Preliminary page and line breaksmcmanuskevin.com/wp-content/uploads/2018/09/iral-2013-0013.pdf · Imparfait German Perfekt Imperfekt and situation were observed in second language

Preliminary page and line breaks

Mouton de Gruyter — 1st proofs

1-iral-51-3 — 2013/7/10 11:37—299— #58—ce

Prototypical influence in second language acquisition:What now for the Aspect Hypothesis1

KEVIN MCMANUS

IRAL 51 (2013), 299–322 0019042X/2013/051-299DOI 10.1515/iral-2013-0013 c©Walter de Gruyter

Abstract

This paper presents empirical evidence on the development of aspect by English-and German-speaking university learners of French L2 collected from a spokennarrative task and a sentence interpretation task. Contrary to the Aspect Hy-pothesis’s predictions, this study’s results suggest that increased use of proto-typical pairings goes in hand with increased L2 proficiency. Following a smallbut growing number of studies, this study questions the route of L2 developmentproposed by the Aspect Hypothesis. It is proposed that the route of develop-ment found in the data can be accounted for by considering the influence of L1background on L2 development.

1. Introduction

For a long time now, research on the development of aspect in a second lan-guage (L2) has almost dutifully considered the predictions made by the AspectHypothesis (Andersen and Shirai 1994, 1996). The hypothesis adopts a bidi-mensional approach to aspect (Bertinetto, 2001, Depraetere 1995, Smith 1997)in which aspectual information is split into two types of universal semantic in-formation: (1) the inherent semantics of verbs/predicates/sentences (situationaspect) and (2) the particular perspective from which situations are presented(viewpoint aspect). In languages that map viewpoint aspect to tense (e.g.,French, Spanish), the Aspect Hypotheses (AH) claims that certain viewpointaspect meanings are initially only used with particular situation aspect types(e.g., perfective viewpoint initially used with telic situation types). A number

1. This research was supported by the School of Modern Languages at Newcastle Universityand the Association for French Language Studies. I would like to thank the anonymous IRALreviewers for their valuable comments as well as Florence Myles, Richard Waltereit, LauraDomínguez and Nicole Tracy-Ventura.



1-iral-51-3 — 2013/7/10 11:37—300— #59—ce

300 Kevin McManus

of studies claim to find empirical support in the AH’s predictions for L2 devel-opment (e.g. Bardovi-Harlig and Bergström 1996, Comajoan 2006). However,more recently, researchers have questioned the AH both in terms of its predic-tions (Ayoun and Salaberry 2005, Domínguez et al., in press, Labeau 2005,Salaberry, 2008) and the rationale underlying its predictions (McManus 2011,Slabakova 2002). The present study supports a small but growing body of ev-idence that fails to support for AH’s predictions. However, it differs markedlyfrom previous research on the development of aspect in a L2 by triangulatingproduction and interpretation data with learners (n=75) from different first lan-guage backgrounds (English and German) learning the same L2 (French) attwo different proficiency levels.

In this paper I analyse empirical evidence on the development of aspect byEnglish- and German-speaking university learners of French L2 collected froma spoken narrative task (Natalie et Albert) and a sentence interpretation task.This analysis shows that contrary to the AH’s predictions, increased use ofprototypical pairings goes in hand with increased L2 proficiency.

2. Background

2.1. Aspect

Aspect deals with two types of universal semantic information: (1) the inherentsemantics of verbs/predicates/sentences (situation aspect, or lexical aspect) and(2) the particular perspective from which situations are presented (viewpointaspect, or grammatical aspect), defined as ‘the semantic domain of the tempo-ral structure of situations and their presentation’ (Smith 1997:01). FollowingSmith (1997), aspect necessarily involves the interaction between viewpointaspect and situation aspect.

The inherent semantics of predicates is traditionally referred to as lexical as-pect, or Situation Aspect (Smith 1997) and deals with Vendler’s (1957, 1967)classification of verbs/predicates into four lexical semantic classes based ontheir inherent semantic features. Table 1 (adapted from Arche 2006: 42)presents Vendler’s lexical classification of predicates with examples.

Situations are categorized on the basis of inherent temporal properties: dura-tivity, dynamicity and telicity. In this respect, situations differ from each otherbased on specific clusters of these temporal features. Based on the seminalclassification proposed by Vendler (1957, 1967), research generally acknowl-edges four basic situation types: Statives, Activities, Accomplishments, andAchievements. A situation may differ from another in terms of its inclusionand/or absence of a particular temporal feature. Table 2 shows the temporalfeatures of each of these situations.



1-iral-51-3 — 2013/7/10 11:37—301— #60—ce

Prototypical influence in second language acquisition 301

Table 1. Vendler’s lexical classification of predicates

Stative Activity Accomplishment Achievement

No-actions thathold in time butdo not take time.They lack anykind of internalstructure.

Events with dura-tion but no end-point.

Actions with a cul-minating endpointthat take durationto be completed.

Instantaneousevents, with anendpoint but noduration

be green, be sick,know, belong, hate,love

Swim, push a cart,write novels, walkaround the park

walk to the beach,build a house, reada chapter

die, be born, fallasleep, arrive,recognize, awaken,collapse, explode

Table 2. The temporal features of the situation types

Situation Dynamic Punctual Telic

Stative [−] [−] [−]Activity [+] [−] [−]Accomplishment [+] [−] [+]Achievement [+] [+] [+]

From Table 2 a number of generalizations over situations can be drawn.Firstly, all situations are dynamic except for statives. Secondly, only achieve-ments are punctual. Lastly, accomplishments and achievements are telic, whilstactivities and statives are atelic.

The second type of aspectual information considered to be universal to hu-man languages refers to the perspective or the viewpoint on a situation thata speaker adopts, as presented in sentences. This means that the temporaltime-course of the same situation may be presented differently. Viewpointsare described as perfective and imperfective. Perfective viewpoint presents asituation inclusive of its endpoints, indicating a ‘complete’ perspective, whilstimperfective viewpoint presents a situation absent of boundaries, indicating an‘ongoing’ or ‘non-complete’ perspective. The sentences in (1) demonstrateperfective and imperfective viewpoint in French.

(1) a. Jean a joué dans le parc‘Jean played in the park’

b. Jean jouait dans le parc‘Jean was playing/ used to play in the park’

Where past time reference is concerned, French marks the perfective-imperfective



1-iral-51-3 — 2013/7/10 11:37—302— #61—ce

302 Kevin McManus

viewpoint distinction morphosyntactically: the Passé Composé (PC) conveysperfective viewpoint (1a) and the Imparfait (IMP) conveys imperfective view-point (1b). In (1a) the situation <jouer dans le parc> is presented as a completewhole, inclusive of its endpoints. In contrast, the same situation in (1b) is un-specified with respect to completion. The sentences in (1) represent exampleswhere viewpoint aspect is marked with aspectual morphemes. However, view-point aspect is not universally expressed by aspectual morphemes. This is bestexemplified in German.

German is devoid of viewpoint aspect morphemes altogether (Bertinetto2001, Bohnemeyer and Swift 2004, Comrie 1976), yet it still conveys view-point aspect. Bohnemeyer and Swift (2004) and Smith (2006) argue that inlanguages that lack viewpoint aspect morphemes, viewpoint information is in-terpreted from the inherent semantics of the sentence (i.e., situation aspect in-formation), or as Bohnemeyer and Swift (2004: 266) put it, ‘clauses and verbalprojections not overtly marked for viewpoint aspect are assigned viewpoint-aspectual operators on the basis of the telicity of their event predicates’. There-fore, in the absence of aspectual morphemes atelic predicates are said to entailimperfectivity and telic predicates entail perfectivity, resulting in sentences re-ceiving preferential interpretations (Bohnemeyer and Swift 2004, Smith 2006,Sonnenhauser 2008), as shown in (2):

(2) a. Als ich Marys Büro betrat, schrieb sie einen Brief.‘When I entered-SP Mary’s office, she wrote-SP a letter’.

b. Als ich Marys Büro betrat, schrieb sie einen Brief. Überraschtblickte sie auf, legte den Stift zur Seite, und lächelte mich an.‘When I entered-SP Mary’s office, she was writing-SP a letter.Surprised, she looked-SP up, put-SP the pen away, and smiled-SPat me’.

The sentences in (2), from Bohnemeyer and Swift (2004: 268–269), showviewpoint aspect contrasts for einen Brief schreiben (‘write a letter’) despitelacking viewpoint aspect morphemes. Einen Brief schreiben is telic and isalso interpreted perfectively unless this information is overridden by other el-ements. For Bohnemeyer and Swift (2004: 269), then, (2a) ‘suggests that thewriting event’s onset coincided with the entering (at least as this reading ismore plausible than the overlap reading) i.e. the writing event is interpretedperfectively’. (2a) is interpreted perfectively because of the predicate’s telicity.In (2b) einen Brief schreiben is still present and it is still telic. However, einenBrief schreiben in (2b) demonstrates that a perfective entailment with telic situ-ation types can be cancelled out. With greater contextual detail on the enteringevent, entering overlaps with letter writing and completion of letter writing isnot entailed. Therefore, telicity appears to entail perfectivity in the absence



1-iral-51-3 — 2013/7/10 11:37—303— #62—ce


of viewpoint aspect morphemes. Entailment only holds if there is no explicitcontextual information to the contrary.

In comparison, English marks viewpoint aspect by tense. However, unlikeFrench, it lacks a general imperfective form. Perfectivity is marked with theSimple Past (SP). Progressivity is marked with the Progressive tense. Habit-uality is marked using would, the periphrasis used to and additionally the SP(Comrie 1976; Tagliamonte and Lawrence 2000; Traugott 1972), as shown in(3):

(3) a. Tony used to play football.b. Tony would play football.c. Tony played football.

Quirk and Greenbaum (1972: 43) claim that used to, would and the SP arealternating forms that can be used to convey habituality in English. To thisend, Tagliamonte and Lawrence (2000) also demonstrate how habitual formscan be alternated in English with no apparent differences in meaning, as in (4)

(4) Well we used to go [/went] every week. It was one of those things wedid [/used to do] every week.

In (4) Tagliamonte and Lawrence (2000: 325) note that ‘went’ may be usedinstead of ‘used to go’ and ‘used to do’ may replace ‘did’ leading to no ‘ap-parent change in meaning’. In a study based on conversation data of BritishEnglish with 92 speakers (totalling approximately 1.5 million words), Taglia-monte and Lawrence (2000) report on how the SP, used to and would are usedin English to convey habitual viewpoint. The results show that ‘the vast major-ity, nearly 70 per cent, of all habitual past contexts in English are realized withpreterit morphology [the SP]. This contrasts with the picture one gets from theliterature where only used to and/or would are cited as habitual past markers’(Tagliamonte and Lawrence 2000: 329).

To summarise these crosslinguistic differences between English, French, andGerman for viewpoint aspect, Table 3 shows the forms used to mark viewpointaspect, where � stands for yes and � for no.

2.2. The Aspect Hypothesis and semantic prototypes

The AH is a hypothesis for L2 development specifically dealing with prototyp-icality. It arose from studies in L1 acquisition arguing that viewpoint markingemerges before marking for time reference (Antinucci and Miller 1976, Bron-ckhart and Sinclair 1973). When learners used tense, it was to mark aspectualdistinctions and not temporal ones. Similar correlations between viewpoint



1-iral-51-3 — 2013/7/10 11:37—304— #63—ce

304 Kevin McManus

Table 3. Form-meaning pairings for perfective and imperfective viewpoint aspect inEnglish, French and German

Form Viewpoint

Perfective Imperfective

Progressive Habitual

English Progressive � � �

Simple Past � � �

Used to � � �

Would � � �

French Passé Composé � � �

Imparfait � � �

German Perfekt � � �

Imperfekt � � �

and situation were observed in second language acquisition (SLA) research(Andersen 1991) and formed the basic postulates of the AH:

– Learners first use past marking (e.g., English) or perfective marking (e.g.,Chinese, Spanish) on achievement and accomplishment verbs, eventuallyextending its use to activities and stative verbs.

– In languages that encode the perfective/imperfective distinction, imperfec-tive past appears later than perfective past, and imperfective past markingbegins with stative verbs and activity verbs, then extending to accomplish-ment and achievement verbs.

– In languages that have progressive aspect, progressive marking begins withactivity verbs, then extending to accomplishment or achievement verbs.

– Progressive markings are not incorrectly overextended to stative verbs.(Andersen and Shirai 1996: 533)

As Slabakova (2002) notes, the AH is a data-driven observation for L2 devel-opment: the postulates in 1–4 above are based on observed patterns of lan-guage use. However, there are theoretical explanations for it, principally inPrototype Theory (Rosch, 1973, Rosch and Mervis 1975, 1978). PrototypeTheory accounts for human categorization, assuming ‘there are good members(prototypes) and marginal members of a category, the goodness being gradi-ent and determined by the commonality with the central members (prototype)of the category’ (Andersen and Shirai 1996: 555). Applied to SLA, learnersacquire a category starting with its prototype and later expand its applicationto less prototypical cases. Andersen and Shirai (1996: 558) claim that ‘ac-tion in process or progress seems to be the prototype of progressive aspect [. . .



1-iral-51-3 — 2013/7/10 11:37—305— #64—ce


and] within the process meaning, prototypical cases are activity and accom-plishment verbs’. In a nutshell, L2 learners start to acquire aspect by pairingprototypes of viewpoint and situation type; for example, progressive viewpointmarkers with activity situation types. Perfective viewpoint is characterized as‘complete’, which is paired with situation types of the same ‘complete’ char-acterization, essentially telic situation types. Imperfective viewpoint has themeaning ‘not complete’ or ‘ongoing’, which is paired with situation types ofthe same meaning, that is atelic situation types (statives and activities). L2development is characterized by increased use of non-prototypical pairings,indicated by Andersen and Shirai’s ‘extended uses’ in the AH. Therefore, ac-cording to the AH, as proficiency increases nonprototypical pairings developfollowing on from prototypical ones: that is, then, perfective viewpoint withatelic situation types and imperfective viewpoint with telic situation types.

2.3. Prototypes in SLA

Greatest support for prototypical influence in L2 development is claimed tocome from Bardovi-Harlig and Bergström (1996) and Comajoan (2006). Bar-dovi-Harlig and Bergström (1996) collected data from written narratives basedon the Charlie Chaplin silent film Modern Times. Learners were split into four‘proficiency’ groups per L2 according to their past tense use in obligatory pasttime contexts. All French L2 learners show prototypical preferences. Table 4(adapted from Bardovi-Harlig and Bergström 1996: 317) presents learners’scores for viewpoint marking in French L2 with the IMP and the PC with sta-tives and achievements.

Group results (Table 4) show that prototypical influence is greatest at thelowest proficiency level: PC with achievements (70.3 %) and IMP with sta-tives (100 %). Then, as proficiency increases, prototypical influence appearsto reduce, whereby the IMP with statives and the PC with achievements isused less. For nonprototypicality, although group results are low, there is

Table 4. French L2 learners’ viewpoint marking in per cent in Bardovi-Harlig andBergström (1996)

Statives Achievements

IMP PC IMP PC

Group 1 (n=4) 100 0 0 70.3Group 2 (n=7) 50 3.3 33.3 68.1Group 3 (n=7) 70 6.5 5 52.3Group 4 (n=5) 45.1 4.6 19.6 63.3



1-iral-51-3 — 2013/7/10 11:37—306— #65—ce

306 Kevin McManus

change. Group 1 learners show no nonprototypical uses (IMP with achieve-ments and PC with statives), whereas from Group 2 onwards nonprototypicaluse increases, although not in a linear route of development, as predicted bythe AH. This is observed for both the IMP with achievements and the PC withstatives. Bardovi-Harlig and Bergström argue that learners use more nonpro-totypical pairings as proficiency increases. Their results suggest that learnersare influenced more by prototypes at the earlier stages of acquisition than lateron. As proficiency increases, prototypicality in production appears to reduceas nonprototypicality increases. These patterns of development are consistentwith the AH’s predictions.

Comajoan (2006) also observes initial prototypical influence in three English-speaking university learners of Catalan L2. In his longitudinal research design,data were collected at various points over seven months from three learners.Results are split into two stages in order to compare development over time.Like Bardovi-Harlig and Bergström (1996), Comajoan (2006) also used Mod-ern Times to elicit narratives, as well as storybook retellings and a personalnarrative.

Comajoan’s results show that L2 learners initially use the Imperfect and thePreterit with all situation types. As predicted by the AH and as found forBardovi-Harlig and Bergström’s French L2 learners, Comajoan (2006) notesthat prototypicality reduces over time. For the Preterit with telic situation types,Daniel’s and Robert’s use decreases between Stage 1 and Stage 2 (100 % →86.5 % and 86.2 % → 80.6 %, respectively), however increases in prototypi-cality are observed for Barbara: 67.4 % → 79.4 %. For the Imperfect withstatives, all learners’ prototypical pairings decrease between Stage 1 and Stage2: Daniel 33.3 % →27 %; Barbara 52.9 % → 36.5 %; Robert 45.8 % → 39.1 %.These results show less of a prototypical influence on L2 development at Stage2 than at Stage 1. Furthermore, nonprototypicality increases over time. Alongthe same lines as Bardovi-Harlig and Bergström’s (1996) French L2 data, Co-majoan’s results generally show that between Stage 1 and Stage 2 prototypi-cality reduces as nonprototypicality increases.

However, Salaberry (1999) data from 20 English-speaking university learn-ers of Spanish L2 is less supportive of the AH. His data were collected at twodifferent points in time, two months apart. Learners were divided into fourdifferent groups based on their level of instruction. Participants also producedspoken narratives based on Modern Times. Salaberry’s results show that alllearners indicate prototypical influence beyond the least proficient group. Ta-ble 5 shows Salaberry’s (1999: 166) results in terms of prototypical and non-prototypical pairings.

The results (Table 5) show that from Group 2, the Preterit and the Imperfectare used with both telic and stative situation types, showing preference forprototypical pairings. Group results for prototypes between (a) the Preterit and



1-iral-51-3 — 2013/7/10 11:37—307— #66—ce


Table 5. Prototypicality and nonprototypicality in per cent in Salaberry (1999)

Time 1 Time 2Telic Stative Telic Stative

Group 1 (n=4)Preterit 60 15 81 12Imperfect 0 100 50 50




telic situation types and (b) the Imperfect and statives generally increase overtime alongside proficiency. Salaberry’s findings show that the more proficientgroups (e.g., Groups 3 and 4) show greater use of prototypical combinationsthan the less proficient groups (e.g., Groups 1 and 2). This is exemplified byGroup 2 and Group 3 learners, who show greater prototypical combinations atTime 2 than at Time 1. Furthermore, in each group at Time 1, the higher theproficiency level, the greater the use of prototypical combinations. It is onlyGroup 4 learners that buck this trend indicating greater prototypical influenceat Time 1 than Time 2.

Overall, Salaberry’s results show that less proficient learners show greateruse of nonprototypes and then this pattern dips as proficiency increases. Theresults then show that nonprototypical use begins to increase again: these pro-totypical patterns of L2 development are not linear as the AH suggests, butrather U-shaped. Salaberry’s results show that less proficient learners are gen-erally less influenced by prototypes than more proficient learners. In addition,lower level learners show greater preference for nonprototypical pairings thanmore advanced learners.

Labeau’s (2005) French L2 results show support for Salaberry’s findings.She also used Modern Times to elicit viewpoint marking in written and spokennarratives. Her data are based on a corpus of 61 English-speaking universitylearners of French divided into three groups based on level of instruction: Year1 (n=21), Year 2 (n=17), and Year 4 (n=22). She notes that uses of the PC and



1-iral-51-3 — 2013/7/10 11:37—308— #67—ce

308 Kevin McManus

Table 6. Tense selection according to situation type in per cent in French L2 in Labeau(2005)

Statives Telic

IMP PC IMP PC

Year 1 (n=21) 58.8 3.9 25.9 75.8Year 2 (n=17) 79.3 1.4 11.3 79.1Year 4 (n=23) 80.3 2.4 5.1 80.2

the IMP are ‘unequally distributed across lexical categories at all levels, whichseems to support the AH inasmuch as it predicts preferential combinations’(Labeau 2005: 144). Labeau’s results for the spoken narratives are shown inTable 6 (adapted from Labeau 2005: 149).

In Labeau’s oral data (Table 6), group results show that prototypicality in-creases with proficiency. The proportions for prototypical combinations be-tween groups are indeed very similar: less than 5 % difference between groupsfor the PC with telic situation types. For nonprototypicality, the results areagain low, but they indicate less use of nonprototypical combinations as profi-ciency increases. These scores suggest that Year 1 learners use more nonpro-totypical combinations than more advanced learners, as observed by Salaberry(1999): nonprototypicality reduces with increases in proficiency. In obligatoryperfective contexts, Labeau notes that learners’ use of the PC increases withproficiency. Use of the PC is generally high, even amongst the Year 1 learners(74 %). In obligatory imperfective contexts, use of the IMP differs betweenlearner groups: for instance, in the spoken narratives, use of the IMP is as fol-lows (from Year 1 to Year 4): 55.2 % → 76.4 % → 81.4 %. She concludesthat ‘the proportion of IMP increases with proficiency, but falls short of nativelevels’3 (Labeau 2005: 182).

Studies investigating the extent to which L2 development is influenced byprototypical combinations appear to suggest mixed evidence. Findings thatprototypical influence is not strongest at the early stages of development arefound by Labeau (2005) and Salaberry (1999), amongst others. These samestudies also show that nonprototypicality fails to increase consistently withproficiency. An emerging pattern of L2 development appears to be that pro-totypical combinations increase with proficiency. Ayoun and Salaberry (2005:269) similarly note such trends in the literature and suggest that ‘the develop-

2. The learner corpus also contains four bilinguals: two English-French bilinguals and oneEnglish-German bilingual in the Year 4 group and one English-German bilingual in the Year1 group.

3. Labeau’s study does not incorporate an independent measure of proficiency.



1-iral-51-3 — 2013/7/10 11:37—309— #68—ce


mental picture of tense-aspect marking originally proposed by Andersen needsto be reviewed and possibly revised’.

3. Research questions and predictions

In this paper, the influence of semantic prototypes on the development of as-pect in a L2 is investigated by examining the production and interpretation ofviewpoint aspect forms by English- and German-speaking university learnersof French L2 at two significantly different levels of proficiency. Specifically,this study aims to address the following research questions:– What role do semantic prototypes have in the L2 development of aspect?– What are the theoretical implications of semantic prototypes for L2 devel-

opment more generally?For the AH, L2 development is characterized by pairing semantic prototypesof viewpoint and situation type (as noted in Section 2.2). As prototypes are hy-pothesised to be a universal feature of human categorization, they are thereforeinnate and predetermined. Because the AH is a hypothesis based on universalsemantic prototypes, it does not make predictions between learners of differentL1 backgrounds. In addition, Andersen and Shirai (1994, 1996), propose thatL2 development is linear: L2 learners begin with prototypical combinationsand develop towards nonprototypical ones in defined stages:– Learners first use perfective marking [the PC] on achievement and accom-

plishment verbs, eventually extending its use to activities and stative verbs.– In languages that encode the perfective/imperfective distinction, imperfec-

tive past [the IMP] appears later than perfective past [PC], and imperfectivepast marking begins with stative verbs and activity verbs, then extends toaccomplishment and achievement verbs.

(Andersen and Shirai 1996: 533)By adopting the AH’s predictions, differences between learners can be pre-dicted for the acquisition of viewpoint in French L2. Firstly, differences arepredicted between learners of different proficiencies and not of different L1backgrounds. Therefore, English and German learners are predicted not to besignificantly different from each other at the same proficiency level. FollowingAndersen and Shirai’s (1996) formulation of the AH, the following predictionsfor the development of viewpoint aspect in French L2 are made:

Low proficiency learners will be influenced more by semantic prototypesthan high proficiency learners. They will use the PC more with telic than atelicsituation types and the IMP more with atelic than telic situation types. Forlow proficiency learners, use of the PC and the IMP will be influenced by thesituation type of the predicate



1-iral-51-3 — 2013/7/10 11:37—310— #69—ce

310 Kevin McManus

Due to increased L2 proficiency and amount of naturalistic exposure toFrench, high proficiency learners will show greater use of nonprototypical pair-ings than low proficiency learners. They will use the PC more with atelic situa-tion types than low proficiency learners. They will use the IMP more with telicsituation types than low proficiency learners. For high proficiency learners, useof the PC and IMP will be influenced less by the situation type of the predicate.

Therefore, whilst French native speakers’ use of the PC and the IMP will notbe influenced by a predicate’s situation type, it can be predicted that learnerswill be significantly influenced by prototypicality. Accordingly, in line with theAH’s predictions for L2 development, low proficiency learners’ tense selectionwill show greater prototypical influence than high proficiency learners.

4. The Study

This study involves 75 learners and a control group of ten native speakers ofFrench. All learners are either (1) English native speakers or (2) German nativespeakers. English-speaking learners were recruited and tested in England andGerman-speaking learners were recruited and tested in Germany.4 French na-tive speakers were all university students from France studying in England onan ERASMUS exchange programme. They were recruited and tested withintwo weeks of arriving in England. All learners were studying French as partof university degree (Bachelor of Arts in England and Magister Artium in Ger-many), recruited from two different levels of instruction: Year 1 and Year 4of a 4-year degree programme. As part of this study’s cross-sectional design,learners were recruited from two different levels of instruction and tested once.However, since learners from the same level of instruction are typically nothomogenous in their proficiency (e.g., Tremblay 2011), proficiency in FrenchL2 was independently measured by a C-Test. This proficiency measure wascomposed of five short texts (mean length of each text: 69 words) and requiredlearners to fill in the second half of every second word, following Daller etal. (2003) and Grotjahn and Eckes (2006). Each participant is scored out of123. Learners scored between 53 and 114 out of 123 with a mean score of83.77 (SD = 22.58). Learners were split into proficiency groups based on theirperformance in the test: learners scoring below the mean are placed in a ‘lowgroup’ (score range: 53–78) and learners scoring above the mean are placed inan ‘advanced group’ (score range: 92–114). These scorings led to four learnersgroups: two ‘low proficiency’ groups and two ‘advanced proficiency’ groups.

4. I am grateful to Richard Waltereit, Philipp Obrist and Max Grosse for helping me contact andrecruit German-speaking learners and to Eugene Stemp and Franck Michel for their assistancewith English-speaking learners.



1-iral-51-3 — 2013/7/10 11:37—311— #70—ce


Table 7. Learners’ background information

Low group Advanced group

English L1 German L1 English L1 German L1

N 19 19 19 18Mean age (years) 18.3 20.5 20.9 23.5Mean exposure to French(years)

7.9 7.9 11 11

Mean time spent in a Fran-cophone country (years)

0.1 0.3 0.5 0.6

Furthermore, each proficiency group was split between L1 English and L1 Ger-man speakers, giving rise to the following learner groups:– English low group (n=19)– German low group (n=19)– English advanced group (n=19)– German advanced group (n=18).A one-way analysis of variance (ANOVA) was carried out to compare the pro-ficiency scores between groups. ANOVA indicates a statistically significantmain effect for C-Test score [F(4,76) = 239.29, p < .001]. Post hoc com-parisons using the Tukey HSD test indicated that the English low group (M =59.47, SD = 6.63) and the German low group (M = 63.95, SD = 6.16) weresignificantly different from the English advanced group (M = 98.47, SD =6.63), the German advanced group (M = 101.89, SD = 6.58) and the controlgroup (M = 122.67, SD = .52). The English advanced group and the Germanadvanced group were also significantly different from the NS. No significantdifferences were found between (1) the English advanced group and the Ger-man advanced group and (2) the English low group and the German low group.

Although there are no significant differences between (1) the low groups and(2) the advanced groups for performance in the C-Test, some differences arepresent in terms of age and time spent in a Francophone country, as summa-rized in Table 7.

4.1. Data collection materials

All participants completed two tasks, as well as a background questionnaireand the C-Test. All tasks were carried out on a one-to-one basis with the sameresearcher. A picture-based spoken narrative was used as well as a sentence



1-iral-51-3 — 2013/7/10 11:37—312— #71—ce

312 Kevin McManus

interpretation task.5 Before data collection, all tasks were piloted with a groupof five native speakers and five university learners of French.

For the spoken narrative (Natalie et Albert), participants narrate a picturestory in the past with viewpoint aspect shifts contrasting perfective and im-perfective viewpoints. Natalie et Albert is about a girl (Natalie) and her petcat (Albert). The story begins with the pair’s daily routine, describing typicalevents in their lives. This part of the story requires the imperfective viewpoint(i.e., the IMP). The storyline switches to a one-time event when Albert sawa dog in the street and ran away. This narrative shift to perfective viewpointrequires the PC. In this task, not only are there perfective-imperfective view-point contrasts, but the different viewpoints are combined with telic and atelicsituation types. Natalie et Albert (picture narrative) differs from Modern Times(silent film) because it was designed to balance semantically prototypical andnonprototypical pairings. As noted by Bardovi-Harlig (2000: 201), ModernTimes lacks such a balance because ‘certain types of predicates occur more fre-quently than others’. As a consequence, the predicate bias inherent to ModernTimes prevents a balanced investigation into prototype effects in L2 develop-ment.

The Sentence Interpretation task tests the meaning implications learners as-sign to sentences. It contains 31 test items. Participants are presented with atest item containing (i) a written context in the participant’s L1 and (ii) twoFrench sentences to rate. Participants are required to rate each French sen-tence in terms of how appropriately it describes the L1 context. One sentenceis perfective and one is imperfective. In each test item, only one sentence isappropriate. Each sentence is rated from inappropriate to appropriate, using athree-point Likert scale: −1 (inappropriate) 0 (don’t know) +1 (appropriate).The Sentence Interpretation task uses both perfective and imperfective view-points across situation types. For imperfective sentences both progressive andhabitual meanings are tested. Test sentences for the task are selected by com-bining different viewpoint types (perfective and imperfective) with differentsituation types (telic and atelic). For example:– Perfective and telic: (lire) le livre ‘read the book’– Perfective and atelic: (avoir) besoin d’aide ‘need help’– Imperfective and telic: (vendre) une guitare ‘sell a guitar’– Imperfective and atelic: (manger) dans le parc ‘eat in the park’Therefore, as in Natalie et Albert, the Sentence Interpretation task is designedto specifically investigate both prototypical (e.g., perfective and telic, ‘readthe book’) and nonprototypical (e.g., imperfective and telic, ‘sell a guitar’)

5. The tasks used in this study were initially created by the SPLLOC 2 research team(Domínguez et al. 2009). They were adapted to French for the present study (for furtherdetails, see McManus, 2011).



1-iral-51-3 — 2013/7/10 11:37—313— #72—ce


combinations. Furthermore, the Sentence Interpretation task complements theproduction task by investigating the specific interpretations learners attribute toviewpoint forms.

5. Results

5.1. Spoken narrative results

Table 9 displays learners’ use of tense in obligatory perfective contexts (OPCs)and obligatory imperfective contexts (OICs) analysed according to a predi-cate’s situation type (telic vs. atelic). The most striking difference betweenparticipants is that NS only ever use the PC in OPCs and only ever use the IMPin OICs. In contrast, learners use a variety of tenses in each of the obligatorycontexts. In OPCs, all learners use the PC more than any other tense. Ger-man low group learners are the only participants to use the perfective PasséSimple6 in this same context. As for OICs, low group learners do not showclear preference for one tense. English low group learners indicate preferencefor the IMP (44.1 %) over the PC (27.3 %); however German low group learn-ers fail to show an equally clear preference for the IMP (35.3 %) over the PC(32.6 %). These findings are not replicated for advanced group learners, whoclearly show preference for the IMP (>71 %) over the PC (<13.3 %) in OICs.

A comparison of telic and atelic predicates with the PC and the IMP inOPCs and OICs in Natalie et Albert (Table 9) shows that learners use theIMP and the PC with both telic and atelic situation types. However, it ap-pears that advanced group learners show greater preference towards proto-typical pairings, especially for the IMP with atelic situation types. In con-trast, the control group’s use of tense does not appear to reflect prototypicalinfluence. These differences between groups were tested. Firstly, in OPCs,ANOVA indicates a statistically significant main effect for the IMP with atelicsituation types [F(4,548) = 11.78, p < .001]. Post hoc comparisons indicatedthat learners in the English low group (M = .258, SD = .439) and Germanlow group (M = .287, SD = .454) differed significantly from learners in En-glish advanced group (M = .429, SD = .497) and German advanced group(M = .426, SD = .497). Significant differences were also found between alllearners and the control group (p < .001). Secondly, in OICs, ANOVA in-dicates a statistically significant main effect for the IMP with atelic situation

6. The Passé Simple is rarely used in spoken French, but still exists in a number of discoursetypes (e.g. formal speeches, fairy tales). In contemporary usage, researchers have argued thatthe Passé Simple has been replaced by the PC in most discourse types (e.g., Labeau 2005,2009; Wilmet 2003).



1-iral-51-3 — 2013/7/10 11:37—314— #73—ce

314 Kevin McManus

Table 8. Tenses used in Natalie et Albert according to obligatory context and situationtype

Perfective Imperfective (habitual)

Telic Atelic Telic Atelic

% n % n % n % n

English LowPC 83.9 (161) 71.9 (95) 24.6 (14) 26 (27)IMP 13.5 (26) 25.8 (34) 45.6 (26) 42.3 (44)PRES 2.1 (4) 2.3 (3) 26.3 (15) 29.8 (31)Other 0.5 (1) 0 3.5 (2) 1.9 (2)

German LowPS 1.9 (4) 1.7 (2) 0 (0) 0 (0)PC 82.3 (177) 60.9 (70) 34.7 (26) 31.3 (36)IMP 8.8 (19) 28.7 (33) 37.3 (28) 33.9 (39)PRES 6.5 (14) 7.8 (9) 28 (21) 33.9 (39)Other 0.5 (1) 0.9 (1) 0 (0) 0.9 (1)

English AdvancedPC 91.6 (206) 53.2 (67) 17.4 (12) 11 (14)IMP 6.2 (14) 42.9 (54) 66.7 (46) 74.8 (95)PRES 2.2 (5) 3.9 (5) 15.9 (11) 14.2 (18)Other 0 (0) 0 (0) 0 (0) 0 (0)

German AdvancedPC 83.1 (207) 48.4 (59) 13.1 (11) 11.4 (14)IMP 8.9 (22) 42.6 (52) 69 (58) 75.6 (93)PRES 6.8 (17) 9 (11) 17.9 (15) 12.2 (15)Other 1.2 (3) 0 (0) 0 (0) 0.8 (1)

ControlPC 100 (192) 100 (97) 0 (0) 0 (0)IMP 0 (0) 0 (0) 100 (32) 100 (100)PRES 0 (0) 0 (0) 0 (0) 0 (0)Other 0 (0) 0 (0) 0 (0) 0 (0)

types [F(4,523) = 34.59, p < .001]. Post hoc comparisons indicated that En-glish low (M = .423, SD= .496) and German low (M = .342, SD= .477) grouplearners differed significantly from English advanced (M = .748, SD = .436)and German advanced (M = .756, SD = .431) group learners. Significant dif-ferences were also found between advanced group learners and the controlgroup (p < .05). Furthermore, ANOVA indicates a significant main effect forthe PC with telic situation types [F(4,299) = 4.67, p < .05]. Post hoc compar-



1-iral-51-3 — 2013/7/10 11:37—315— #74—ce


isons indicated that German low group (M = .347, SD= .479) learners differedsignificantly from German advanced group (M = .131, SD = .339) learners.

In sum, results from Natalie et Albert show that advanced and low grouplearners use the PC in OPCs more than any other tense. In OICs, advancedgroup learners use the IMP more than any other tense, whilst low group learn-ers fail to show clear preference for a single tense. With respect to prototyp-ical influence, statistically significant main effects for (i) the IMP with atelicsituation types and (ii) the PC with telic situation types were found betweengroups at different proficiency levels. Advanced group learners show greatereffects for prototypical influence than low group learners. This is most clearfrom the use of the IMP (see Table 9), where its use with atelic situation typesby advanced group learners is significantly different from low group learners(p < .001) and NS (p < .05).

5.2. Sentence Interpretation task results

Table 10 displays participants’ selection of OPCs and OICs. As found in Na-talie et Albert, NS always select the PC in OPCs and always select the IMP inOICs. Indeed in Sentence Interpretation task learners show these same trends,although to a less absolute extent. A comparison of telic and atelic predicateswith tenses in OPCs and OICs in the Sentence Interpretation task (Table 10) There is no Table 10 in

your file.shows that, as in Natalie et Albert, NS select the PC and the IMP independentlyof a sentence’s situation type. For learners, the results indicate prototypical in-fluence.

In OPCs, both low groups show greater use of the PC with telic than withatelic situation types and greater use of the IMP with atelic than with telic sit-uation types. Therefore, both low groups appear to show prototypical effectsin OPCs. These findings appear to be replicated for the advanced group learn-ers. As for OICs, prototypical effects appear minimal for low group learnersbut large for advanced group learners, consistent with findings in OICs fromNatalie et Albert. These differences between groups for prototypical influencewere tested. In OPCs, ANOVA indicates a statistically significant main effectfor the PC with telic situation types [F(4,319) = 5.35, p < .05]. Post hoccomparisons indicated that English low group learners (M = .803, SD = .401)differed significantly from English advanced group learners (M = .947, SD =.225). ANOVA also indicates statistically significant differences between lowgroup learners and the control group for the selection of the IMP with atelicsituation types in habitual contexts (p < .05). However, no statistically sig-nificant main effects for tense selection were found between learners in OICs.This result contrasts with findings from Natalie et Albert, where a statisticallysignificant main effect was found both for the PC with telic situation types and



1-iral-51-3 — 2013/7/10 11:37—316— #75—ce

316 Kevin McManusTa

ble

9.Te

nses

sele

cted

inth

eSe

nten

ceIn

terp

reta

tion

acco

rdin

gto

oblig

ator

yco

ntex

tand

situ

atio

nty

pe(p

erce

nt;r

awfig

ures

inpa

rent

he-

sis)

Perf

ectiv

eIm

perf

ectiv

e

Hab

itual

Prog

ress

ive

Telic

Ate

licTe

licA

telic

Telic

Ate

lic

%n

%n

%n

%n

%n

%n

Eng

lish

low

PC80

.3(6

1)46.1

(35)

10.5

(4)

42.1

(32)

33.3

(19)

14.7

(14)

IMP

19.7

(15)

53.9

(41)

89.5

(34)

57.9

(44)

66.7

(38)

85.3

(81)

Ger

man

low

PC89

.5(6

8)53.9

(41)

15.8

(6)

30.3

(23)

38.6

(22)

24.2

(23)

IMP

10.5

(8)

46.1

(35)

84.2

(32)

69.7

(53)

61.4

(35)

75.8

(72)

Eng

lish

adva

nced

PC94

.7(7

2)55.3

(42)

0(0

)26

.3(2

0)61

.4(3

5)7.

4(7

)IM

P5.

3(4

)44.7

(34)

100

(38)

73.7

(56)

38.6

(22)

92.6

(88)

Ger

man

adva

nced

PC98

.6(7

1)56.9

(41)

8.3

(3)

8.3

(6)

55.6

(30)

11.1

(10)

IMP

1.4

(1)

43.1

(31)

91.7

(33)

91.7

(66)

44.4

(24)

88.9

(80)

Con

trol

PC10

0(4

0)10

0(4

0)0

(0)

0(0

)0

(0)

0(0

)IM

P0

(0)

0(0

)10

0(2

0)10

0(4

0)10

0(3

0)10

0(5

0)



1-iral-51-3 — 2013/7/10 11:37—317— #76—ce


the IMP with atelic situation types.

6. Discussion

Predictions from the AH were adopted for the present study to account for therole of semantic prototypes in the L2 development of aspect (as presented inSections 2 and 3). The AH predicts a linear route of L2 development: proto-typical → nonprototypical. The AH’s predictions for the L2 development ofviewpoint aspect, as presented in Section 3, are as follows:– Learners first use perfective marking [the PC] on achievement and accom-

plishment verbs, eventually extending its uses to activities and stative verbs.– In languages that encode the perfective/imperfective distinction, imperfec-

tive past [the IMP] appears later than perfective past [the PC], and imperfec-tive past marking begins with stative verbs and activity verbs, then extendingto accomplishment and achievement verbs.

(Andersen and Shirai 1996: 533)In contrast to NS, the results show that in general learners’ use of tense is pro-totypically influenced by a predicate’s situation type. Learners select the PCsignificantly more with telic than atelic situation types, as predicted by the AH.The results from Natalie et Albert revealed a statistically significant main ef-fect for the PC with telic situation types between different proficiency levels(p < .05). This finding is also corroborated in the Sentence Interpretation task(p < .05). A statistically significant main effect for the IMP with atelic situ-ation types between different proficiency levels was also found in Natalie etAlbert (p < .001). However, this finding was not replicated in the SentenceInterpretation task. Where significant main effects for prototypical influencewere found, these showed that advanced group learners were affected signifi-cantly more by semantic prototypes than low group learners. This is particu-larly the case with IMP-atelic pairings in Natalie et Albert. Therefore, in con-trast to the AH, this study’s results suggest a route of development different tothe one predicted by the AH: prototypical influence correlates with increasedprototypical influence. In other words, as L2 proficiency increases so doesprototypical influence. Moreover, learners indicated significant prototypicalinfluence despite the tasks being specifically designed to balance prototypicaland non-prototypical pairings (in contrast to the Modern Times method).

However, why is it that advanced group learners are influenced significantlymore by prototypicality than low group learners? If there is a prototypical biasin the input (as suggested by Andersen and Shirai), then it is arguable thatprototypicality strengthens with proficiency because more proficient learnershave generally had more exposure to L2 input than less proficient learners.However, the use of prototypical combinations is entirely dependent on what



1-iral-51-3 — 2013/7/10 11:37—318— #77—ce

318 Kevin McManus

learners use tense for. If prototypicality influences L2 development as the AHproposes, then for PC-telic and IMP-atelic combinations to be made, it is nec-essary that (in French) perfectivity be mapped to the PC and imperfectivity bemapped to the IMP. This appears to be an essential condition for semanticallyprototypical combinations to be made, which is overlooked by the AH. Theconsistent use of the PC in OPCs and the IMP in OICs by advanced grouplearners in Natalie et Albert is mirrored by statistically significant prototypi-cal combinations in these obligatory contexts. Variability of tense use by lowgroup learners in OICs results in no statistically significant prototypical combi-nations in these same contexts. This study’s results provide evidence to supportthe claim that prototypicality influences L2 development once there is greaterstability in the viewpoint meanings mapped to tense (as measured by the useof tenses in obligatory contexts).

Although the AH predicts that prototypicality will initially influence L2 de-velopment, it can also be stated that the AH could be used to judge whetherlearners have mapped viewpoint aspect to tense. Advanced group learners useprototypical combinations in both OPCs and OICs (PC-telic and IMP-atelic)because they appear to have mapped perfectivity to the PC and imperfectivityto the IMP. In short, the route of L2 development as predicted by the AH isentirely dependent on the PC and the IMP marking viewpoint aspect. At nopoint, however, are L2 form-meaning mappings discussed by Andersen andShirai (1994, 1996). Shirai (2010: 172) claims that ‘although there is no con-sensus regarding the explanation for such semantic bias [prototypical combi-nations], this phenomenon has been attested in various languages’. It is arguedthat when L1-L2 form-meaning differences for viewpoint aspect are analysed,then clear predictions for prototypical influence can be made. Crucially, as thepresent study’s results suggest, the use of prototypical combinations is entirelydependent on the mapping of viewpoint aspect to tense.

Furthermore, Andersen and Shirai (Andersen and Shirai 1994, 1996, Shirai2009, 2010) see the AH as a theory of L2 development rooted in part on se-mantic prototypes and in part on input frequency prototypes. That said, theAH founders seem to argue that prototypical combinations are largely due tothe frequency of prototypes in the input:

In earlier work, I have proposed prototype formation based on distributional learn-ing as an explanation for the semantic bias in early tense-aspect morphology, andin other aspects of grammatical development [. . . ] In short, the semantic biascomes from biased frequency distribution in the input and learners’ prototype for-mation based on such biased input. (Shirai 2010: 184–186)

It is plausible (although as yet unverified) to assume that semantic prototypi-cality is more frequent in the input than semantic nonprototypicality, thereforestudies (such as this one) find correlations between semantic prototypicality



1-iral-51-3 — 2013/7/10 11:37—319— #78—ce


and high L2 proficiency. Indeed Hendricks (1999) also argues that prototypi-cal combinations reflect a distributional bias in the input. However, it is opento debate to what extent ‘biased frequency distribution in the input’ can ac-count for this apparently universal route of L2 development, especially as thisroute of inquiry has not been actively investigated. This is especially the casewhen learners’ exposure to the L2 can vary substantially (Gass 2003, Gassand Mackey 2007, VanPatten et al. 2004). In which case, learners will notnecessarily be exposed to the same prototypical combinations. Despite such alarge number of studies on the AH (for an extensive review, see Bardovi-Harlig2000), researchers are no wiser as to why prototypicality strengthens as profi-ciency increases. The link between frequency and semantic prototypicality isarguably very relevant for such an understanding. However, this link is oftenimplicitly assumed rather than empirically verified. A relevant correlation inthis study is the difference between learner groups in terms of their declaredexposure to French in a naturalistic setting. Advanced group learners declaredmore time spent abroad than low group learners and show significantly greaterprototypical effects. The present study’s results tentatively point to a close re-lationship between L2 proficiency, prototypicality and input. Further researchis required on the connection between input and the use of prototypical combi-nations.

There is also empirical support that L2 development with reference to proto-typicality is not linear (e.g., Bardovi-Harlig and Bergström 1996; Labeau 2005;Salaberry 1999, 2008). Perhaps the low group learners in this study are tooproficient to show stronger prototypical combinations than the advanced grouplearners, in which case L2 development could be hypothesised to be more U-shaped than linear (Kellerman 1985, Lightbown 1985, Lightbown 2003, Long1990). Indeed, findings suggesting that the AH fails to reliably typify L2 de-velopment appears to be acknowledged by Shirai (2004):

[The AH] is supported by most studies, but its developmental component – namely,the prediction that beginning learners are more restricted by inherent aspectualvalue than more advanced learners – may need to be revised. (Shirai 2004: 106)

The present study offers empirical support for a revision of the AH’s ‘devel-opmental component’. However, Shirai (2004) also argues that the extent towhich the AH supports L2 development is dependent on a particular study’smethodology. In contrast to many studies on the L2 development of aspect,the present study avoided Modern Times because it has been reported to failin balancing prototypical and nonprototypical combinations, instead favouringprototypical ones (Bardovi-Harlig 2000; Bardovi-Harlig and Bergström 1996).As a consequence, when Modern Times is avoided (e.g., Domínguez et al., inpress) or combined with other methods (e.g., Labeau 2005), then maybe it is



1-iral-51-3 — 2013/7/10 11:37—320— #79—ce

320 Kevin McManus

no surprise that the AH’s predictions for prototypical influence fail to be sup-ported.

7. Conclusion

The present study points to a route of L2 development different to the one pre-dicted by the AH: the most proficient learners (advanced group learners) appearto be influenced significantly more by prototypicality than less proficient learn-ers (low group learners). Furthermore, low group learners show greater useof nonprototypical pairings than advanced group learners. In fact, this studyhas indicated a route of L2 development opposite to the one predicted by theAH. This different route of L2 development (i.e., increased use of prototypicalpairings alongside increased L2 proficiency) may be explainable by consider-ing what learners use tense for. Unless viewpoint aspect has been mapped totense in the L2, it remains difficult to see how viewpoint and situation typeprototypes can be paired (although this is the assumption made by the AH).This reasoning explains why low group learners show prototypical influence inOPCs but not in imperfective OICs: perfectivity has been mapped in the L2 butimperfectivity has not. It therefore becomes much clearer to understand whyadvanced group learners are prototypically influenced in both OPCs and OICs:both perfectivity and imperfectivity have been mapped in the L2. Thereforeviewpoint mapping in the L2 appears to be an essential condition for proto-typical influence to take effect (as demonstrated in the advanced group data).However, more research on form-meaning mapping with both very similar andvery different L1-L2 combinations and at very different proficiency levels isrequired in order to further substantiate this claim.

University of [email protected]

ReferencesPlease re-submit thissection in the IRAL format(see IRAL style sheet.) Andersen, Roger W. 1991. Developmental sequences: The emergence of aspect marking in second

language acquisition. In Crosscurrents in Second Language Acquisition and Linguistic Theo-ries, eds. Thom Hübner and Charles A. Ferguson, 305–324. Amsterdam: John Benjamins.

Andersen, Roger W., and Shirai, Yasuhiro. 1994. Discourse motivations for some cognitive acqui-sition principles. Studies in Second Language Acquisition 16:135-156.

Andersen, Roger W., and Shirai, Yasuhiro. 1996. Primacy of aspect in first and second languageacquisition: the pidgin/creole connection. In Handbook of Second Language Acquisition, eds.William C. Ritchie and Tej K. Bhatia, 527-570. San Diego, CA: Academic Press.

Antinucci, Francesco, and Miller, Ruth. 1976. How children talk about what happened. Journalof Child Language 3:169-189.

mailto:[email protected]



1-iral-51-3 — 2013/7/10 11:37—321— #80—ce


Arche, María J. 2006. Individuals in Time. Tense, aspect and the individual/stage distinction.Amsterdam/Philadelphia: John Benjamins.

Ayoun, Dalila, and Salaberry, M. Rafael. 2005. Towards a Comprehensive Model of the Ac-quisition of L2 Tense-aspect in the Romance Languages. In Tense and Aspect in RomanceLanguages, eds. Dalila Ayoun and M. Rafael Salaberry, 253-281. Amsterdam/Philadelphia:John Benjamins.

Bardovi-Harlig, Kathleen, and Bergström, Anna. 1996. Acquisition of tense and aspect in secondlanguage and foreign language learning: learner narratives in ESL and FFL. The CanadianModern Language Review / La Revue canadienne des langues vivantes 52:308-329.

Bertinetto, Pier Marco. 2001. On a frequent misunderstanding in the temporal-aspectual domain:‘The Perfective = Telic Confusion’. In Semantic Interfaces: Reference, Anaphora and Aspect,eds. Carlo Cecchetto, Gennaro Chiecrchia and Maria Teresa Guasti, 177-210. Stanford: CSLIPublications.

Bohnemeyer, Jürgen, and Swift, Mary. 2004. Event realization and default aspect. Linguistics andPhilosophy 27:263-296.

Bronckhart, Jean Paul, and Sinclair, Hermina. 1973. Time, tense, and aspect. Cognition 2:107-130.

Comajoan, Llorenç. 2006. The aspect hypothesis: Development of morphology and appropriate-ness of use. Language Learning 56:201-268.

Comrie, Bernard. 1976. Aspect. An Introduction to the Study of Verbal Aspect and Related Prob-lems. Cambridge: Cambridge University Press.

Depraetere, Ilse. 1995. On the necessity of distinguishing between (un)boundedness and (a)telicity.Linguistics and Philosophy 18:1-19.

Domínguez, Laura, Tracy-Ventura, Nicole, Arche, María J., Mitchell, Rosamond, and Myles, Flo-rence. in press. The role of dynamic contrasts in the L2 acquisition of Spanish past tensemorphology. Bilingualism: Language and Cognition.

Eckes, Thomas, and Grotjahn, Rudiger. 2006. A closer look at the construct validity of C-tests.Language Testing 23:290-325.

Gass, Susan M. 2003. Input and Interaction. In Handbook of Second Language Acquisition, eds.Catherine. J. Doughty and Michael. H. Long, 224-255. Oxford: Blackwell.

Gass, Susan M., and Mackey, Alison. 2007. Data Elicitation for Second and Foreign LanguageResearch. Mahwah, N.J.: Lawrence Erlbaum Associates.

Kellerman, Eric. 1985. If at first you don’t succeed... In Input in Second Language Acquisition,eds. Susan M Gass and Carolyn Madden, 345-353. Rowley, MA: Newbury.

Labeau, Emmanuelle. 2005. Beyond the Aspect Hypothesis: Tense-Aspect Development in Ad-vanced L2 French. Oxford: Peter Lang.

Labeau, Emmanuelle. 2009. Le PS: cher disparu de la rubrique nécrologique? Journal of FrenchLanguage Studies 19:61-86.

Lightbown, Patsy M. 1985. Input and acquisition for second-language learners in and out of class-rooms. Applied Linguistics 6:263-273.

Lightbown, Patsy. M. 2003. SLA research in the classroom/ SLA research for the classroom.Language Learning Journal 28:04-13.

Long, Michael H. 1990. The least a second language acquisition theory needs to explain. TESOLQuarterly 24:649-666.

McManus, Kevin. 2011. The Development of Aspect in a Second Language. Unpublished PhDthesis: Newcastle University.



1-iral-51-3 — 2013/7/10 11:37—322— #81—ce

322 Kevin McManus

Quirk, Charles Randolph, and Greenbaum, S. 1972. A University Grammar of English. London:Longman.

Rosch, Elaine. 1973. On the internal structure of perceptual and semantic categories. In CognitiveDevelopment and the Acquisition of Language, ed. Timothy E. Moore, 111-144. New York:Academic Press.

Rosch, Elaine, and Mervis, Carolyn. 1975. Family resemblances: studies in the internal structureof categories. Cognitive Psychology 7:573-605.

Rosch, Elaine. 1978. Principles and categorization. In Cognition and Categorization, eds. ElaineRosch and Barbara Lloyd, 27-48. Hillsdale, NJ: Lawrence Erlbaum.

Salaberry, M. Rafael. 1999. The development of past tense verbal morphology in classroom L2Spanish. Applied Linguistics 20:151-178.

Salaberry, M. Rafael. 2008. Marking Past Tense in Second Language Acquisition: A TheoreticalModel. London: Continuum.

Shirai, Yasuhiro. 2004. A multiple-factor account to form-meaning connections in the acquisitionof tense-aspect morphology. In Form-meaning Connections in Second Language Acquisition,eds. Bill VanPatten, Jessica Williams, Susan Rott and Mark Overstreet, 91-112. Mahwah, NJ:Erlbaum Associates.

Shirai, Yasuhiro. 2009. Temporality in first and second language acquisition. In The Expression ofTime, eds. Wolfgang Klein and Ping Li, 167-194. Berlin: Mouton de Gruyter.

Shirai, Yasuhiro. 2010. Semantic bias and morphological regularity in the acquisition of tense-aspect morphology. What is the relation? Linguistics 48:171-194.

Slabakova, Roumyana. 2002. Recent research on the acquisition of aspect: an embarrassment ofriches? Second Language Research 18:172-188.

Smith, Carlota S. 1997. The Parameter of Aspect (2nd Edition). Dordrecht Kluwer Academic.Smith, Carlota S. 2006. The pragmatics and semantics of temporal meaning. In Proceedings of the

2004 Texas Linguistics Society Conference, eds. Pascal Denis, Eric McCready, Alex Palmerand Brian Reese, 92-106. Sommerville, MA: Cascadilla Press.

Sonnenhauser, Barbara. 2008. Aspect interpretation in Russian - a pragmatic account. Journal ofPragmatics 40:2077-2099.

Tremblay, Annie. 2011. Proficiency assessment standards in second language acquisition research:"Clozing” the gap. Studies in Second Language Acquisition 333:339-372.

VanPatten, Bill, Williams, Jessica, and Rott, Susan. 2004. Form-meaning connections in sec-ond language acquistion. In Form-Meaning Connections in Second Language Acquisition,eds. Bill VanPatten, Jessica Williams, Susan Rott and Mark Overstreet, 1-26. Mahwah, NJ:Lawrence Erlbaum Associates.

Vendler, Zeno. 1957. Verbs and times. The Philosophical Review 66:143-160.Vendler, Zeno. 1967. Linguistics in Philosophy. Ithaca: Cornell University Press.Wilmet, Marc. 2003. Grammaire critique du français. Brussels: Duculot.

Documents

Preliminary page and line breaksmcmanuskevin.com/wp-content/uploads/2018/09/iral-2013-0013.pdf · Imparfait German Perfekt Imperfekt and situation were observed in second language