and 33.3%nonpast in the foreground but in the background shows 25.0% use of past and 58.3%nonpast in the background. Thus, he shows both greater use of past in foregroundthan in background and a dominant use of past in the foreground. Learner 17 (W40,LI Japanese) provides an example of a learner who shows dominant use of past inthe foreground (79.2% past, 20.8% nonpast) where the background shows equal useof past and nonpast (50.0% each). Learner 31 (W90, LI Arabic) shows dominant useof past in both foreground and background with 95.7% use of past in the foregroundand 76.5% in the background.The competition between forms is important from the perspective of narrative.In the foreground the essential competition between forms is between the simplepast and base (or unmarked) forms. This is true in both written and oral narratives.However, the interlanguage background, like the target language, shows greaterdiversity, as discussed in the following section.278100Kathleen Bardovi-Harlig10-30 40 50 60 70 80 90GroupFigure 5. The distribution of past and nonpast in background (written narratives).BackgroundThe use of past emerges as the dominant tense somewhat later in the backgroundthan in the foreground. In the written narratives, it is not until W60 that past clearlypulls ahead of nonpast (Figure 5). The same pattern is found in the oral narratives,where simple past becomes the most used tense/aspect in O60 (Figure 6). Unlikethe foreground, however, where learners primarily use two verb markers, the simplepast, and the base or 0-marker, in the background learners use three: the simplepast, base, and simple present. The use of present in the background comes primarilyfrom the use of the copula. Calculating the total number of verbs in the simplepresent in the oral narratives yields 105 verb tokens; 82, or 78.1%, of those were thecopula. In the written corpus, of the 71 verb tokens in simple present (76.1%)were the copula. (In contrast, only 39.4% [80/203] of the background verbs in thepast are due to the copula in the oral narratives, with 46.4% [ 103/222 ] in the writtennarratives).\" Be in its base form appears only rarely in this corpus, perhaps becauseof its relative infrequency in input. The complementary use of simple present andbase is represented in Figures 5 and 6 where the Nonpast bars show a combineddistribution of base (the shaded portion of the bar) and simple present (the unshadedportion). Also in the nonpast category are uses of present progressive, 0-progressive,and (present) perfect, which, although relatively rare, occur more frequently in thebackground than the foreground.I2 These forms are not represented in Figures 5and 6, but their distribution is given in Tables 2 and 3.The background also sees the emergence of the past progressive in both writtenand oral narratives. Because the tense that has been established for the narratives isA Narrative Perspective on Tense/Aspect in SLA2791009080Past • Base ^ Present10 20 30 40 50 60 70 80GroupFigure 6. The distribution of past and nonpast in background (oral narratives).past (recall that past is the dominant tense for background for W/O60 and higher),the use of 0-progressive and present progressive may be interpreted as attempts atpast progressive according to the grammatical aspect encoded. The pluperfectemerges late, as expected (Bardovi-Harlig, 1994b), and exclusively in the writtenbackground. Increased use of pluperfect by even more advanced learners in thisstudy would be expected to cut into the use of simple past.The use of progressive by learners in the written narratives is similar to the useof progressive by native speakers who completed the same retell task as the learners(Reynolds, 1994). Reynolds found that the progressive was used by native speakersexclusively in the background. Progressives constituted 12.6% (21/167) of all back-ground verbs in written native speaker narratives. Relatively similar use is found inthe background by W70 (13.8%), W80 (13.3%), and W90 (16.2%) and by O70(12.9%). The use of progressive in the foreground is rare, under 4% by the samelearner groups (2.4, 3.8 and 2.9%, respectively).Some of the propositions lacked a main verb. These are indicated on Tables 2and 3 as \"No Verb.\" Except in the case of Group 10 in oral narratives where 33.3%(4/12) of the propositions expressed in the background lacked a verb, missing verbsare rather rare. However, out of the 42 occurrences in oral and written narrativescombined, 38 (or 90.5%) of all such cases of no verb occur in the background, as inExamples (10) and (11).10. O10, Learner 1, LI KoreanForegroundMan eat, many many food.BackgroundBut... no money28011. W30, Learner 5, LI ArabicForegroundand he tell to Policeand the Police cought himKathleen Bardovi-HarligBackgroundSo Charlie Chaplin a generlemanIn contrast to the foreground, which exhibits simple past and base almost exclu-sively, the background shows diversity in tense/aspect. Simple past competes withbase and present tense (including simple present and present progressive) in thebackground in the earliest stages of tense/aspect development. Propositions withoutverbs are also found in the background. In later stages of development, past progres-sive and pluperfect emerge in the background and occur with simple past as innative speaker narratives. In the intermediate stages of development, the foregroundmay show fairly consistent use of past tense while the background shows a range ofverbal morphology, as Example (12) shows.12. W70, Learner 17, Japanese LIForeground[1] First the woman was watchin inthe bread shop.[ 2 ] She stole a lofe of bread[3] and run away.14 ] and said to shopes owner.'The woman took your bread.\"[ 5 ] He fry to arrest her.[6] She was runningand crashed to masuturshe man.[mustache man]When they crashed.[8] shopes owner came to their place.[9] A policman came there too.[10] The shopes owner said, \"she stolemy bread.\"[11] The man said, \"she didn't steal it. Idid it.\"He is masuturshe man.[12] The policman arrest him.But the woman sow [saw] all of things[13] and she said \"she didn't steal abread she did it.\"[14] The policman arrest woman.The masuturshe man try to help of her.[15] He ate many food in shop[16] and he didn't pay many.13[17] The policman arrest him too.[18] Then policman took him into thetruck.[19] After took the seat into the truck[20] the woman came into truck.[21] They met in the truck.She wanted run away from the polic.BackgroundThe other woman has lookedA Narrative Perspective on Tense/Aspect in SLA 281[22] Suddenly the truck rocked.[23] they droped on the road.[24] They run away together.[25] They sow 1 couple:The couple looks so happy.[26] The masuturshe man said, \"Let'simagine, When we get a house.\"[27] \"she said I'm hunglly.\"They were imagineing about meal,[28] finishThe learner's use of past in the foreground is 79.2% and in the background 50.0%.For the simple past alone, he shows 70.8% use in the foreground and 37.5% use inthe background (see Tables 2 and 3 for group means). The use of past can be said tocharacterize the foreground because it is the dominant form, but diversity character-izes the background where the learner shows use of base, present, present perfect,14simple past, and past progressive. The background performs a variety of functionsincluding the reporting of prior action (the woman has looked), description (he ismustache man, the couple looks happy), and explanation (she wanted to run awayfrom the police).Imagination Sequence, in contrast to the isolated background clauses whichare interspersed throughout the narrative, the imagination sequence presents anextended portion of backgrounded text. In this episode the poverty and homeless-ness of the protagonists is contrasted to what their lives might be like in middle-classsuburbia. A work-clothes-clad Chaplin and a pinafore-clad wife in a tidy houseprovides the backdrop for a scene in which the conveniences of the middle class areexaggerated.In the him excerpt, the imagination sequence follows the escape sequence and isintroduced by the title \"Can you imagine us in a little home like that?\" Although itfollows the escape episode, it cannot be interpreted as being on the same time axisas the main narrative, a requirement of foregrounding; as Dry (1983) states,\"Foreground events are presented as actually occurring in the narrative world, asopposed to being merely talked of, expected, or hypothesized\" (p. 21).Just under half of the narratives (18 written narratives and 18 oral narratives)devote one or more finite background clauses to the dream sequence, with 11written and 9 oral narratives elaborating the imagined scene in three or moreclauses. The accounts of the imagined scene are essentially descriptive, althoughfour of the narratives (one written and three oral) included an embedded narrativeof the dream sequence.In general, the use of nonpast in the imagination sequence is higher than in thebackground as a whole for all subjects in the written narratives where 51.6% use ofnonpast in the dream sequence contrasts with only 27.6% use of nonpast in thebackground as a whole (excluding the imagination sequence). The oral narrativesshow approximately the same use in the dream (49.3%) as in the background overall(44.3%).282 Kathleen Bardovi-HarligSome learners distinguish the imagination episode chiefly by tense, as in Example(13). The writer of Example (13) shows 87.5% use of past in the background (exclu-sive of the dream) but used nonpast exclusively in the dream. The writer also usedparentheses to set the embedded narrative apart from the text.13. W80, Learner 30, LI SpanishForeground Background[14] They sat down green grass[15] and saw a familyThey seemed to very happy happy, chaplinand the woman envied the family[16] and imagined future of them(chaplin is the woman's husbund. It is deal.There are many food in their home. Theylook like very happy.)[17] but the polise man came.The imagination sequence may be marked in ways other than tense, as well. For amore proficient learner, a change in tense may be an unsatisfactory marker becausethe sequence, although imagined, was completed in the film. This raises a conflictbetween using tense to mark the sequence to distinguish it from the foreground—oreven the rest of the background—(which suggests the use of nonpast) and reportingthe fact that the imagination sequence has ended (which suggests the use of thepast). For a less proficient learner who only rarely marks tense, a tense contrastcannot be implemented. In Example (14) a learner delimits the imagination se-quence from the rest of the narrative by explicitly marking the beginning, they areto daydream, and the end, they wake up.14. O40, Learner 11, LI ArabicForeground BackgroundThen they are to daydream,[I:Mm-hm.]Where he lives in this, type of house, beautifulhouse. And, she cook for him ... [I: Mm-hm.]The wife is coming back from his job. [I: Mm-hm.] [laughs] [////]The- the girl put in the- dinner of rice... [I:Mm-hm.] [////]So, he took the, [////], food ... [I: Mm-hm.]Then he call the cow, to, to squeeze milk, heget the milk from [////]And they wake up from their dream-The boundaries of the dream are explicitly marked at the beginning and the end by11/18 written narratives and 11/18 oral narratives. The explicit boundary markersoccur at all levels, as the following examples show. The numbers in parentheses givethe number of intervening verbs.15. the beginning to dream (2 verbs)after the dream [O20, Learnerl2, LI Spanish]A Narrative Perspective on Tense/Aspect in SLA 28316. And chaplin saw the little house life (6 verbs)they imagined this... [O60, Learner 32, LI Japanese]17. they just imagined to live there (2 verbs)when they realized [W90, Learner 35, LI Korean]18. talked to each other about their dream (5 verbs)However, their happy time did not went so long time [W90, Learner 36, LI Japanese]Distinguishing the dream sequence from the rest of the narrative seems to beparticularly important when the account of the dream itself contains a narrativesequence. In Example (19), Learner 26 offers a lengthy account of the dream (17verbs) which is bounded at the beginning with in the imagine and at the end withthe imagine over. In fact, just before the first clause of the embedded foregroundclause he came (line 3), the storyteller cautions the listener again about the upcom-ing sequence that it's in imagine. (Verbs marked with an asterisk were excludedfrom calculations due to the phonological environment; tripped was supplied by theinterviewer and was not counted.)19. O70, Learner 26, LI Japanese[1] Um... they were in the, in the imagine the, the, uh, they having, uh... they arehaving good time, in the house. [I: Mm-hm.][2] They were doing- They [laughs] [////] something. Uh, yeah. Uh...[3] and.. .it's in - imagine. He came, to- came in the house, and he- he-[4] 'tried to hug her, but he - how do you say -he... [I: Uh, tripped?][5] He tripped.. .[I: Mm-hm.][6] .. .and uh, whoah... I think there was, dinner time - yeah. And, h-[7] Chaplin, pick - orange? - was some fruit. [I: Mm-hm.][8] And, he was... eating, fruit.[9] And, I think he didn'Mike it.[10] He *kick the, orange, outside, yeah. And... in the kitchen, uh, they have,[11] they are h- they are having, they were having - dinner, at there, and[12] uh, Chaplin realiz-ed there is no milk... [ I: Mm-hm. ][13] And, uh, he *call the cow. And uh, the cow - I can't explain, oh UmAnyway, uh... uh... mm [laughs] [I: What did the cow do?][14] Uh, well, I am trying to say: Chaplin, uh, received milk... [I: Mm-hm.][15] ... from the cow. And, uh, milk was... fr-... milk was fall in the - glass or some-thing? [I: Mm-hm.] Pot? [I: In the pitcher.][16] And.. .and... then, they are eating same steak, same time ... [I: Mm-hm.][17] They are cutting their meat.Seems like, I think that the, [imagine] over. [I: Mm-hm.]And, then, I think cop was behind of them. [I: Mm-hm.]I think that's all. That's it.Unlike the narrative in Example (13) in which the learner delimits the dream by useof nonpast (i.e., present and base), the narrative in Example (19) does not use tensealone to mark the sequence. In fact, Learner 26 shows no use of simple present andonly one instance of a base form but shows a very high use of progressive, 46.1%,compared to her use of progressive elsewhere in the background of 19.4% of theverbs. The use of progressive in the dream is split between past progressive (19.2%)and present progressive (26.9%). Present progressive occurs nowhere else in the284 Kathleen Bardovi-Harligbackground. Thus, the dream sequence is set off by two markers, high use of pro-gressive, and explicit boundary markers.Examining the accounts of the dream which provide extended background textemphasizes the fact that it is a greater challenge to make associations betweentense/aspect forms and their use in the background than in the foreground.Summary. The background is the site of tense/aspect diversity, not only innative speaker narratives but also in learner narratives. As an environment for theuse of simple past, background lags behind the foreground. The simple past becomesthe dominant tense in the background later than in the foreground but never reachesthe same high level of use because other past tense forms such as the past progres-sive and perfect emerge and are used appropriately in the background. Even takingall past forms into account, the use of past in the background still lags behind the useof the foreground until W90, where the use of past is over 90% appropriate use forboth. (The background use of past does not catch up to the foreground in the oralnarratives whose highest group is lower than the written group.) The imaginationsequence emphasizes the complexity of the background and the difficulty learnershave in making form-use associations in that environment.DISCUSSIONAnalyzing tense use from the perspective of the narrative shows that tense use ininterlanguage is influenced by narrative structure. The foreground is the environ-ment in which the simple past first appears with lexical verbs (i.e., noncopularverbs). The foreground shows substantially greater rates of use of simple past thanbackground in both oral and written narratives and across levels.Although the use of simple past does not characterize the foreground at firstoccurrence, simple past does come to be the emerging dominant form in the fore-ground by the time that learners exhibit 40% past in past-time contexts (Group 40)and strongly dominant at 50% use (Group 50) in both modalities. In the backgroundit takes somewhat longer for the past to overtake the nonpast.Looking at the relation of tense and grounding in these two ways may explainsome of the differences in the reports of tense distribution in previous studies. Inaddition, there is a clear effect of proficiency across the groups. As the figures in thepreceding section show, the profile of grounding changes as learners become moreproficient with respect to their tense/aspect systems. Thus, proficiency also accountsfor some of the differences reported in the literature. For example, Kumpf's (1984)learner seems to be a relatively low-level learner whose use of past in the foregroundhad only begun to emerge but who had already acquired the use of the tensedcopula in the background. Flashner's three learners, one beginner, intermediate, andadvanced learner, exemplify the middle stages of development which show varyingproportions of marking of the foreground over the background. Finally, it is clearthat the 16 intermediate learners studied by Bardovi-Harlig (1992b) do not constitutea group at a uniform stage of development but would be spread out across thegroups used to divide the learners in this study. The most advanced learner in thatA Narrative Perspective on Tense/Aspect in SLA 285study shows the equal, high use of past in foreground and background (90%), whichis characteristic of W90, the most advanced group in this study. Thus, we see thatthe features of grounding in interlanguage are not static but that they change asinterlanguage develops.An analysis based on narrative structure also suggests a functional analysis oftense distribution. The temporal criteria for the foreground of sequentiality, punctu-ality, and completeness (Reinhart, 1984) offer an explanation of why the simple pastdominates the foreground. The foreground is functionally simple: The foregroundmoves the narrative along and in essence tells the story. In contrast, the backgroundis functionally complex. It does any number of things that can be characterized asproviding supportive material for the foreground. The background may look ahead,look backward, evaluate, orient, abstract, or hypothesize, and it may even recountsequential events on a time axis different from that of the main event. Even if thedominant tense of the background is the simple past for native speaker narratives,background use of simple past clearly pushes the use of simple past beyond any coreunderstanding of it for the learner.Moreover, although the dominant use of simple past is expected in a narrativeoriented to the past (vs. an historical present narrative), in the background there isa much greater morphological diversity than in the foreground. Nevertheless, thedegree of diversity is dependent on how the story is crafted. Although the pluperfectand past progressive occur in the background in native speaker texts, their occur-rence there is optional. For example, if a story is told in chronological order, thepluperfect is unlikely. If all events are reported as sequential, there will be no needfor the past progressive, which is more typically used if actions are simultaneous.Thus, for native speakers and learners alike, the use of tense/aspect forms dependson how the text is constructed. It is therefore difficult to predict with what frequencyperfect or progressive will be used in any given text, but it is possible to comparelearner frequency of use to native speaker frequency for the same text. For example,as noted earlier, the use of progressive in the learners' written narratives in GroupsW70-W90 compared favorably to the native speakers' narrative for the same film.Further investigation of the background seems promising for the study of perfectand progressive because one could study which events or situations are reported inthe background and how they are encoded. One such investigation of the back-ground reports on the expression of reverse order reports and the acquisition ofpluperfect (Bardovi-Harlig, 1994b).Investigating the acquisition of tense/aspect from the perspective of narrativestructure does not preclude other analyses such as an analysis based on lexicalaspect. Following the influential work of Andersen (1986, 1991; Andersen & Shirai,1994), a number of studies have begun to investigate the influence of lexical aspectin shaping the tense/aspect system (see, e.g., Bardovi-Harlig & Bergstrom, in press;Bardovi-Harlig & Reynolds, 1995; Bayley, 1994; Bergstrom, forthcoming; Hasbiin,1994; Ramsay, 1990). The next step in the investigation would seem to be to analyzethe same set of narratives by lexical aspect. An overlap between the analyses isapparent when one considers the temporal criteria for foregrounding. The subcriter-ion of punctuality is the denning feature of achievements, and completedness relates286 Kathleen Bardovi-Harligto both achievements and accomplishments (telic verbs). Sequentiality is not directlyrelated to aspectual class, but only events that are reported as completed can besequenced (Dowty, 1986). The narrative hypothesis and the aspect hypothesis do infact make different predictions and can be distinguished, but such distinctions maybe too fine-grained for a study of interlanguage (Bardovi-Harlig, 1994a).Examining narratives also reveals what type of things seem to be most easilyexpressed. Sequenced actions are clearly easy to express. The simplest narrativeshave foregrounds but may not have background. The hypothetical episode, relatedthrough talk as a daydream, is difficult to encode, particularly because the filmshows the episode and so it is in one sense a string of past events, yet it is somehowless real than the rest of the film because it was imagined. It is not clear how thenative speakers would mark the imagination sequence if it included several eventsbecause the native speakers tested by Reynolds (1994) only nominalized the event,thereby avoiding any use of tense or verbal forms. Nevertheless, learners do attemptto delineate the imagined sequence, through both tense and the marking of bound-aries.Grouping learners separately for oral and written production by overall use ofpast in past-time contexts shows little effect for modality. By and large, oral andwritten narratives were quite similar when compared to equivalent groups acrossmodality: The proportion of background to foreground verbs, the frequency ofuse of the copula in the background, and the progression of tense/aspect use areremarkably similar. This group as a whole showed somewhat lower use of past inthe oral narratives, but there were also individual learners who showed higher useorally (see Table 1 and the Appendix). Nevertheless, the patterns of tense/aspectdevelopment with respect to narrative structure are the same.CONCLUSIONThis paper has shown that narrative structure plays a key role in the development ofthe tense/aspect system. Grouping learners according to the use of past in past-timecontexts allows us to address the question of how learners employ the limited tense/aspect marking that they can produce. After an early stage in which nonpast is thefavored form in both foreground and background, learners mark foreground eventsfor past first and use a variety of forms in the background, progressing towarda more nativelike distribution with increasing proficiency. The results permit theassimilation of earlier findings into a clear developmental sequence in the acquisitionof the tense/aspect system.NOTES1. In the literature a distinction is made between narrative discourse and narrative context (or narrativeclauses). Narrative discourse encompasses the entire narrative text, or what this paper simply calls thenarrative, whereas the narrative context or clauses refer to the foreground specifically (Dahl, 1984; Labov,1972; Schiffrin, 1981).2. For a discussion of the additional advantages of the use of film, see Chafe (1980) and Bardovi-Harlig(1994a).3. This was one of six films used in the collection of narratives over a 2-year period (Bardovi-Harlig,A Narrative Perspective on Tense/Aspect in SLA 2871992a). Modern Times was selected for analysis because we had collected more oral and written narrativepairs for this film than for any of the others. The ESF project also used a similar excerpt of Modern Times(Klein & Perdue, 1992), which should facilitate crosslinguistic comparisons in future studies.4. In order the titles were as follows: (a) Alone and hungry, (b) \"She stole a loaf of bread.\" (c) \"No, shedidn't. I did.\" (d) \"It was the girl—not the man.\" (e) \"Remember me—and the bread?\" (f) \"Now is your chanceto escape.\" (g) \"Where do you live?\" (h) \"No place—anywhere.\" (i) \"Can you imagine us in a little home likethat?\" (j) \"I'll do it! We'll get a home, even if 1 have to work for it.\"5. For example, a study of the narratives produced by 16 intermediate-level learners found that appro-priate use of past ranged from 15 to 90% in oral production and from 32 to 98% in written production(Bardovi-Harlig, 1992b).6. The learners were then grouped according to their oral production. The resultant grouping placedonly five learners in the 60% range and only two in the 70% range. Thus, three additional speakers wereincluded on the basis of their oral scores, placing one in Group 60 and two in Group 70. This balancing of theoral groups resulted in a relatively large written Group 70.7. Following established practice for narrative analysis also eliminates a potential problem in that theonly language included in the silent film was the speech of the actors. Thus, any potential verbal influence ofthe film was eliminated.8.1 took the most conservative position on the exclusions, combining Bayley's and Wolfram's exclu-sions. Bayley excludes homophonic (same voicing value) stops and interdentals, and Wolfram excludes onlyhomorganic stops.9. Present perfect and 0+ perfect are quite rare in this corpus. Only 10/2,813 verbs (0.4%) fell into thiscategory.10. Group 10 oral shows only one verb (1/12) used in the past in the background, and that was a tensedcopula, uxis.11. The background shows very high use of the copula. Just over half of all the tensed verbs (simple pastor simple present) are copula verbs in both oral (52.6%) and written (53.6%) narrative. This is due in part tothe descriptive function of the background.12. The one exception was Group 10 in the oral narratives, whose use of present progressive in theforeground at 17.1% is second only to the use of base forms at 48.6%, but equal to other marked, butuninterpretable forms labeled \"Other.\"13. It has been argued that negative events are not foreground, but background. However, in the case ofstating the nonpayment by the main character, we followed Fleischman (1985) and van Dijk (1975), whoclaim that such cases must be considered to be events when they are expected or part of a script. Paying forfood in a restaurant is part of a script. This was the only negative event in the narratives. Seventeen of thenarratives (6 written and 11 oral) reported the nonpayment in this way.14. This use of present prefect can be found in pluperfect contexts as an early marker of reverse-orderreports (Bardovi-Harlig, 1994b).REFERENCESAndersen, R. W. (1986). El desarollo de la morfologi'a verbal en el Espanol como segundo idioma. In J. M.Meisel (Ed.), Adquisicion de languaje/Aquisicao da linguagem (pp. 115-138). Frankfurt: Vervuert.Andersen, R. W. (1991). Developmental sequences: The emergence of aspect marking in second languageacquisition. In C. A. Ferguson & T. Huebner (Eds.), Second language acquisition and linguistic theories(pp. 305-324). Amsterdam: Benjamins.Andersen, R. W., & Shirai, Y. (1994). Discourse motivations for some cognitive acquisition principles. Studiesin Second Language Acquisition, 16,135-156.Bamberg, M. (1987). The acquisition of narratiues. Berlin: Mouton de Gruyter.Bamberg, M., & Marchman, V. (1990). What holds a narrative together? The linguistic encoding of episodeboundaries. Papers in Pragmatics, 4, 58-121.Bardovi-Harlig, K. (1992a). Final report to the National Science Foundation (Grant DBS-19616).Bardovi-Harlig, K. (1992b). The telling of a tale: Discourse structure and tense use in learners' narratives. InL. Bouton & Y. Kachru (Eds.), Pragmatics and language learning (Vol. 3, pp. 144-161). Urbana: Divisionof English as an International Language, University of Illinois.Bardovi-Harlig, K. (1992c). The use of adverbials and natural order in the development of temporal expres-sion. International Review of Applied Linguistics in Language Teaching, 30, 299-320.Bardovi-Harlig, K. (1994a). Anecdote or evidence? Evaluating support for hypotheses concerning the devel-opment of tense and aspect. In E. Tarone, S. M. Gass, & A. Cohen (Eds.), Research methodology insecond language acquisition (pp. 41-60). Hillsdale, NJ: Erlbaum.288 Kathleen Bardovi-HarligBardovi-Harlig, K. (1994b). Reverse-order reports and the acquisition of tense: Beyond the principle ofchronological order. Language Learning, 44,243-282.Bardovi-Harlig, K., & Bergstrom, A. (in press). The acquisition of tense and aspect in SLA and FIX: A studyof learner narratives in English (SL) and French (FL). Canadian Modern Language Review.Bardovi-Harlig, K., & Reynolds, D. W. (1995). The role of lexical aspect in the acquisition of tense andgrammatical aspect. TESOL Quarterly, 29,107-131.Bayley, R. (1994). Interlanguage variation and the quantitative paradigm: Past tense marking in ChineseEnglish. In E. Tarone, S. M. Gass, & A. Cohen (Eds.), Research methodology in second languageacquisition (pp. 157-181). Hillsdale, NJ: Erlbaum.Bergstrom, A. (forthcoming) The expression of past temporal reference by English-speaking adult learners ofFrench. Unpublished doctoral dissertation, Pennsylvania State University, University Park.Chafe, W. (Ed.). (1980). The Pear stories. Norwood, NJ: Ablex.Dahl, 0. (1984). Temporal distance: Remoteness distinctions in tense-aspect systems. In B. Butterworth, B.Comrie, & 6. Dahl (Eds.), Explanations for language universals (pp. 105-122). Amsterdam: Mouton.Dowty, D. (1986). The effects of aspectual class on the aspectual structure of discourse: Semantics orpragmatics? Linguistics and Philosophy, 9, 37-61.Dry, H. (1981). Sentence aspect and the movement of narrative time. Text, 1, 233-240.Dry, H. (1983). The movement of narrative time. Journal of Literary Semantics, 12, 19-53.Flashner, V. E. (19). Transfer of aspect in the English oral narratives of native Russian speakers. In H.Dechert & M. Raupach (Eds.), Transfer in language production (pp. 71-97). Norwood, NJ: Ablex.Fleischman, S. (1985). Discourse functions of tense-aspect oppositions in narrative: Toward a theory ofgrounding. Linguistics, 23, 851-882.Hasbun, L. (1994). The effect of lexical aspect on the acquisition of tense and grammatical aspect in Spanishas a foreign language. Unpublished manuscript, Indiana University, Bloomington.Hopper, P. J. (1979). Aspect and foregrounding in discourse. In T. Givon (Eds.), Syntax and semantics:Discourse and syntax (pp. 213-241. New York: Academic Press.Hopper, P. J., & Thompson, S. A. (1980). Transitivity in grammar and discourse. Language, 56, 251-299.Klein, W., & Perdue, C. (Eds.). (1992). Utterance structure,-Developing grammars again. Amsterdam: Benja-mins.Kumpf, L. (1984). Temporal systems and universality in interlanguage: A case study. In F. Eckman, L. Bell, &D. Nelson (Eds.), Universals of second language acquisition (pp. 132-143). Rowley, MA: NewburyHouse.Labov, W. (1972). The transformation of experience in narrative syntax (chapter 9). In Language in the innercity (pp. 3-396). Philadelphia: University of Pennsylvania Press.Labov, W., & Waletzky, J. (1967). Narrative analysis: Oral versions of personal experience. In J. Helm (Ed.),Essays on the verbal and visual arts (pp. 12-44). Seattle: University of Washington Press.Longacre, R. (1981). A spectrum and profile approach to discourse analysis. Text, 1, 337-359.Ramsay, V. (1990). Developmental stages in the acquisition of the perfective and the imperfective aspects byclassroom L2 learners of Spanish. Unpublished doctoral dissertation, University of Oregon, Eugene.Reinhart, T. (1984). Principles of gestalt perception in the temporal organization of narrative texts. Linguis-tics, 22, 779-809.Reynolds, D. W. (1994 )lexical aspect and grounding in NS narrative: Baseline data for the acquisition oftemporal morphology. Unpublished Ph.D. qualifying paper, Indiana University, Bloomington.Robison, R. E. (1990). The primacy of aspect: Aspectual marking in English interlanguage. Studies in SecondLanguage Acquisition, 12, 315-330.Schiffrin, D. (1981). Tense variation in narrative. Language, 57, 45-62.Schumann, J. H. (1987). The expression of temporality in basilang speech. Studies in Second LanguageAcquisition, 9, 21-41.Silva-Corvalan, C. (1983). Tense and aspect in oral Spanish narrative. Language, 59, 760-780.Tannen, D. (19). Talking voices: Repetition, dialogue, and imagery in conversational discourse. Cambridge:Cambridge University Press.Tomlin, R. S. (1984). The treatment of foreground-background information in the on-line descriptive dis-course of second language learners. Studies in Second Language Acquisition, 6, 115-142.van Dijk, T. A. (1975). Action, action description, and narrative. New Literary History, 6, 273-294.Veronique, D. (1987). Reference to past events and actions in narratives in L2: Insights from North Africanlearners' French. In C. W. Pfaff (Ed.), First and second language acquisition processes (pp. 252-272).Cambridge, MA: Newbury House.Wolfram, W. (1985). Variability in tense marking: A case for the obvious. Language Learning, 35, 229-253.A Narrative Perspective on Tense/Aspect in SLA2APPENDIXThe left-hand column of the tables identifies the subjects by group. LI and Intensive EnglishProgram (IEP) instructional level follows. \"Written/Oral Types\" gives the percentage of use ofpast in past-time contexts for verb types. The column labeled \"Type/Token\" gives the type-token ratios for each narrative, which is the ratio of the number of distinct realizations of verbsto the total number of verbs used in past-time contexts. \"No. Verbs\" gives the total number ofverbs (in all contexts) that were used by learners. The right-hand column gives the learner'sgroup in the opposite modality.SUBJECT PROFILES BY GROUP—WRITTEN NARRATIVESWrittenGroup1010203030404040505050606060607070707070707070708080808080SubjectNo.12345671011121314151617181920212223242526272829LIKoreanTaiwaneseKoreanSpanishArabicArabicArabicKoreanArabicArabicArabicSpanishSpanishKoreanKoreanKoreanJapaneseArabicArabicArabicArabicArabicArabicJapaneseSpanishJapaneseSpanishKoreanArabicIEPLevel1215253132233221145134WrittenTypes13.3 (2/15)13.6 (3/22)28.6 (6/21)33.3(11/33)33.3 (9/27)41.0(16/39)41.7(10/24)47.4 (9/19)50.0(12/24).5(12/22)55.6(25/45)61.5 (8/13)61.5(16/26)65.4(17/26)66.7(12/18)70.4(19/24)71.4(15/21)71.4(10/14)73.3 (33/45)73.7 (28/38)75.0(12/16)76.2(16/21)76.9(20/26)78.8 (26/33)82.4(14/17)82.4(28/34)83.8(31/37)84.6(22/26)85.0(17/20)WrittenType/Token0.75(15/20)0.67 (22/33)0.66(21/32)0.79(33/42)0.(27/42)0.63 (39/62)0.63 (24/38)0.83(19/23)0.59(24/41)0.81 (22/27)0.60 (47/75)0.62(13/21)0.59 (27/46)0.67 (26/39)0.36(18/50)0.66(27/41)0.55(21/38)0.56(14/25)0.66(45/68)0.63 (38/60)0.76(16/21)0.53 (21/40)0.52 (26/50)0.62 (33/53)0.81(17/21)0.69 (34/49)0. (37/69)0.70 (26/37)0.67 (20/30)WrittenOralNo. VerbsGroup1534334245674119502876214940574338277068224452532149693932O10O30O10O40O30O20O30O10O20O40O40O20O40O70O20O20060O70O60O40O30O70O50O50O50O70O50O20O40353252329080809090909090903031323334353637KoreanArabicJapaneseJapaneseKoreanKoreanJapaneseSpanish23566585.7(18/21)88.5 (23/26)92.0 (23/25)95.0(19/20)97.3 (36/37)100(26/26)100(16/16)100(42/42)Kathleen Bardovi-Harlig0.84(21/25)0.62 (26/42)0.81(25/31)0.65(20/31)0.84 (37/44)0.79 (26/33)0.(16/25)0.63 (42/67)2942323245332567O30O50O60O30O60O60O80O60SUBJECT PROFILES BY GROUP—ORAL NARRATIVESOralGroup101010202020202020303030303030404040404040505050505060606060SubjectNo.381162812915623353072111420131029233127252419323717LIKoreanKoreanKoreanKoreanKoreanSpanishArabicKoreanArabicTaiwaneseJapaneseArabicKoreanArabicArabicArabicSpanishArabicSpanishArabicArabicArabicArabicSpanishSpanishJapaneseArabicJapaneseSpanishJapaneseIEPLevel1111233252322332513233653551Oral Types6.7 (1/15)10.0 (1/10)15.4 (1/13)21.2 (4/19)23.1 (3/13)24.1 (7/29)25.0 (9/36)25.0 (6/24)28.2(11/39)33.3 (6/18)33.3 (7/21)34.6 (9/26)35.7 (5/14)38.1 (8/21)38.6(17/44)40.7 (22/)42.9(15/35)42.6(20/47)44.1(15/34)47.6(10/21)47.6(10/21)50.0(10/20)50.0(16/32)53.6(15/28)55.3(26/47)59.3(16/27)60.9 (28/46)62.1(18/29)62.5(20/32)63.0(17/27)OralType/Token0.88(15/17)0.91(10/11)0.72(13/18)0.76(19/25)1.0 (13/13)0. (29/45)0.66(36/55)0.65 (24/37)0.55(39/71)0.(18/28)0.62 (21/34)0.55 (26/47)0.82(14/17)0.60(21/35)0.46 (44/96)0.63 (/86)0.70 (35/50)0.42(47/112)0.59 (34/58)0.84(21/25)0.72(21/29)0.57 (20/34)0.43 (32/74)0.74 (28/38)0.49 (47/95)0.69 (27/39)0.61 (46/74)0.60 (29/48)0.44 (32/73)0.57(27/47)OralWrittenNo. VerbsGroup1711182613505943803036562042107991253232377439120768347W20W40W10W70W80W60W50W60W40W10W90W30W80W40W70W50W30W70W60W50W80W70W80W80W80W70W70W90W90W70A Narrative Perspective on Tense/Aspect in SLA6060707070708034352622141836KoreanKoreanJapaneseArabicKoreanArabicJapanese562424666.7(18/27)66.7(16/24)70.0 (30/43)72.7 (24/33)75.0(12/16)75.0(18/24).5(17/19)0.81 (26/32)0.59(24/41)0.69 (34/49)0.60 (33/55)0.80(16/20)0.67 (24/36)0.70(19/24)33477562204024291W90W90W80W70W60W70W90