Research in Second Language Acquisition : Empirical Evidence across Languages [1 ed.] 9781443810951, 9781443809610

Citation preview

Research in Second Language Acquisition

Research in Second Language Acquisition: Empirical Evidence across Languages

Edited by

Jörg-U. Keßler and Dagmar Keatinge

Research in Second Language Acquisition: Empirical Evidence across Languages, Edited by Jörg-U. Keßler and Dagmar Keatinge This book first published 2009 Cambridge Scholars Publishing 12 Back Chapman Street, Newcastle upon Tyne, NE6 2XX, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2009 by Jörg-U. Keßler and Dagmar Keatinge and contributors All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-4438-0961-6, ISBN (13): 978-1-4438-0961-0

CONTENTS About this Book Jörg-U. Keßler & Dagmar Keatinge......................................................... 1 Part I: Second Language Processing: Contributions to Theory Development The development of plural marking and plural agreement in child English L2 acquisition Yumiko Yamaguchi .................................................................................... 9 Development of discourse functions in Japanese and English bilingual first language acquisition Yuki Itani-Adams ..................................................................................... 41 The acquisition of the passive voice in English as a foreign language: production and perception Dagmar Keatinge & Jörg-U. Keßler....................................................... 69 Acquiring the Passive Voice: Online Production of the English Passive Construction by Mandarin Speakers Kenny Wang ............................................................................................ 95 Part II: Second Language Grammars across Languages Morphological development in Swedish learners of French: Discussing the processability perspective Malin Ågren........................................................................................... 121 Postverbal subject in Italian L2 – a Processability Theory approach Camilla Bettoni, Bruno Di Biase & Elena Nuzzo.................................. 153 Applying Processability Theory and its Extension to Serbian as a Family and Community Language in Australia Lucija Medojevic ................................................................................... 175 Acquisition of non-canonical order in Japanese as a second language: The case of causative structure Satomi Kawaguchi................................................................................. 213 Case, Word Order and Thematic Role Elena Ossella......................................................................................... 241 Profiling child ESL acquisition: Practical and methodological issues Kristin Kersten ...................................................................................... 267 Appendix Contributors........................................................................................... 295

ABOUT THIS BOOK

1 Introduction Since Pienemann’s first publication on Processability Theory (henceforth PT) in 1998, PT has become an established theory in the field of second language acquisition research and the publication of Pienemann (ed.) (2005) marked an extension to the existing theory, including the newer models of Lexical- Functional Grammar (Bresnan 2001) and its implications to second language development within the theory. Since the publication of the 2005- volume, various projects have been conducted to apply the extensions made to Processability Theory to a number of new theoretical issues, typologically different languages and pedagogical contexts. Additionally, the scope of PT has been widened to the application of the theory to the field of Applied Linguistics and Language Teaching. All this research has added additional support to the theoretical implications and the explanatory power provided in Pienemann 1998 and 2005. This volume is a follow- up volume to Mansouri (2007) Second Language Acquisition Research. Theory Construction and Testing and Keßler (2008) Processability Approaches to Second Language Development and Second Language Learning both also published by Cambridge Scholars Publishing, Newcastle upon Tyne. Like the latter, the contributions in this volume focus on the application of PT to various settings in second language acquisition research. They provide an overview of the manifold approaches taken to investigate L2 acquisition within the PT framework and combine a more theoretical approach in order to further extend the theory with a number of studies utilizing PT to further investigate bilingual language acquisition and language development in natural and institutional settings.

2

About this Book 

The structure of this book Taking into consideration the various aspects of research within the PT framework, this book is organised in two parts. Part 1 “Second Language Processing: Contributions to Theory Development” contains a number of papers discussing the inclusion of further theoretical aspects into PT, focusing on English as a second language. In Part 2 “Second Language Grammars across Languages”, PT is applied to a number of typologically different languages and contexts.

Part 1: Second Language Processing: Contributions to Theory Development A lot of research within the PT framework focuses on the development of English as a second or foreign language. Part 1 includes four papers working on the inclusion or extension of further theoretical aspects into PT. Leading on from her results in Yamaguchi (2008), Yamaguchi explores the development of plural marking and plural agreement in the context of child English L2 acquisition and provides another longitudinal study confirming the hierarchical nature of L2 processing. Itani- Adams works on simultaneous bilingual language acquisition within a PT context. Her paper adds support to the separate development hypothesis (deHouwer 1990) by utilizing the PT stages as an instrument to describe different developmental paths in the languages acquired by her informant. Keatinge and Keßler present a new extension to the PT framework. Their study reports on the inclusion of ESL interlanguage structures of the passive voice into the framework. Their findings confirm previous research (Pienemann 2005) by showing that the passive voice develops at a later point in time in the acquisition process and they provide an overview of interlanguage development of the passive voice in English. In addition to this study, Wang presents results from an online production study, also exploring the acquisition of the passive voice in English by speakers of Mandarin Chinese. His findings support the results of Keatinge and Keßler.

Jörg-U. Keßler & Dagmar Keatinge

3

Part 2: Second Language Grammars across Languages The second part of this volume examines interlanguage development and a number of morpho- syntactic structures across a variety of typologically different L2s. Ågren reports on her findings of morphological development in Swedish learners of French. Her study features a new innovative method of diagnosing language development online as well as one of the first studies which investigate the interlanguage development of French as L2 within a PT framework. Bettoni et al. analyse the development of a particular Italian construction, namely the postverbal plural and add a new feature to the language specific developmental hierarchy of Italian as L2. Medojevic’s study features a first attempt to apply PT to Serbian. Her study investigates the acquisition of Serbian as a family and community language by Serbian/ Australian bilingual teenagers and provides empirical evidence for the interlanguage development of Serbian. Kawaguchi explores the acquisition of causative structures in L2 Japanese by native speakers of English and discusses the inclusion of this particular structure into the PT framework. Her findings confirm previous research on this particular structure, showing that causative constructions – like the passive voice – are acquired at a later point in time of the acquisition process. Ossella investigates case, word order and thematic role in German as a foreign language by native speakers of Italian. Kersten’s study provides an insight into further methodological issues of using linguistic profiling to investigate L2 development.

Bibliography Bresnan, J. (2001). Lexical- Functional Syntax. Malden, MA: Blackwell. deHouwer, A. (1990). The acquisition of two languages from birth. A case study. Cambridge: CUP. Kawaguchi, S. (2005): “Argument structure and syntactic development.” In: Pienemann, M. (ed.):, pp. 253 -298. Keßler, J.-U. (ed.) (2008): Processability Approaches to Second Language Development and Second Language Learning. Newcastle upon Tyne: Cambridge Scholars Press. Mansouri, F. (ed.) (2007): Second Language Acquisition Research. Theory-Construction and Testing. Newcastle upon Tyne: Cambridge Scholars Press.

4

About this Book 

Pienemann, M. (1998): Language Processing and Second Language Development. Processability Theory. Amsterdam: John Benjamins. Pienemann, M. (ed.) (2005): Crosslinguistic Aspects of Processability Theory, Amsterdam, John Benjamins.

Jörg-U. Keßler & Dagmar Keatinge

5

Acknowledgements The contributions to this book are mostly based on papers presented at the 7th symposium on Processability Approaches to Language Acquisition Research and Teaching (PALART), hosted by the Universities of Newcastle and Paderborn in Newcastle upon Tyne in September 2007. The symposium and this volume would not be possible without the support of the core group of international researchers on Processability Theory. We would like to thank Manfred Pienemann (Paderborn and Newcastle), Bruno Di Biase and Satomi Kawaguchi (Sydney), Gisela Hakansson (Lund) and Camilla Bettoni (Verona) for their helpful discussions, support and advice during the symposium and the process of compiling this volume. We would also like to thank Benjamin AdeThurow and Patrice Hübsch (both Ludwigsburg), Miriam Hasse and Sybille Wills (both Paderborn) for their help in style editing and formatting the chapters of this book. Taking up a tradition from the 2008 volume (Keßler 2008) we asked an artist to create a cover art for the book. We are grateful to Verena Wendel (Ludwigsburg) who designed the cover art especially for this volume. Her print "mnİmonikȐ" (lat.) is extracted from an Intaglio series broaching the issue of Language Acquisition by illustrating its order and disorder. We would also like to say thank you to Camilla Field (Paderborn) for her support in proof reading and style editing of this volume. Special thanks go to Amanda Millar and Carol K. from Cambridge Scholars Publishing for their support and patience during the making of this compilation. Ludwigsburg and Paderborn, March 2009 Jörg-U. Keßler Dagmar Keatinge



PART I SECOND LANGUAGE PROCESSING: CONTRIBUTIONS TO THEORY DEVELOPMENT





THE DEVELOPMENT OF PLURAL MARKING AND PLURAL AGREEMENT IN CHILD ENGLISH L2 ACQUISITION

1

YUMIKO YAMAGUCHI

This chapter aims to examine grammatical development, focusing on morphological plural marking -son nouns and NP (Noun Phrase) plural agreement, in child English as a second language (L2) acquisition based on Processability Theory (PT; Pienemann, 1998, 2005). Plural marker -s on nouns is predicted to emerge at stage 2 in PT, since it only requires lexical procedure, namely, form variation of one constituent. In contrast, NP plural agreement is claimed to occur at stage 3, as it requires phrasal procedure, that is, exchange of grammatical information between two constituents. However, the developmental sequence of English L2 plural marking and plural agreement predicted in PT has been recently challenged by claiming that phrasal plural agreement with numerals emerges before lexical plural marker -s (e.g., Charters and Jansen, 2007). Hence, the current study addresses this issue with fresh evidence from a two-year longitudinal case study of a Japanese L1 primary school aged (5;8 to 7;8) child acquiring English as L2 in Australia, which is work in progress for my PhD thesis. In order to elicit the child speech production, several tasks, such as semi-structured interviews, narratives, and communication games, were performed in this study. Distributional analysis of plural marking in all contexts is carried out and compared with previous English L2 studies (e.g., Pienemann, 1998, 2005). The results demonstrate that plural marker -s on nouns (e.g., books) is acquired before NP plural agreement with numerals (e.g., two books) or other quantifiers 1 An early version of this paper was presented at the 7th International Symposium on Processability, Second Language Acquisition and Bilingualism, held in Newcastle, U.K., 10 September, 2007. I would like to thank Bruno Di Biase, Satomi Kawaguchi, and the participants of the symposium for their suggestions and comments.

10

The development of plural marking and plural agreement in child English L2 acquisition

(e.g., a lot of books), as predicted in PT. In addition, NP plural agreement is found to occur with numerals before it appears with other quantifiers.

1. Introduction This chapter aims to investigate the developmental sequence of morphological plural marker -son nouns and NP (Noun Phrase) plural agreement in Japanese L1 child learner acquiring English L2 acquisition within the framework of Processability Theory (PT; Pienemann, 1998, 2005). PT (Pienemann, 1998, 2005) assumes that L2 learners can produce only those linguistic forms for which they have acquired the necessary processing prerequisites and predicts which grammatical structures can be processed by the L2 learner at a given level of development, based on Levelt’s (1989) speech model and Lexical-Functional Grammar (LFG; Kaplan and Bresnan, 1982; Bresnan, 2001). Processability hierarchy for L2 morphology and syntax predicated in PT have been supported by empirical evidence from recent studies in various language contexts, such as Italian, Japanese, German, Arabic, English, Swedish and Chinese (e.g., Di Biase & Kawaguchi, 2002; Keßler, 2006, Mansouri, 2005; Pienemann, 1998; Pienemann & Håkansson, 1999, Zhang, 2005). However, the original processability hierarchy for English L2 was based on the findings from only cross-sectional studies (Johnston, 1985; Pienemann and Mackey, 1993). Moreover, the developmental sequence of plural marker -s on nouns and NP plural agreement in English L2 has been challenged by recent research (e.g., Charters and Jansen, 2007). The present chapter attempts to address this issue with fresh evidence from a longitudinal study of Japanese L1 child learner acquiring English L2, focusing on the developmental sequence of plural marker -s and NP plural agreement. In particular, more detailed analysis of the development of NP plural agreement is carried out by categorizing NP plural agreement into two types, such as numerical and non-numerical agreement. PT currently does not distinguish between these two types of NP plural agreement. In the following, plural marking systems in English and Japanese are briefly described first. Then, previous findings in English L2 studies on the acquisition of plural morpheme are summarised. Next, Processability Theory (PT; Pienemann, 1998, 2005) is outlined, as well as the findings in previous PT-based studies on English L2 morphology. Then, the current study, including research methods and the findings, is presented. Finally, the findings of the study are discussed based on PT.

Yumiko Yamaguchi

11

2. Plural marking in English and Japanese In English, there are a number of morpho-syntactic properties that reflect whether we are speaking about one thing or more than one thing (e.g., Ferenz and Prasada, 2002). As for regular English nouns, plural can be formed by adding a plural marker -s (or sometimes -es) (e.g., cats, dogs, books, boxes). In contrast, plural marking is not obligatory in Japanese and occurs only when the speaker wants to draw the listener’s attention to the quantity (e.g., Corbett, 2000, Maynard, 1990). The following is an example in which the plurality is not clear. (1)

Heya ni neko ga room in cat SUBJ ‘There is a cat/are some cats in the room.’

iru be

In order to specify quantity, Japanese employs a set of counters (e.g., Maynard, 1990). While the number of neko ‘cat’ is not specified in example (1), large quantities can be expressed by adding the quantifier, takusan ‘much, many’, after the subject marker, ga, as in example (2). (2)

Heya ni neko ga takusan room in cat SUBJ many ‘There are many cats in the room.’

iru be

Also, a suffix -tachi can be used to make plural forms of personal pronouns (e.g., watashi-tachi ‘we’) and other nouns referring to persons (e.g., otona-tachi ‘adults’) (e.g., Maynard, 1990). In example (3), “more than one cat” is emphasized with this suffix -tachi.. (3)

Heya ni neko-tachi ga room in cat-PLURAL SUBJ ‘There is more than one cat in the room.’

iru be

In addition, there are three other suffixes, such as -domo, -gata, and -ra, which can be used as plural markers for personal pronouns or human nouns 2 . In Japanese, whether or not NP head type is animated affects the choice between marking plurality and not marking it. Downing (1996) 2

Please see Makino and Tsutui (1986) for further explanations about these suffixes.

12

The development of plural marking and plural agreement in child English L2 acquisition

summarises the factors which affect the choice between marking plurality and not marking it, as in Table 1.

NP head type

referent type pronoun proper noun common noun

human required required possible

other animate required rare rare

inanimate required impossible impossible

Table 1: The use of plural markers in Japanese (after Downing, 1996: 205)

3. The acquisition of English plural morpheme The English plural morpheme -s is one of the first bound morphemes acquired by English L1 children (Jia, 2003). The results in English L1 studies (e.g., Brown, 1973; de Villiers and de Villiers, 1973) suggest that “most normally developing children master the plural morpheme no later than the 3.0 to 3.5 mean length of utterance (MLU) range or 29-33 months of age, as defined by 80% or 90% correct use in obligatory contexts across two or three consecutive testing sessions” (Jia, 2003, p.1297). In English L2 studies (e.g., Bialystok and Miller, 1999; Flege, YeniKomshian, and Liu, 1999; Jia, Aaronson, and Wu, 2002), it was shown that the native-level mastery of the plural morpheme was hardly achieved by immigrants, even after many years of exposure to English. In particular, studies on speakers of Asian languages, which lack the morphology system for plurality (e.g., Flege et al, 1999; Jia et al, 2002), found that plural morpheme was one of the hardest grammatical structures to acquire. With regard to studies on Japanese L1 speakers, Hakuta (1976, 1978) examined a 5 year old girl acquiring English L2 in the United States over 7 months. In his study, the acquisition points of grammatical morphemes were defined as the first of three consecutive two-week samples in which the morphemes were supplied in over 90% of obligatory contexts. He found that the Japanese child learner, who was exposed to English L2 at 5 years old, achieved only 60% accuracy of the plural morpheme use in the end of his 7 month longitudinal study. He claimed that the reason why the plural morpheme appeared relatively later than other grammatical morphemes, such as progressive -ing, in the Japanese child learner’s English L2 acquisition was that “the notion of plurality (number) does not exist in the Japanese grammar” (Hakuta, 1978, p. 143). More recently, Jia (2003) conducted a 5 year longitudinal study on the acquisition of the English plural morpheme by 10 Chinese L1 children,

Yumiko Yamaguchi

13

who immigrated to the United States between ages 5 and 16. Jia (2003) used the criterion of 80% correct use of the morpheme in obligatory contexts across three consecutive testing sessions based on L1 literature (Lahey, Liebergott, Chesnick, Menyuk, and Adams, 1992) and demonstrated that only 7 of these 10 learners mastered the plural morpheme after 5 years of English exposure. However, accuracy-based criteria used in these previous studies are often claimed to be ambiguous (e.g., Larsen-Freeman and Long, 1991; Pienemann, 1998). In fact, it is difficult to determine whether or not the learner has acquired the grammatical structure based on the accuracy, since even 80 to 90% accurate use across three consecutive testing sessions does not guarantee that the same or higher accuracy use continues in learner’s production after those sessions. In order to address the issue on the accuracy-based criteria, the emergence criterion, which “identifies the point of first emergence of a structure in an interlanguage system” (Pienemann, 2005, p.54), was developed by Meisel, Clahsen, and Pienemann (1981). As the emergence criterion has been applied in much L2 research based on PT 3 (e.g., Pienemann, 1998; Di Biase and Kawaguchi, 2002), the current study also considers the first emergence of plural maker -s and NP plural agreement as the acquisition points of those structures. More detailed description of PT-based analysis is provided in the following sections.

4. Processability Theory Processability Theory (PT; Pienemann, 1998, 2005) is one of the major current approaches in second language acquisition (SLA) studies and a universal framework which is able to predict developmental sequences for any L2. According to Pienemann (1998, 2005), the logic underlying PT is that the learner at any stage is able to produce and comprehend only those L2 linguistic structures that current stage of language processor can handle. Therefore, it is important to understand the architecture of the language processor and how it handles an L2 in order to predict the developmental sequences of linguistic structures in L2 acquisition. PT follows the view on language production proposed by Levelt’s (1989) speech model, which overlaps to some extent with Kempen and 3

Besides PT-based studies, Yamamoto (2005) has shown that three Japanese L1 children, aged 3, acquired plural marker -s on nouns 24 to 36 months after the children began to be exposed to English L2, although the acquisition criterion in her study was not clearly explained.

14

The development of plural marking and plural agreement in child English L2 acquisition

Hoenkamp’s (1987) and Garrett’s (e.g., 1976, 1980, 1982) work. The basic assumptions of the language processing in PT are as follows: I. Processing components are relatively autonomous specialists which operate largely automatically; ii. Processing is incremental. iii.The output of the processor is linear, while it may not be mapped onto the underlying meaning in a linear way. iv. Grammatical processing has access to a grammatical memory. (Pienemann, 2005: 4 ff.)

Based on Kempen and Hoenkamp’s (1987) Incremental Procedural Grammar (IPG), a set of grammatical encoding procedures is formed according to their sequence of activation in the language production process. The processing procedures and routines in PT hierarchy are as follows: I. the lemma, ii. the category procedure (lexical category of the lemma), iii. the phrasal procedure (instigated by the category of the head), iv.the S-procedure and the target language word order rules, v. the subordinate clause procedure - if applicable. (Pienemann, 2005: 9)

PT claims that this sequence follows an implicational pattern where each procedure is a necessary prerequisite for the following procedure. L2 learners are claimed to build all of these procedures, apart from the first one (de Bot, 1992). “A word needs to be added to the L2 lexicon before its grammatical category can be assigned. The grammatical category of a lemma is needed before a category procedure can be called. Only if the grammatical category of the head of phrase is assigned can the phrasal procedure be called. Only if a phrasal procedure has been completed and its value is returned can Appointment Rules determine the function of the phrase. And only if the function of the phrase has been determined can it be attached to the S-node and sentential information be stored in the S-holder.” (Pienemann, 1998:80)

In order to explain the above hierarchy in relation to grammatical structures in individual languages, PT has incorporated Lexical-Functional Grammar (LFG; Kaplan and Bresnan, 1982; Bresnan, 2001), which is a typologically and psychologically plausible grammar theory. According to Pienemann (1998), LFG, which shares the key aspects with Kempen

Yumiko Yamaguchi

15

and Hoenkamp’s (1987) IPG, is efficient to analyze the psycholinguistic process of grammatical information exchange. LFG consists of a constituent structure (c-structure), a lexicon, a functional structure (fstructure) and an argument structure (a-structure). The most prominent characteristic of LFG is ‘feature unification’, namely information matching between parts of the sentence. Figure 1 shows the example of three parallel structures and feature unification in LFG.

Figure 1: Three parallel structures and feature unification in LFG (After Pienemann, Di Biase, and Kawaguchi, 2005: 200) In the sentence in Figure 1 (‘Peter sees a dog’), the lexical entries of the noun ‘Peter’ and the verb ‘sees’ have the values ‘3’ (third) and ‘SG’ (singular). In order to achieve subject verb-agreement, the grammatical information is passed on to the NP (Noun Phrase) procedure and VP (Verb Phrase) procedure and then two sets of information are passed on to the S (sentence) procedure where the diacritic features of PERSON and NUMBER is matched or unified. In PT, processability hierarchy is formed based on the points of unification, as follows: 1. No exchange of grammatical information (= no unification of features). 2. Exchange of grammatical information within the phrase 3. Exchange of grammatical information within the sentence. (Pienemann, 2007: 143)

16

The development of plural marking and plural agreement in child English L2 acquisition

Table 2 shows hypothesized processability hierarchy for English L2 (cf. Pienemann, 2005: 24, Di Biase and Kawaguchi, 2004: 80). In this section, processability hierarchy for English L2 is explained focusing on the morphological structures for plural marking, which the present study examines. Please see Pienemann (1998, 2005) for further explanations on processability hierarchy for English L2. stage 6 5

Processing procedure

Morphosyntactic structures

Examples

Subord.cl.procedure (inter-clausal) S-procedure (inter-phrasal)

Indirect Q

I wonder where he went.

Wh-Do Q? Wh-Aux Q? SV agreement (= 3rd person -s) Copula-Sub Q? Wh-Copula Q? Tense agreement (Aux - V) Do-SV(O) Q? Wh - SV(O) Q? ADV+SV(O) NP agreement SVO Q? SVO Plural -s Past -ed Progressive -ing Invariant forms formulae

What did you eat? Where can she play? Mary drinks coffee

4

VP procedure (inter-phrasal)

3

Phrasal procedure (phrasal information)

2

Category procedure (lexical)

1

Word/lemma

Is he in the car? Where is John? He has seen Tom. Do you like tea? Where you go? Today I study. I have two apples You like cats? I drink juice We read books Sam walked Daddy is sleeping Dog Hello How are you?

Table 2: Processing hierarchy for English L2 (After Pienemann, 2005: 24, Di Biase and Kawaguchi, 2004: 80) In PT, plural marker -s is called ‘lexical’. In order to assign lexical morphemes, such as plural -s, regular past -ed, and progressive -ing, on nouns or verbs, no exchange of grammatical information is required, since the diacritic feature is to be marked in one constituent only. In other words, as diacritic features such as “number” and “tense” are listed in the lexical entries of word, marking on nouns or verbs can be achieved directly from conceptualization. In PT (Pienemann, 1998, 2005), as lexical morphemes can be produced without phrasal procedures, they are

Yumiko Yamaguchi

17

claimed to develop before phrasal procedures. Hence, they are hypothesized to emerge at stage 2, as shown in Table 2. With regard to NP agreement, diacritic features are to be marked in several constituents and their value information has to be matched between constituents. The following example shows the relevant lexical entries for the phrase “many dogs” (Pienemann, 2005: 25). many:

DET

dogs:

N

(SPEC) (NUM) (PRED) (NUM)

= = = =

MANY PL DOG PL

In this example, as the value of DET (determiner) (i.e., “many”) for NUM (number) is PL (plural), the value of the head N (noun) (i.e., “dog”) for NUM needs to be PL with plural marker -s for NP agreement. In other words, “the value for the diacritic feature ‘number’ has to be unified between DET and N” (Pienemann, 2005: 25). For this process, the grammatical information is stored in a phrasal procedure. As “this type of morpheme becomes available to the language learner once phrasal procedures have been developed for the L2” (Pienemann, 2005, p.12), it is called ‘phrasal’ in PT. Accordingly, it is hypothesized to emerge at stage 3.

5. PT-based studies in English L2 Pienemann (1998, 2005) tested Processability hierarchy for English L2 against empirical data from Johnston’s (1985) study of Vietnamese and Polish adult learners and Pienemann and Mackey’s (1993) study of child learners. Johnston’s (1985) cross-sectional study examined the development of 12 English grammatical structures in 24 Polish and Vietnamese adult immigrants in Australia. Pienemann (1998) has shown that the 12 grammatical structures were acquired by Vietnamese and Polish adult learners in the order described in PT and has claimed that Johnston’s (1985) study strongly supports English L2 processability hierarchy. Another study, which Pienemann (1998) examined in order to test English L2 processability hierarchy, was a cross-sectional study of child English L2 acquisition by Pienemann and Mackey (1993). In their study, the speech samples from 13 children aged 8 to 10 years old were collected using various communication tasks, such as habitual actions, story completion, informal interview, picture sequencing, picture differences,

18

The development of plural marking and plural agreement in child English L2 acquisition

and meet partner. Although the implicational analysis of the children’s speech corpus was presented in the same way as that for Johnston’s (1985) data, two additional structures were included in Pienemann and Mackey’s (1993) study. As in Johnston’s (1985) study, no contradictory evidence to the hierarchy predicted in PT has been found in the child English L2 acquisition. Hence, the empirical data from Pienemann & Mackey’s (1993) study has shown additional support for English L2 processability hierarchy. It means both English L2 studies (Johnston, 1985; Pienemann and Mackey, 1993) have shown that plural marker -s is acquired before NP plural agreement as predicted in PT. However, empirical evidence for English L2 processability hierarchy is still limited, since studies by Johnston’s (1985) and Pienemann and Mackey’s (1993) were both cross-sectional and PT has never been tested in longitudinal studies in English L2 context. Moreover, more detailed analysis on the English L2 developmental sequence for plural marker -s and NP plural agreement is required, since the sequence predicted in PT has been challenged by some recent research (Charters and Jansen, 2007; Dao, 2007). In Dao’s (2007) study, spontaneous oral production of 36 Vietnamese instructed learners of English L2, aged between 13 and 18, was elicited by tasks and their morphological development was examined. The results of her study showed that plural marker -s had occurred with numerals (e.g., five books), before plural marker -s without any quantifiers (e.g., books) emerged. More specifically, 6 Vietnamese learners, who did not acquire lexical plural morpheme -s, were found to have already acquired NP plural agreement with numerals in Dao’s (2007) study. This suggests that the developmental sequence for plural marker -s and NP plural agreement in Vietnamese learners’ English L2 acquisition is not consistent with the sequence predicted in PT. Based on the findings in Dao (2007), Charters and Jansen (2007) have claimed that as numerals facilitate the acquisition of plural marking by highlighting the conceptual transparency, ‘phrasal plural’ with numerals (e.g., “I have two apples”) emerges before ‘lexical plural’ (e.g., “I like apples”). However, as Dao (2007) examined instructed L2 learners crosssectionally, more detailed analysis, in particular, in longitudinal studies on naturalistic L2 learners needs to be conducted to investigate the discrepancy. In addition, PT currently does not distinguish between two types of NP plural agreement (i.e., NP plural agreement with numerals and other quantifiers). Hence, the development sequence for each type of NP plural agreement should be analysed separately, as Charters and Jansen (2007) suggested.

Yumiko Yamaguchi

19

6. Empirical Study The present study is based on a work in progress for my PhD thesis. The study takes the form of a two-year longitudinal case study of a Japanese L1 primary school aged (5;8 to 7;8) child acquiring English as L2 in Australia. In order to investigate the developmental sequence for English L2 plural marker -s and NP plural agreement predicted in PT and to examine the discrepancy between PT and recent English L2 research (e.g., Charters and Jansen, 2007), the present study addresses the question whether plural marker -s on nouns (e.g., books) is acquired before NP plural agreement, in particular, NP plural agreement with numerals (e.g., two books), in English L2 acquisition of a Japanese child.

6.1 Participant The participant, Kumi 4 , is the second daughter of Japanese native speaker parents. Their conversation at home is always in their native language, Japanese. The family moved to Australia when she was five years old and she was enrolled in a local primary school where all subjects were taught in her second language, English. She had learned some English before she moved to Australia. However, her English skills were very limited at the time when she started attending the local primary school. She was able to produce only basic sentences such as “I don’t know” and “thank you”. Such sentences are recognized as formulae or chunks at stage 1 in processability hierarchy, as they require no procedural skills.

6.2 Procedures Data collection started one month after Kumi started attending local primary school. In other words, she had been exposed to English for one month in Australia by the time of the first recording. Kumi’s speech in English was tape-recorded from the time she was 5 years 8 months old until 7 years 8 months old, namely over two years. Data was collected fortnightly for the first two months, after that, every two months for the rest of the first year, and every three months in the second year. For speech elicitation, various tasks, including semistructured interviews, narratives, and communication games were utilized 4

Kumi is a fictitious name.

20

The development of plural marking and plural agreement in child English L2 acquisition

and performed with native or near native speakers of English. Each recording session lasted around 20 to 30 minutes. Kumi’s speech production was transcribed.

6.3 Data Analysis Based on the transcribed speech corpus for two-year longitudinal study on Japanese L1 child learner acquiring English L2, a distributional analysis of plural marker -s on nouns and two types of NP plural agreement (i.e., NP plural agreement with numerals and non-numeric quantifiers) was conducted. The obligatory plural contexts were determined based on Jia (2003), which examined English L2 plural morpheme acquisition by Chinese children. Jia (2003) adopted linguistic cues and contextual cues as the criteria for the obligatory contexts for plural morpheme, following Cazden (1968): The obligatory context indexed by linguistic cues refers to count nouns that occur after (a) determinatives, such as some, all, both; (b) cardinal numbers, such as two and three; (c) noun phrases, such as a dozen, or fractions, such as two thirds; and (d) plural demonstratives, such as these and those. The contextual cues included those derived from sentential contexts (e.g., break into pieces or use chopsticks), discourse contexts (e.g., her stepsisters were jealous of her beauty, after mentioning that Cinderella’s stepmother had two daughters), and pictorial contexts (the number of a particular object portrayed on a picture). (Jia, 2003, p.1301)

In this study, as lexical plural (i.e., plural marker -s) and phrasal plural (i.e., NP plural agreement) are examined separately, based on PT. Hence, the present study determines the obligatory plural contexts for plural marker -s based on the contextual cues and those for NP agreement based on the linguistic cues. Following Jia (2003), the current study excludes contexts for routine uses of plural marking. These include “the names of cartoon characters (e.g., Power Rangers and Goose Bumps), names of places (e.g., Six Flags), nations (e.g., United States), and school subjects, (e.g., Social Studies)” (Jia, 2003, p.1301). Jia (2003) claims that the learner may produce plural forms of these nouns through rote memory and that the singular forms of these nouns are not used in English, in contrast to other nouns with plural forms. Moreover, if there were any doubts about whether the plural marker was obligatory or not, it was excluded from the analysis. Echoic

Yumiko Yamaguchi

21

expressions, which occurred when the learner repeated the interlocutor’s utterances, were also excluded. Based on previous PT studies, emergence criterion was applied to determine the acquisition points of target structures. A structure was considered to have been acquired if the rule is supplied more than once in lexically and structurally varied contexts, following Di Biase & Kawaguchi (2002). Although most previous studies on the acquisition of grammatical morphemes (e.g., Brown, 1973; Cazden, 1968; Hakuta, 1976, 1978; Jia, 2003) determined the acquisition point based on the accuracy, it has been pointed out that the accuracy of morpheme insertion cannot be guaranteed to develop steadily (e.g., Larsen-Freeman and Long, 1991; Pienemann, 1998).

6.4 Results The results of the distributional analysis on Kumi’s two-year longitudinal study on English L2 acquisition of morphological plural marking are presented in Table 3. The first row shows the different points in time (T1, T2…) in the corpus, while the far left column indicates the hypothesized morphological structures for plural marking, that is, plural marker -s on nouns, NP plural agreement with numerals, and NP plural agreement with non-numeric quantifiers. In this table, the number after “+” shows the total number of the occurrences of the hypothesized structures in obligatory plural contexts. The number after “-” indicates the total number of the contexts in which Kumi failed to produce the hypothesized structures in obligatory plural contexts. In addition, “>” shows the instances when Kumi oversupplied plural marker -s on nouns in singular contexts or on irregular plural nouns. There is no “oversupply” for NP plural agreement, as all the NPs analyzed had numerals or non-numeric quantifiers. Table 3 indicates that Kumi acquired plural marker -s on nouns in Time 3, namely 2 months after she was exposed to English L2 in Australia. NP plural agreement with numerals is shown to have emerged in Time 7, that is, after 7 month exposure of English L2. Then, NP plural agreement with other quantifiers appeared in Time 9, namely 11 months after her first exposure to English L2.

㻌

+0 -6

+0 -6

0

㻌

0

0

0

T3

T2

T1

㻌

0

0

T4

㻌

+1 -4

0

+0 -1

㻌

0

㻌

T6

T5

㻌

+2 -1

+0 -3

T7

+1 -2

0

T8

+7 -8

+1 -3

+3 -2

T9

+10 -2

+6 -6

+6 -3

T10

Table 3. Distributional analysis of Kumi’s plural marking acquisition

+2 plural marker +0 +2 +2 +2 +22 +2 +2 -7 -s on nouns -5 -6 -4 -4 -2 -4 -2 >1 Note: “+ = supplied in plural context, “-” = not supplied in plural context, “>” = oversupply (marked in singular context or overregularizations)

structures NP plural agreement with other quantifiers 㻌 NP plural agreement with numerals 㻌 +6 -2

+3 -2

+2 -0

T11

+9 -1

+2 -0

+6 -1

T12

+14 -12 >1

+0 -1

+6 -4

T13

+39 -3 >1

+5 -1

+9 -0

T14

22 The development of plural marking and plural agreement in child English L2 acquisition

Yumiko Yamaguchi

23

All the instances involving three hypothesized structures, such as plural marker -s on nouns, NP plural agreement with numerals, and NP plural agreement with non-numeric quantifiers, in Kumi’s speech production are presented in Appendix A, B, and C respectively. Most of the obligatory contexts for plural marking observed in Kumi’s early English L2 production occurred when she talked about pictures (i.e., ‘pictorial contexts’). In Time 1, namely, one month after Kumi started attending local primary school in Australia, Kumi never supplied plural marker -s on nouns, although there were 5 obligatory contexts. There was no context for NP plural agreement with either numerals or non-numeric quantifiers in Time 1. In Time 2, namely after 1.5 month English L2 exposure, Kumi produced “sunglasses” twice. However, as they occurred in the same lexical contexts, Time 2 was not considered as the emergence point for plural marker -s. As for NP plural agreement, there were 6 obligatory contexts with numerals (e.g., “two light”). However, Kumi never produced NP plural agreement with numerals. In Time 3, Kumi supplied plural marker -s on nouns twice (i.e., “books”, “shoes”). As they occurred more than once in lexically and structurally varied contexts, Time 3 was considered as the acquisition point for plural marker -s. Time 3 was 2 months after Kumi had started attending local primary school in Australia. At the same time, obligatory contexts for NP plural agreement with numerals occurred 6 times (e.g., “two mountain”, “three tree”). However, she failed to produce NP plural agreement successfully. As shown in Table 3, it is clear that Kumi started supplying plural marker -s on nouns constantly in obligatory contexts from Time 3. However, NP plural agreement was not acquired until Time 7, although Kumi produced NP plural agreement with a numeral once in Time 5 (i.e., eight babies). As this was only instance for successful NP plural agreement with numerals, Time 5 was not considered as the acquisition point. There was no context for NP plural agreement in Time 6. In Time 7, namely 7 months after Kumi was exposed to English L2 in Australia, NP plural agreement occurred twice in lexically and structurally varied contexts (i.e., “one hundred bees” and “two mans”). Although the plural marker -s in “two mans” is oversupplied on an irregular plural noun 5 , it can be considered as a positive occurrence for NP plural agreement, as this indicates that Kumi attempts to express plurality 5

Irregular plural nouns, such as “men” and “feet” are called “vowel-change plurals” (Jia, 2003). See Jia (2003) for further information on irregular plural nouns.

24

The development of plural marking and plural agreement in child English L2 acquisition

with plural marker -s. Hence, Time 7 was considered as the acquisition point for NP plural agreement with numerals. However, Kumi failed to produce NP plural agreement with non-numeric quantifiers in two obligatory contexts (i.e., “lots of baby”, “many difference”) in Time 7. In Time 9, that is, after 11 months exposure to English L2 in Australia, Kumi produced NP plural agreement with non-numeric quantifiers in three contexts (i.e., “most kids”, “a lot of girls”, “a lot of letters”). It means that Kumi acquired NP plural agreement with nonnumeric quantifiers in Time 9. In sum, Kumi acquired plural marker -s in Time 3, namely 2 months after she was exposed to English L2 in Australia. Then, she acquired NP plural agreement with numerals in Time 7, that is, after 7 month exposure to English L2. Finally, in Time 9, namely 11 months after Kumi was exposed to English L2, NP plural agreement with non-numeric quantifiers was acquired.

7. Discussion The results of Kumi’s longitudinal study show that she first acquired plural marker -s on nouns, which requires lexical procedure, and then NP plural agreement, which requires phrasal procedure, in her English L2 acquisition. In other words, Kumi’s acquisition of English L2 plural morphemes was implicational: lexical procedure > phrasal procedure, as shown in Table 4. It suggests that the developmental sequence of plural marker -s on nouns and NP plural agreement in English L2 acquisition of the Japanese L1 child is compatible with the sequence predicted in PT. structures/time NP plural agreement

T1

T2

T3

T4

T5

T6

T7

T8

T9

T10

T11

T12

T13

T14

/

-

-

/

-

/

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

(phrasal) plural marker -s on nouns (lexical)

Note: “+ =acquired, “-” = not acquired, “/” = no context Amount of L2 exposure time: T1=1 month, T2=1.5 months, T3=2 months, T4=2.5 months, T5=3 months, T6=5 months, T7=7 months, T8=9 months, T9=11 months, T10=13 months, T11=16 months, T12=19 months, T13=22 months, T14=25 months

Table 4: Implicational analysis of Kumi’s plural marking acquisition

Yumiko Yamaguchi

25

Based on these findings, it can be argued that English L2 processability hierarchy for morphological plural marking is supported by longitudinal data as well as cross-sectional data (Johnston, 1985; Pienemann and Mackey, 1993). Moreover, the present study, which analyses the development of NP agreement with numerals and non-numeric quantifiers separately, has added valuable evidence for Processability hierarchy for English L2 morphology, since PT currently does not distinguish NP plural marking by quantifiers (i.e., numerals or non-numeric quantifiers). As shown in Table 3 in the result section, plural marker -s on nouns is acquired before both NP agreement with numerals and non-numeric quantifiers in Kumi’s English L2 acquisition. In fact, Kumi kept failing to produce NP plural agreement with numerals until Time 7 (i.e., after 7 months exposure to English L2) after she had acquired plural marker -s in Time 3 (i.e., after 2 months exposure to English L2). In other words, NP plural agreement with numerals was acquired 5 months later than plural marker -s without numerals. Accordingly, it is clear that plural marker -s is acquired before NP agreement with numerals and the findings of the present study falsified Charters and Jansen (2007), who have argued that plural marking with numerals is acquired earlier than plural marking without numerals. One of the possible reasons for this discrepancy is a methodological difference. While the present study is longitudinal, Dao’s (2007) study presented in Charters and Jansen (2007) is cross-sectional. However, the developmental sequence for English L2 plural marking found in the present study is compatible with the sequence predicted in PT, which is based on evidence from two cross-sectional studies (Johnston, 1985; Pienemann and Mackey, 1993). As PT has been mostly tested by crosssectional studies so far, more evidence from longitudinal studies is required to determine whether or not both longitudinal and cross-sectional studies provide identical developmental sequences in L2 acquisition. In addition to the difference between longitudinal and cross-sectional studies, learner’s speech production may be affected by speech elicitation techniques. Although Pienemann’s (1998) ‘steadiness hypothesis’ claims that “basic nature of the grammatical system of an IL (interlanguage) does not change in different communicative tasks as long as these are based on the same skill type in language production” (Pienemann, 1998, p. 273), it is possible that materials used for tasks, such as pictures, facilitate learners’ production of certain grammatical structures, in particular, plural morpheme. Early morpheme studies (e.g., Dulay and Burt, 1974), which used Bilingual Syntax Measure (BSM), have also been criticized for using

26

The development of plural marking and plural agreement in child English L2 acquisition

a developmental sequence which was a mere artifact of BSM (e.g., LarsenFreeman, 1975). Another possible reason is that the participants in Dao’s (2007) study are instructed ESL learners, while those in the present study as well as in Johnston’s (1985) and Pienemann and Mackey’s (1993) studies are naturalistic learners of English L2. In other words, ESL instruction, which the Vietnamese learners in Dao’s (2007) study received, may have influenced the development of NP plural agreement with numerals. However, according to the Teachability Hypothesis (Pienemann, 1984, 1989), stages of acquisition cannot be skipped through formal instruction and instruction will be beneficial if it focuses on structures from ‘the next stage’ (Pienemann, 1998). One of the recent studies which show support for this prediction is Mansouri and Duffy’s (2005) experimental study on the effect of ESL instruction on the developmental sequence of English syntactic structures. Mansouri and Duffy (2005) found that L2 learners, who were exposed to ESL instruction based on the developmental sequence predicted in PT, were able to produce the English syntactic structures more accurately than those who were exposed to the reversed order. This finding has suggested that instruction can not change the developmental sequence of English L2 syntax predicted in PT. However, instructed English L2 learners in Dao’s (2007) study are found to acquire NP plural agreement, which is claimed to require phrasal procedure in PT, before plural marker -s on nouns, which does not require phrasal procedure. As Mansouri and Duffy’s (2005) study focused on syntactic structures, further experimental studies, focusing on morphological structures, are needed in order to examine whether or not instruction can change the developmental sequence of English L2 morphology. It should be noted that some previous English L2 studies on Asian learners acquiring English grammatical morphemes (e.g., Hakuta, 1976, 1978; Jia, 2003; Jia et al, 2002) have argued that the plural morpheme appears relatively later than other grammatical morphemes. For instance, Hakuta (1976, 1978), who examined English L2 acquisition of 5 year-old Japanese girl, showed that the plural morpheme was not acquired by the end of his 7 month longitudinal study. Jia (2003) found that only 7 of 10 Chinese L1 children mastered English plural morpheme after 5 years of English exposure. However, the present study, which also examined English L2 acquisition of 5 year-old Japanese girl, shows that the acquisition point of plural marker -s was 2 months after the participant was exposed to English L2. The reason for this discrepancy should be the methodological difference between PT-based studies and other English grammatical morpheme studies. The present study, which is based on PT,

Yumiko Yamaguchi

27

applies emergence criterion to determine the acquisition points of grammatical structures, while other morpheme acquisition studies depended on accuracy. As in Table 5, indicating the percentages of the accurate use of plural marker -s on nouns by Kumi, the same or higher accuracy use never continues in her production from Time 1 to 14, namely over two years. Hence, the emergence criterion seems more suitable to determine the acquisition points of the grammatical structures, since the accurate-based criteria were ambiguous (e.g., Larsen-Freeman and Long, 1991; Pienemann, 1998). time %

T1 0

T2 20

T3 25

T4 33.3

T5 42.8

T6 91.6

T7 33.3

T8 50

T9 46.6

T10 83.3

T11 75

T12 90

T13 48.2

T14 90.6

Table 5: The percentages of the accurate use of plural marker -s on nouns

8. Conclusion The present study has demonstrated that the developmental sequence of plural marker -s on nouns and NP plural agreement in English L2 acquisition by a Japanese L1 child learner is consistent with the sequence predicted in PT. Moreover, the current study contributed to the advancement of the processability hierarchy for English L2 morphology since it found that NP plural agreement occurred with numerals first, then with non-numeric quantifiers later. Furthermore, the present study on a Japanese L1 child acquiring English L2 has provided new evidence for English L2 morpheme acquisition studies. However, the present study was a case study of a Japanese L1 child acquiring English L2 in Australia. As other recent English L2 research (e.g., Charters and Jansen, 2007) obtained the controversial results that plural marking with numerals is acquired before plural marking without numerals, more studies on developmental sequence of English L2 plural marking based on PT need to be conducted in various contexts to investigate whether or not the sequence is varied by the differences of research methodologies, learning environments, L1 backgrounds and so on. A further issue needing to be addressed in future studies is whether producing NPs with numerals really requires phrasal procedure, that is, grammatical information exchange between two constituents, as predicted in PT (e.g., Charters and Jansen, 2007). It can be argued that NPs involving numerals require conceptual information processing, while NPs involving non-numeric quantifiers require grammatical information

28

The development of plural marking and plural agreement in child English L2 acquisition

exchange between the head noun and the quantifiers (e.g., Itani-Adams, in progress). The treatment of plural marking with numerals in English L2 processability hierarchy may require further examination in future PTbased research.

9. Bibliography Bresnan, J. (2001). Lexical-Functional Syntax. Oxford: Blackwell. Brown, R. (1973). A first language: The early stages. Cambridge, MA: Harvard University Press. Bialystok, E., and Miller, B. (1999). The problem of age in secondlanguage acquisition: Influences from language, structure, and task. Bilingualism: Language and Cognision, 2, 127-145. Cazden, C. B. (1968). The acquisition of noun and verb inflections. Child Development, 39, 433-448. Charter, H., and Jansen, L. (2007). Do you agree?: The SLA of number concepts, features, morphs and agreement. Paper presented in EUROSLA 2007, Newcastle, U.K. Corbett, G. (2000). Number. Cambridge, U.K.: Cambridge University Press. Dao, L. (2007). Foreign Language Acquisition: Processes and Constrains. PhD Thesis, Australian National University, Australia. de Bot, K. (1992). A bilingual production model: Levelt's 'Speaking' model adapted. Applied Linguistics, 13(1), 1-24. de Villiers, J. G., and de Villiers, P. A. (1973). A cross-sectional study of the acquisition of grammatical morphemes in child speech. Journal of Psycholinguistic Research, 2, 267-278. Di Biase, B., and Kawaguchi, S. (2002). Exploring the typological plausibility of Processability Theory: language development in Italian second language and Japanese second language. Second Language Research, 18(3), 274-302. Di Biase, B., and Kawaguchi, S. (2004). Second Language Acquisition: Notes, Readings, & Exercises. Sydney: University of Western Sydney. Downing, P. A. (1996). Numeral Classifier systems: the Case of Japanese (Studies in Discoure and Grammar 4). Amsterdam: John Benjamins. Dulay, H. C., & Burt, M.K. (1974). Natural Sequences in Child Second Language Acquisition Language Learning, 24(1), 37-53. Ferenz, K. S., and Prasada, S. (2002). Singular or plural? Children's knowledge of the factors that determine the appropriate form of count nouns. Journal of Child Language, 29, 49-70. Flege, J. E., Yeni-Komshian, G. H., and Liu, S. (1999). Age constraints on

Yumiko Yamaguchi

29

second-language acquisition. Journal of Memory and Language, 41, 78-104. Garrett, M. F. (1976). Syntactic processes in sentence production. In R. J. Wales, & Walker, E. (Ed.), New approaches to language mechanism. (pp. 231-256). Amsterdam: North-Holland. Garrett, M. F. (1980). The limits of accommodation: Arguments for independent processing levels in sentence production. In V. A. Fromkin (Ed.), Errors in linguitic perfomance: Slips of the tongue, ear, pen, and hand. (pp. 263-272). New York: Academic Press. Garrett, M. F. (1982). Production of speech: Observations from normal and pathological language use. In A. W. Ellis (Ed.), Normality and pathology in cognitive functions. (pp. 19-76). London: Academic Press. Hakuta, K. (1976). A case study of a Japanese child learning English as a second language. Language Learning, 26(2), 321-351. Hakuta, K. (1978). A report on the development of grammatical morphemes in a Japanese girl learning English as a second language. In E. M. Hatch (Ed.), Second Language Acquistion: A Book of Readings. Rowley, M.A.: Newbury House Itani-Adams, Y. (in progress). One child two languages: Bilingual first language acquisition in Japanese and English. PhD Thesis, University of Western Sydney, Sydney, Australia. Jia, G. (2003). The Acquisition of the English Plural Morpheme by Native Mandarin Chinese-Speaking Children. Journal of Speech, Language, and Hearing Research, 46, 1297-1311. Jia, G., Aaronson, D., and Wu, Y. H. (2002). Long-term language attainment of bilingual immigrants: Predictive factors and language group differences. Applied Psycholinguistics, 23, 599-621. Johnston, M. (1985). Syntactic and Morphological Progressions in Learner English. Canberra: Commonwealth Dept of Immigration and Ethnic Affairs. Kaplan, R., & Bresnan, J. (1982). Lexical-functional grammar: a formal system for grammatical representation. In J. Bresnan (Ed.), The Mental Representation of Grammatical Relations (pp. 173-281). Cambridge, MA: MIT Press. Keßler, J.-U. (2006): Englischerwerb im Anfangsunterricht diagnostizieren. Linguistische Profilanalysen und der Übergang von der Primar- in die Sekundarstufe I. Giessener Beiträge zur Fremdsprachendidaktik. Tübingen: Narr. Kempen, G., & Hoenkamp, E. (1987). An incremental Procedural Grammar for sentence formulation. Cognitive Science, 11, 201-259. Levelt, W. J. M. (1989). Speaking. Cambridge, Massachusetts: The MIT

30

The development of plural marking and plural agreement in child English L2 acquisition

Press. Lahey, M., Liebergott, J., Chesnick, M., Menyuk, P., and Adams, J. (1992). Variability in the use of grammatical morphemes: Implications for understanding language impairment. Applied Psycholinguistics 13, 373-398. Larsen-Freeman, D., & Long, M.H. (1991). An Introduction to Second Language Acquisition Reseach. London: Longman. Makino, S., and Tsutsui, M. (1986). A dictionary of basic Japapanese grammar. Tokyo, Japan: The Japan Times. Mansouri, F. (2005): Agreement morphology in Arabic as a second language: Typological features and their processing implications. In M. Pienemann, (Ed.), Cross-Linguistic Aspects of Processability Theory (pp. 117- 153). Amsterdam: John Benjamins. Mansouri, F. & L. Duffy (2005). The pedagogic effectiveness of developmental readiness in ESL grammar instruction. Australian Review of Applied Linguistics, 28(1), 81-99. Maynard, S. K. (1990). An introduction to Japanese grammar and communication strategies. Tokyo, Japan: The Japan Times. Meisel, J. M., Clahsen, H., & Pienemann, M. (1981). Determining developmental stages in natural second language acquisition. Studies in Second Language Acquisiton, 3, 109-135. Pienemann, M. (1984). Psychological constraints on the teachability of languages. Studies in Second Language Acquisiton, 6(2), 186-214. Pienemann, M. (1989). Is language teachable? Psycholoinguistic experiments and hypotheses. Applied Linguistics, 10(1), 52-79. Pienemann, M. (1998). Language Processing and Second Language Development: Processability theory. Amsterdam: John Benjamins. Pienemann, M. (2005). Cross-Linguistic Aspects of Processability Theory. Amsterdam: John Benjamins. Pienemann, M. (2007). Processability Theory. In B. VanPatten, and Williams, J. (Eds.), Theories in second language acquisition: An Introduction. Mahwah, New Jersey: Lawrence Erlbaum Associates. Pienemann, M. & Håkansson, G. (1999). A unified approach towards the development of Swedish as L2: Processability account. Studies in Second Language Acquisition, 21, 383-420. Pienemann, M., & Mackey, A. (1993). An empirical study of children's ESL development and rapid profile. In P. McKay (Eds.), ESL Development: Language and literacy in Schools (Vol.2, pp. 115-259). Canberra: Commonwealth of Australia and National Languages and Literacy Institute of Australia. Yamamoto, A. (2005). Kodomo no Eigo Gakushu: Shutoku katatei no

Yumiko Yamaguchi

31

puroto taipu (Children's Engish Learning: the prototype of acquisition process). Tokyo: Kazama shobo. Zhang, Y. (2005): Processing and formal instruction in the L2 acquisition of five Chinese grammatical morphemes. In M. Pienemann, (Ed.), Cross-Linguistic Aspects of Processability Theory (pp. 155-177).

“Do you have a man wear the sunglasses?”

“we go the there. we we have a books to read”(talking about a bookshop) “girl find under the shoes”

“I play with my teddy and blocks” “daddy climb the rocks”

“do you have tables?” “boy went climb in the rocks” “nuts”** (in response to “what kind of cooking did you do?”)

T3

T4

T5

supply

T2

T1

*“we did with crayon for make the (X)” *“children cut out and stick..and draw the leaf for flower” *“bee come out” *“bee is following the a dog”

non-supply *“bee comes out” x 2 *“dog put er hand re r in the er tree” *“then bee chase to the dog” *“frog had a baby” *“brown stick and light is white” *“bird egg is er falling down” *“today..in kindygarten..I got.. easter chocolate.. and easter chocolate from my friend” *“your light is off?” *“my light is off.. again” *“brown stick.. and light is white” *“today we need a hat” *“I have a butterfly” *“I ask you you the bird in sky” *“children have a children book” *“bee come out” *“bee chase the dog” *“turtle has a baby” *“this has a chair inside” *“do you have a picture of boat?” *“picture of boat..yacht” *“cat crash the egg”

*“girl look.. the bottles” (referring to one bottle)

over-supply

The development of plural marking and plural agreement in child English L2 acquisition

Appendix A: Samples of plural marker -s on nouns

32

T8

T7

T6

supply “beautiful shop is a kind of a beautiful things’ shop” “this is the for the girls and boys.. big girls and boys” “I went to Gold Coast in the holidays” “do you have a green bumps?” x 2 “the bumps” “bumps..leaves’ bumps” “do you have a nothing on the leaves?” “bumps mean like this or (xx)” “that’s leaves’ bumps” “do you got a nothing on the bumps?” ”do you got a man wearing brown pants?” “do you got a man wearing the black shoes?” x 2 “she look under the shoes” “and she took she’s turtle and say bye to other turtles” “we play guessing game and we do the rocks” “this so yummy and and this is in a natural things” “her slippers fell down” “he look and there was a parents” *“then bee go” *“let’s follow the bee” *“bee was chasing dog already” *“baby came out..there were lot of baby” *“bee fell down” *“bee going to chase the dog”

non-supply *“we give the kangaroo a kangaroo food” *“bee was chasing the dog”

Yumiko Yamaguchi over-supply

33

34

T10

T9

supply “when we went to wild land. we saw ducks” “there was very big mountains” “my sister said I don’t want to see mountains anymore” “his parents again” “this has important things on it” “no.. but he’s wearing a brown pants under the coat” “I have a brown pants” “he open the window with his muscles up” “the bees were very angry about a dog” “when they begin to play. the dads came. they laugh at them” “they have colours on” “but in Japan they there are more much colours brown” “we did treasure hunt and Lilly’s mum Kathy tell the instructions and her sister’s friend and her cousin and her sister’s friend did did reading the clues” “we read the clues” “that was magnify glass and sunglasses” “the dollars are not real” *“they was just friend” *“my favourite prize was the gun and the phone”

non-supply *“cow horn was too big” *“sometime they are lady” *“they have hat on” *“they have tie on” x 2 *“they wear jacket and they have something looks like stick” *“I have broken bottle”

over-supply

The development of plural marking and plural agreement in child English L2 acquisition

T12

T11

supply “she look in the boots” “they are best friends” “today in assembly in the school there.. only kindergartens and year 1 and year 2 goes “you sit on a small seat and put your hands on the handle” “you cut with this.. not papers” “if you cut papers with this. it will be dirty” “in Saturday school. today was the day that mother and fathers come to look” “we did how we write numbers and Japanese letter thing” “the bees came out” “the lady fell down.. even her shoes was off” “this is the thing that you put lovely flowers in” “you want to put flowers in it” “this is the place that you you can have a look at lovely stories” “you can put your computers on” “you need to use the chairs for it” *“in Saturday school. today was the day that mother and fathers come to look”

non-supply *“this has a number from one to twelve” *“these cover so much of your leg”

Yumiko Yamaguchi over-supply

35

36

T14

T13

supply “my sister she cut the cardboard of hollies” “we did a er dress up in pajamas” “pens and cardboards and something like that” “there was beads as well” “we opened the presents” “this is something you put things in” “this is something that you put things in” “lots of people put mobile phones “is he putting earphones?” “is your grandma putting her glasses down?” “is that a normal shoes?” “I think I don’t have questions” “is your mum wearing a shoe.. shoes?” “the bees was so angry” “the dog was chased by the bees” “he didn’t know there was a deer horns” “he took his frog and said thank you good bye to the frog parents” “one of the boys took a sword from his box and cut the legs of the table” “it was too short.. it was up to the legs” “first I didn’t know it was apartment because it was like mat and stairs and *“you take picture and the pictures is real things” *“because he’s putting them (= his hands) in his pocket or something” *“when..like.. people isn’t putting seatbelt on even they’re driving”

non-supply *“he took it and said good bye to the other snail” *“lots of people put mobile phones and handkerchief and a little bit money case” *“we ate lamb chop” *“we made chocolate cream and put it on and bush de noel had cream and strawberry inside” *“your g.grandma picture wearing a slipper with a curly thing?” *“white curly thing on it?” *“a slipper with white curly thing on it” *“it’s not high heel” *“just like slipper” *“is your mum crossing her leg?”

*“then two young gentlemens came laughing..”

over-supply *“there were lots of child. childrens”

The development of plural marking and plural agreement in child English L2 acquisition

supply there was only floor place for rooms to live” “for snack we ate watermelons” “there is something that you can open and write stories in any stories” “monkeys love to eat them” “it’s normally in cities” “it has forty steps and things like that” “people can go into lots of steps and but stairs and elevator” “it’s tall and has elevators and windows” “it’s made made by mans” “you can take pictures of people” “the pictures is real things” “it’s really new and real trees and birds and normally you should bring it to holidays” “what flowers?” “I don’t have flowers” “in school days she dress dressed up” “on weekends she normally works or plays” “normally at afternoon on school days she like.. works in school” “in afternoon at home she eats snacks” “she.. like.. organizes works and stuff”

non-supply

Yumiko Yamaguchi over-supply

37

T11 T12 T13

T4 T5 T6 T7 T8 T9 T10

T1 T2 T3

two pumpkin nine hundred bee, seven baby two plane, ten time x 2 ten time, three time x 2, four time, two ring, three one dollar three time, two man

one hundred bees, two mans six babies ten times three times (2), ten times, two boys, nine years, five dollars two crabs, six kids, two wheels two tortoises, two kids, two snail

two man x 4

two light x 3, two bin x 2, two crab kid two tree x 2, two mountain, two grass, six grass, three tree

negative instances

eight babies

positive instances

Appendix B: Samples of NP plural agreement with numerals

over-supply

The development of plural marking and plural agreement in child English L2 acquisition

supply non-supply “after the schools she like.. marking works and like.. cleaning tables and stuff and putting art works on the walls” “he like.. goes outside and. like.. organizes roads and streets” “he need to take a ticket when people like.. parking at the bus stops and things.. they’re not supposed to park” “**”may be the default form, as no singular forms are observed at this time or before.

38

two frogs, forty steps, two stairs, two young gentlemens

six button

T14

T13

T11 T12

T1 T2 T3 T4 T5 T6 T7 T8 T9 T10

most kids, a lot of girls, a lot of letters lots of ants, some lands, any cars, lots of papers, any crowns, big bunch of the bees any words, any covers a lot of bunch of bees, many kind of things, lots of flowers x2, a lot of things, all the teachers lot of childrens, lots of stories, lots of magic things, lots of things, a lot of ladies, some tissues a bunch of bees, a lot of babies, one of the boys, lots of white pages, a lot of windows, lots of times, lots of steps, lots of buttons, lots of things

positive instances

all the thing, a little bit money case, any light, many difference

some of the car

lot of time, only little bit boy some child, more prize, one of a ring

lots of baby, many difference

lot of bee

negative instances

Appendix C: Samples of NP plural agreement with non-numeric quantifiers

T14

Yumiko Yamaguchi

39

DEVELOPMENT OF DISCOURSE FUNCTIONS IN JAPANESE AND ENGLISH BILINGUAL FIRST LANGUAGE ACQUISITION

YUKI ITANI-ADAMS In the field of bilingual first language acquisition (BFLA), one of the central issues is about the relationship between the two developing languages of a child who is regularly exposed to both languages from birth. In other words, researchers are interested in whether the two languages of such a child influence on each other during the paths of their development or whether they develop independently from each other. Earlier studies of acquisition of simultaneous acquisition of two languages (e.g., Taeschner, 1983; Volterra & Taeschner, 1978) hypothesised that bilingual children began with one linguistic system for both lexicon and syntax which was made up of elements from both languages. This one linguistic system would later separate into two different language systems. This hypothesis was named the Unitary Language System Hypothesis (ULS) (Genesee, 1989). However later empirical studies of BFLA involving different language constellations (e.g., De Houwer, 1990; Meisel, 1990; 1994; Mishina, 1997; Paradis & Genesee, 1996) showed different results. The results from these studies showed that bilingual children develop their two languages separately from early on supporting a hypothesis known as Separate Development Hypothesis (SDH) (De Houwer, 1990). The researchers of the studies supporting SDH compared the bilingual children’s development path of each of the two languages with that of monolingual children of the language in question. The focus of the majority of these studies was the development of morphosyntactic

I wish to thank Bruno di Biase and Satomi Kawaguchi for their advice at different stages during the writing of this paper. I would also like to thank the annonymous reviewer who provided valuable suggestions to improve the paper. I would like to acknowledge the School of Humanities and Langauges at the University of Western Sydney for the publication stipend that enabled the writing of this paper. The errors that remain are entirely mine.

42

Yuki Itani-Adams

structure (see Genesee, 2000). They reported that the developmental path of each of the two languages of bilingual children was similar to that of monolingual children. They did not find an evidence for transfer between the two languages of bilingual children. Some of the studies also found that timing of acquisition of certain morphological structures by bilingual children was similar to that of monolingual children, thus concluding having one extra language did not delay or accelerate the rate of acquisition. While there seems to be a general consensus nowadays that bilingual children’s two languages develop independently from each other, some researchers do not totally eliminate the possibility of interaction between the two languages during the course of development. It has been reported that some bilingual children produce some morphosyntactic structures that do not appear in monolingual children’s linguistic developmental path. Such structures are thought to be the product of interaction between two languages and they are called cross-linguistic structures. Interesting areas for research arise from the cross-linguistic structures are the extent to which the two languages interact and the conditions that allow such interaction between two languages to occur (e.g., Döpke, 2000; Hulk, 2000; Hulk & Müller, 2000; Müller & Hulk, 2001; Yip & Matthews, 2000). Genesee (2005) comments that these structures are temporary phenomena in the course of bilingual language development because in the end as adults, bilingual children acquire the target languages appropriately. Therefore, while the cross-linguistic structures are interesting, they are not counter-evidence for the hypothesis that the two languages of bilingual children develop separately. In this paper, I report on a study which investigated how one Japanese-English bilingual child developed discourse functions in the two languages. Rather than focusing only on the child’s syntactic development of the two languages, this study considers the interface between syntax and pragmatics. I pay particular attention to the assignment of topic (TOP) by the child in Japanese and English. As it will be explained below, the assignment of TOP is very closely connected to pragmatics. Japanese and English use different methods to assign TOP. I will investigate whether the Japanese-English bilingual child develops to assign TOP in each language in a language specific manner. Further, I will address the relationship between the two languages of the child by comparing the development of the two languages. In order to achieve this comparison, I will apply the Topic Hypothesis (Pienemann, Di Biase, & Kawaguchi, 2005) to the bilingual data. The Topic Hypothesis is an extension of Processability Theory (PT) (Pienemann 1998) and

Development of discourse functions in Japanese and English bilingual first language acquisition

43

focuses on non-canonical mapping of constituent structure onto grammatical functions in relation to pragmatics. This hypothesis predicts the developmental sequence of a second language (L2) learner’s topic assignment. This paper is organised in the following manner. In the next section, I will first describe the typology of Japanese and English. Then I will briefly describe the developmental sequence proposed by the Topic Hypothesis (Pienemann et al., 2005). This is followed by some examples of structural outcomes of Japanese and English predicted to appear in each step of the Topic Hypothesis. The remainder of the paper presents my study, beginning with a description of the methods used in the study, followed by the results obtained for each language, then a comparison of the results obtained for the each language. The last section summarises the findings.

1

Description of Japanese and English grammars

Typologically English is an SVO language. The grammatical functions of arguments are encoded by the word order. The subject (SUBJ) of the sentence is obligatory in the declarative sentences. On the other hand, Japanese is typologically an SOV language. Unlike English, Japanese does not rely on the word order to encode the grammatical functions of arguments. Japanese utilises post-nominal particles to encode the grammatical or semantic functions of arguments. The particles relevant in this paper are the nominative (NOM) marker -ga, the accusative (ACC) marker -o, and the topic (TOP) marker -wa. The nominative (NOM) marker -ga typically marks SUBJ by default. The accusative (ACC) marker -o marks OBJ(ect) by default. The topic (TOP) marker -wa marks SUBJ by default (see Bresnan, 2001) but can topicalise other arguments. (1a) and (1b) below show that both -ga and -wa can mark the SUBJ of the sentences expressing the same event ‘Haruko ate an apple’. (a)

a. b.

Haruko-ga ringo-o Haruko-NOM apple-ACC ‘Haruko ate (an) apple.’ Haruko-wa ringo-o Haruko-TOP apple-ACC ‘Haruko ate (an) apple.’

tabe-ta eat-PAST tabe-ta eat-PAST

44

Yuki Itani-Adams

Particles can be omitted where information is already in the discourse model. There are no person, number and gender features in Japanese. The Japanese word order is flexible to a certain extent due to the fact that the particles encode the grammatical functions. However, V must be placed in the final position (Shibatani, 1990). Japanese also allows ellipsis of an argument where information is already in the discourse model. The examples sentences in (2) below illustrate the nominal ellipsis of the sentence. Nominal ellipsis allows utterances consisting of a verb only, as illustrated in (2d). This occurs when SUBJ, OBJ and/or other information are understood by the listener. This type of utterance is common in TL Japanese. (b) Examples of nominal ellipsis a. no ellipsis Haruko-ga ringo-o Haruko-NOM apple-ACC ‘Haruko ate (an) apple.’

tabe-ta eat-PAST

b. ‘Who ate the apple?’ Haruko-ga Ø Haruko-NOM Ø ‘Haruko ate (an apple).’

tabe-ta eat-PAST

c. ‘What did Haruko eat?’ Ø ringo-o Ø apple-ACC ‘(Haruko) ate (an) apple.’

tabe-ta eat-PAST

d. ‘What did Haruko do to the apple?’ Ø Ø tabe-ta Ø Ø eat-PAST ‘(Haruko) ate (an apple).’ Japanese verbs are composed of the combination of the stem and agglutinative morphemes. The verb stem never occurs alone. In other words, it is always suffixed by at least one morpheme. For example, for the Japanese verb meaning to eat, the past tense form is tabe-ta (= ate) and the non past tense form is tabe-ru (= eat). Tabe-ta (= ate) has the past tense verb morpheme -ta (PAST) suffixing the stem tabe, whereas tabe-ru (= eat) has the non past verb morpheme -u (NONPAST) suffixing the stem.

Development of discourse functions in Japanese and English bilingual first language acquisition

45

Appendix A presents a list of verb morphemes and their gloss abbreviation that appear in this paper.

2

Description of the Topic Hypothesis

The Topic Hypothesis (Pienemann et al., 2005) predicts a developmental path which a second language learner follows with regard to the assignment of TOP. In this paper, I limit the explanation of the Topic Hypothesis to the description of the predicted developmental sequence. I refer readers to Pienemann et al. (2005) for a detailed theoretical explanation. The original PT (Pienemann, 1998) and the Topic Hypothesis look at the developmental path of language learners. While the original PT looks at the developmental path from the point of view of feature unification, the Topic Hypothesis looks at it from the discoursepragmatic point of view. TOP is assigned to the most prominent position within the sentence, which is the sentence initial position. Table 1 shows the developmental sequence proposed by the Topic Hypothesis. Processing procedure S-procedure

Phrasal procedure

Category procedure

Discourse principle Topicalisation of core arguments

c- to f- mapping

Structural outcomes

TOP=OBJ

Ĺ XP adjunction

Ĺ TOP=ADJ

Ĺ Canonical order

Ĺ SUBJ=default TOP

The TOP function is assigned to a core argument other than SUBJ Ĺ Initial constituent is a circumstantial adjunct or a FOCUS WH-word. TOPIC is differentiated from SUBJECT. Ĺ TOPIC and SUBJECT are not differentiated

Table 1: Predicted developmental sequence of the Topic Hypothesis (after Pienemann et al. 2005, Kawaguchi, 2005) There are three steps that L2 learners are predicted to follow in the development of assignment of topic. Initially L2 learners do not

Yuki Itani-Adams

46

differentiate between TOP and the subject (SUBJ) of predicate. In other words, they use SUBJ as the TOP and this is indicated by SUBJ=default TOP on the bottom row of Table 1. In this step, the SUBJ is assigned to the most prominent position, i.e., the sentence initial position, therefore forming a canonical word order such as SVO or SOV. In order to form the canonical word order, the learners use direct mapping of the conceptual structure (astructure in lexical functional grammar (LFG) (Bresnan 2001) term) onto the grammatical functions (f-structure in LFG term) then further onto the linguistic form (c-structure in LFG term) (Pinker, 1984). For example, this direct mapping for the arguments, a cat and a fish in a sentence a cat ate a fish is shown in (c). The processing procedure required for this step to emerge is the category procedure. (c)

agent ʜ SUBJ ʜ cat

theme/patient ʜ OBJ ʜ fish

semantic roles (a-structure) grammatical functions (f-structure) linguistic form (c-structure)

After this initial step, the L2 learners become able to add an ADJUNCT (ADJ) to a canonical string, i.e., XP + canonical word order. Examples of ADJ are expressions for time or place such as yesterday or at school, and focus wh-words such as what. However in the Topic Hypothesis ADJ also includes varieties of clauses that are positioned in the sentence initial position (Di Biase, personal communication August 2006). During this step, ADJ is topicalised and it is indicated by TOP=ADJ. The phrasal procedure is necessary for this step to emerge. The addition of ADJ will lead the learners to the third step, where they become able to differentiate between TOP and SUBJ. Such differentiation will result in the topicalisation of core arguments other than SUBJ. One such example is object (OBJ) topicalisation. This step is indicated by TOP=OBJ in Table 1. The S-procedure is required for this operation. Di Biase (2005), Kawaguchi (2005) and Zhang (2005) demonstrated that adult second language learners of Italian, Japanese and Chinese respectively developed syntax according to the sequence predicted in the Topic Hypothesis. In this study, I apply the Topic Hypothesis to the Japanese-English bilingual data. The application of the

Development of discourse functions in Japanese and English bilingual first language acquisition

47

Topic Hypothesis enables the direct comparison of the development of the two typologically different languages—If each language of the bilingual child develops following the Topic Hypothesis path, the timing of acquisition of each ‘step’ can be compared across the two languages. Different acquisition times of the same ‘step’ across languages would infer that the two languages are processed separately. Different acquisition times would add further support to De Houwer’s (1990) SDH. I present some examples sentences predicted to appear in each step of the Topic Hypothesis in Japanese and English respectively. All of the Japanese examples in (e), (g) and (i) are taken from Kawaguchi (2005). For the initial TOP=SUBJ step, (d) is an English example showing an SVO sentence in which the SUBJ is the TOP (STOP). (e) is a Japanese example showing that the STOP is encoded by the use of -wa, as in Tanakasan-wa. (d) TOP=SUBJ: English: (STOPVO) I read this book. (e) TOP=SUBJ: Japanese: (STOPOV) Tanaka-san-wa kono tegami-o kai-ta Tanaka-Mr-TOP this letter-ACC write-PAST ‘Mr Tanaka wrote this letter’ (Kawaguchi, 2005) Three English examples are presented for the TOP=ADJ step: (6a) indicates the sentence where the ADJ for time is placed in the initial position followed by a SVO sentence; (6b) shows an example of a focus WH-word being placed in the initial position followed by a SVO sentence; (6c) is another example of ‘ADJ+canonical word order’ (Bruno Di Biase, personal communication, August 2006). Although 6(b) is not target language (TL) like, this type of sentence occurs in interlanguage (IL). (f) TOP=ADJ (English) a. Yesterday I read this book. b. What you want it? c. As for Mary, John likes her. The Japanese example (g) shows that an ADJ kinoo (= yesterday) is placed in the initial position and topicalised by the suffixation of –wa. The ADJ is followed by the SOV.

Yuki Itani-Adams

48

(g)TOP=ADJ (Japanese) Kinoo-wa Tanaka-san-ga kono tegami-o kai-ta yesterday-TOP Tanaka-Mr-NOM letter-ACC write-PAST ‘Yesterday, Mr Tanaka wrote this letter.’ (Kawaguchi, 2005) In the final, TOP=OBJ step, the examples for English and Japanese are given in (h) and (i) respectively. For English, the OBJ is placed in the initial position of the sentence. For Japanese, the OBJ is topicalised by the suffixation of –wa and placed in the initial position of the sentence. (h)TOP=OBJ (English) This book, I will read. (i)TOP=OBJ (Japanese) Kono tegami-wa Tanaka-san-ga kai-ta this letter-TOP Tanaka-Mr-NOM write-PAST ‘This letter, Mr Tanaka wrote.’ (Kawaguchi, 2005)

3 3.1

The study methods Informant and data

The informant for this study, Haru 1 , is a first-born female child of a family of an English-speaking father (coded as Father) and a Japanesespeaking mother (coded as Mother). Haru was raised in a ‘one parent one language’ home environment from birth. Father spoke English and Mother spoke Japanese to Haru. English was the language of communication between Mother and Father. I recorded (audio and video) the interaction between Haru and Japanese-speaking and English-speaking adults. Each set of recordings consisted of two sessions—one of approximately 45 minutes of interaction with a Japanese-speaking adult (most of the time with Mother) and one of approximately 45 minutes of interaction with an English-speaking adult (most of the time with Father). The recordings were conducted to capture naturally occurring interaction between the participants. Recordings at monthly intervals were transcribed 1

Haru is a pseudonym.

Development of discourse functions in Japanese and English bilingual first language acquisition

49

orthographically from age 1;11 (one year and eleven month old) to 4;10. Altogether 38 pairs of recordings consisting of 72 recording sessions, i.e., 38 Japanese sessions and 38 English sessions, were transcribed. Transcribed data sets were identified by the sequential number of the transcribed session (1 to 38) followed by the letter J or E for Japanese or English respectively. Monthly pairs of sessions from 1;11 to 2;10, and at tri-monthly 2 intervals from 2;10 to 4;10 were selected for analysis in the appropriate languages (sessions 1 to 13 followed by 16, 19, 22, 25, 28, 31, 34, and 38). Haru’s spontaneous utterances in Japanese and English from selected transcribed sessions formed the corpus for the study. Data were analysed monthly during the first year of the investigation in order to obtain a detailed description of the language development during the period in which monolingual children are reported to show rapid linguistic development. The total duration of transcribed recording was 925 minutes for Japanese data and 1120 minutes for English data. Appendix B provides the coding for each recording session analysed and Haru’s age at each session.

3.2

Analysis

The aim of this study was to investigate whether Haru assigned TOP in Japanese and English in a language specific manner. In order to achieve this aim, I examined whether the different types of topicalisation illustrated in examples from (d) to (i) above were found in each language of Haru’s data. I used the emergence criteria (Pienemann 1998) to determine the acquisition of a structure in each developmental step. I then compared the timing of the acquisition of each step within one language to determine the sequence of development. Finally, to establish whether the developmental steps happen in the two languages at the same time, I compared the timing of acquisition of each step between the two languages. I paid particular attention to two issues when analysing the Japanese corpus. One was the treatment of nominal ellipsis and particle ellipsis. As described above, Japanese does not need to express the arguments of the predicates overtly when the information is already in the 2

i.e., every third month.

50

Yuki Itani-Adams

discourse model. In addition, due to the particle ellipsis the argument assigned to be the TOP need not necessarily be overtly encoded. In the analysis of the corpus, one needs to be careful not to exclude utterances that may have such ellipsis. The other issue relates to a characteristic of child Japanese (JL1) found in previous JL1 research. Hakuta (1982), Hayashibe (1975) and Miyahara (1974) found that Japanese children 3 initially use word order, rather than particles, to mark the grammatical functions for both comprehension and production. I also found this characteristic in Haru’s corpus (Itani-Adams, 2005). For both Haru’s Japanese and English corpora, I first examined the position of V for main clauses with at least one overt argument. For the Japanese corpus, I included all the utterances that had a lexical verb and at least one overt argument in the analysis, regardless of the availability of the marking of the argument(s). Secondly, I examined the availability of an adjunct in the initial position. Thirdly I examined the semantic relation and the grammatical function of the first NP in her utterances. Finally, for the Japanese corpus only, I examined the position and the grammatical function of NPs suffixed by -wa (TOP) in any other clauses. These included single word utterances and equational sentences. The next section presents the results of the analysis. I shall present results from Haru’s English corpus first followed by those from her Japanese.

4 4.1

Results English

TOP=SUBJ For this part of the analysis, I used Haru’s utterances from 1;11 (recording session 1E) to 2;4 (6E), the first six months of the period of investigation. The utterances from the first six-month period provided sufficient data for this analysis. I selected utterances containing a lexical verb with at least one overt argument, and examined the semantic functions of the argument(s) and the position of V in relation to the

3

I use the term ‘Japanese children’ to mean monolingual children who acquire Japanese as their first language.

Development of discourse functions in Japanese and English bilingual first language acquisition

51

arguments. Haru’s English data during this period presented utterances with one argument and utterances with two arguments. Where it was not clear if the noun in the initial position was a vocative or the agent of the predicate (e.g., daddy in daddy hold this), such a noun was not included in the analysis. I will describe the one-argument utterances first. In the one-argument utterances, the V was placed either before or after the argument. There were seven instances where the V was placed after the argument. All of these arguments were SUBJ of the predicate. The first sample occurred at age 2;1 and is presented in(j) 4 . (j)

H

Open . Haru open

2;1,9 (3E, 339)

There were 13 samples in which the V was placed before the argument. These arguments were OBJ of the predicate. The first example is given in (k). (k)

H

Catch ball . catch ball catch ball

2;2,17 (4E, 206)

With regard to Haru’s one-argument utterances, she placed V after SUBJ and before OBJ. This was also the case with her two-argument utterances during the same period. There were 14 utterances with two arguments. All of them had the SVO word order. The first such instance occurred at 2;2 (4E). Two examples of SVO are presented in (l). (l)

a. b.

H H

I can see it . now I’m do it Me do it u:m h

2;2,17 (4E, 641) 2;3,16 (5E, 276)

We can conclude from the above analysis that Haru used the SVO word order from the time she began to produce two word utterances involving a lexical verb. Further, the SUBJ of the predicate was placed in the sentence initial position as the TOP. In other words, SUBJ was the default TOP this period.

4

In this paper English examples are presented thus: example number, then speaker code (H for Haru), followed by the utterance, then age (years;months,days) , session number, and turn number within the session.

Yuki Itani-Adams

52

TOP=ADJ For this part of analysis I first examined the availability of utterances that had an ADJ or WH-word in the initial position followed by an SVO string, forming XP+SVO. The earliest sample is presented in (m). At 2;2, a month after her first one-argument utterance, Haru produced the utterance: now I’m do it. The relevant part of her turn is shown in bold. It can be seen that this utterance has a word expressing time now followed by an SVO I’m do it. (m)

H

I can see it . now I’m do it

2;2,17 (4E, 641)

After the above utterance, there was no further example of Haru producing ADJ followed by SVO until she was 3;9. However, she produced SVO utterances headed by a WH-word during the intervening period. (n) shows two such utterances. In these instances, Haru referred to the hole her father had dug in the ground the day before. (n)

a b

H H

what you do one? what you dug that one?

2;7,28 (11E, 462) 2;7,28 (11E. 464)

These examples showed that Haru began to produce ADJ or WHword in the sentence initial position as the TOP. The first instance occurred when she was 2;2. TOP=OBJ For this part of the analysis I examined instances of object topicalisation in Haru’s English corpus. There were two instances in the corpus. They are presented in (o). The first example (15a), that one I like, occurred when Haru was 4;0, while playing a game of chess with Father. From the context, that one refers to the knight (horse) chess piece. This was the first instance of object being in the sentence initial position followed by SV. During the same session Haru produced 38 other twoargument utterances and none of them had OBJ in the pre-SUBJ position. From this we can conclude that Haru began to use object topicalisation at the age of 4;0. (o)

a. b.

H H

um that one I like im I like horsey Spring I don't know

4;0,15 (28E, 438) 4;7,25 (34E, 288)

Development of discourse functions in Japanese and English bilingual first language acquisition

53

Summary of Haru’s English development Haru’s English sentences were initially formed in the canonical word order of SVO. At 2;1 she produced an SV utterance. Her first instances of SVO occurred at the age of 2;2. At this age Haru also formed an utterance with an ADJ. Then at 4;0 Haru produced an utterance where the object was topicalised by being placed in the sentence initial position. In Topic Hypothesis terminology, the results of analysis showed that Haru reached the TOP=SUBJ step at the age of 2;1, followed a month by the TOP=ADJ step at age 2;2. She reached the TOP=OBJ step at the age of 4;0. The relative timing of her arrival in each step is shown in Figure 1. It is clear that Haru developed her syntax in the sequence predicted by the Topic Hypothesis. An interesting point to observe is that Haru moved from TOP=SUBJ to TOP=AJ rather quickly while the progression from TOP=ADJ to TOP=OBJ took a longer time.

TOP=OBJ

TOP=ADJ

TOP=SUBJ

1;0

2;0

4;0

3;0

5;0

Age (years;months)

Figure 1 Assignment of TOP in Haru’s English

4.2

Japanese

For Japanese, I first examined clauses with a lexical verb that had at least one overt argument. This type of clause began to appear in the corpus from the time Haru was 2;2 (4J). Before that, when Haru used a verb, she used it in single word utterances forming verb-only utterances. In the 285 clauses with a lexical verb and at least one overt argument found in the corpus, all of them had the V in the final position. In these clauses some of the arguments were suffixed with particles but the others were not (i.e., null-marking). I determined the grammatical function of the

Yuki Itani-Adams

54

argument which was placed in the sentence initial position in these clauses. In addition, I examined clauses which contained the -wa (TOP) particle even when they did not contain a V, in order to determine the position (in relation to other constituents within the clause) and the grammatical function of the NP suffixed by -wa (TOP). I first present the results from the argument with the null-marking, followed by the results from the arguments with the -wa (TOP) marking. Arguments without the TOP marker -wa TOP=SUBJ The first occurrence of a lexical verb with an argument was found when Haru was 2;2, and it was a one-argument utterance. It is presented in (p) 5 . The character I indicates the null marking of argument. In this example, the grammatical function of the argument was SUBJ. Haru used her own name haru to refer to herself when young. This phenomenon accords with the findings of JL1 research (Clancy, 1985). This example shows that Haru produced an utterance with SV word order. (p)

H

shu-u haru-I Haru do-NONPAST ‘Haru (will) do (it).’

2;2,13 (4J, 87)

Two months after the first instance of SV shown above, the first instance of one-argument utterances with OBJ was found. The OBJ in these utterances had no particle marking and they were placed in the preverb position. (q) presents an example. (q)

H

mi-te kotchi-I this look-REQ ‘Look at this.’

2;4,14 (6J, 266)

5 In this paper Japanese examples are presented thus: example number, then speaker code (H for Haru), followed by the utterance. The Japanese example utterance is followed by an English gloss underneath, which is then followed by an English translation and age and turn number as for English examples as described above.

Development of discourse functions in Japanese and English bilingual first language acquisition

55

With one-argument utterances, we observed that both SUBJ and OBJ were placed in the pre-verb position respectively (i.e., SV or OV). This position accords with the Japanese canonical word order. In order to find out the word order between SUBJ and OBJ, it is necessary to examine utterance in which both SUBJ and OBJ occur. The first instance of twoargument utterance was found when Haru was 2;6 (8J). It is presented in (r). In this example, the word konchan is the name of a fox that appears in a story. Haruchan ‘Haru + diminutive suffix’ (= Haru) is the SUBJ and konchan ‘Kon + diminutive suffix’ (= Kon, the fox) is the OBJ of the predicate. Although neither argument has any particle marker, the word order is SOV. In other words it is in the canonical word order and the SUBJ is assigned as the TOP in the sentence initial position. (r)

H

kaiteru konchan-I Haruchan-I Haru Kon, ‘the fox’ draw- PROG-NONPAST ‘Haru (is) am drawing Kon, the fox.’ 2;6,18 (8J, 10)

It can be seen from the above analysis that Haru used the Japanese canonical word order as soon as she began to produce utterances with argument and a lexical verb. While the arguments were not suffixed by -wa (TOP), the SUBJ was placed in the sentence initial position. TOP=ADJ There were five instances of ADJ in Haru’s corpus. The first instance occurred when Haru was 3;0: Haru was talking about a film called ‘Chitty Chitty Bang Bang’ referring to the scene where the people in the film went to the beach and took their clothes off to change to swimming suits (s). The phrase ‘In Chitty Chitty Bang Bang’ is placed in the initial position followed by an OV sentence. (s)

H

chiki chiki ban ban toki ne ne ne pantsu-I . chitty chitty bang bang when m m m pants nugu desho? take-off-NONPAST don’t they? In Chitty Chitty Bang Bang (they) take off pants, don’t they? 3;0,22 (16J, 665)

Yuki Itani-Adams

56

The other instances of ADJ all expressed time: kyoo (= today) and ashita (= tomorrow). The first of these is shown in (t). Here the utterance consists of three arguments with grammatical functions SUBJ, OBJ and OBL(ique). The ADJ kyoo (= today) had to compete against three other elements in the utterance, and was placed in the sentence initial position. (t)

H

kyoo sera-I ne kore-I haruchan-I ne today Haru Sarah this yotte 6 iku kara ne take-COMP go-NONPAST ‘Today, Haru (will) take this to Sarah.’ 3;6,14 (22J, 704)

The analysis shows that Haru began to place ADJ with a SOV sentence at the age of 3;0, placing it in the initial position. Among the data without the -wa (TOP) marker, I did not find any instance where OBJ was placed in front of SUBJ. The section below describes the data that had the -wa (TOP) marking on an argument. Arguments with the TOP marker -wa There were 50 instances of -wa (TOP) in Haru’s corpus. The first occurred when Haru was 2;2. Of all the nominal particles, -wa (TOP) was the first to appear in Haru’s corpus. The first usage of -wa (TOP) occurred in single word utterances such as shown in (u). (u)

H

haru-wa? Haru-TOP ‘How about Haru?’

2;2,13 (4J, 937)

According to Clancy (1985), -wa (TOP) is also the first particle to be acquired by Japanese monolingual children and it is initially used with a single noun. Until Haru was 3;6, this is the only way she used -wa (TOP). With regard to the nominative and accusative markers, Haru used -ga (NOM) for the first time at 2;7 and –o (ACC) at 3;6. In other words, Haru

6

In TL Japanese this word is pronounced motte (take-COMP), however, Haru pronounced this word as yotte when young.

Development of discourse functions in Japanese and English bilingual first language acquisition

57

began to use -wa first followed by -ga, with -o last. This order of emergence of particles accords with that found in JL1 studies (Clancy, 1985). At the age of 3;6 Haru began to use -wa (TOP) in equation sentences. The Japanese equation sentence takes the form of A-wa B desu (= A is B). The word B can be a noun, adjective or adjectival noun and the copula desu is placed in the sentence final position. However, in spoken Japanese (TL) the copula desu can be omitted. Haru used this type of sentence with the omission of the copula. (v) shows two examples of Haru’s utterances: (22a) involved a noun and (22b) an adjective. (v)

a.

b.

mama, kore-wa chuurippu mummy this-TOP tulip ‘Mummy, this (is) (a) tulip.’

3;6,14 (22J, 443)

haruchan-wa hayai Haru-TOP fast ‘Haru (is/am) fast.’

3;6,14 (22J, 630)

Until Haru was 4;7, her use of -wa (TOP) was limited to the type of usage shown in examples (u) and (v) above. It suffixed the first or only noun in the utterances. The following section describes Haru’s other usage of -wa (TOP) from the time she was 4;7. TOP=SUBJ The first instance of -wa in an utterance with a verb appeared when Haru was 4;7. It is shown in (w). The relevant part of her turn is in bold. This utterance forms the canonical word order and consists of three arguments: the SUBJ watashi (= I), the OBJ kore (= this) and the OBL dare (= who). Of these three arguments the SUBJ was suffixed with -wa (TOP), and placed in the sentence initial position. (w)

H

dare-ga saki? dare-ga saki? who-NOM first who-NOM first agetai? watashi-wa kore-I dare-I I-TOP who this give- DES ‘Who’s first? Who’s first? Who do I want to give this? 4;7,26 (34J, 480)

Yuki Itani-Adams

58

TOP=ADJ During the same session where we found evidence for the TOP=SUBJ step, Haru also suffixed ADJ kondo (= next time) with -wa (TOP), as shown in (x). This utterance also had an OBJ, although it was not suffixed with a particle. (x)

H

shitara kondo-wa X yo then next time-TOP X tabemashoo pai-I pie eat-POL-COH ‘Then next X. Next, let’s eat a pie’

kondo-wa next time-TOP 4;7,26 (34J, 344)

TOP=OBJ According to the Topic Hypothesis, this third developmental step sees the learners topicalise core arguments other than SUBJ. One such example is OBJ topicalisation. In order to determine the OBJ topicalisation, one needs evidence that the speaker chose to topicalise OBJ over the SUBJ. In other words, it is necessary for data to show the competition between OBJ and SUBJ for the assignment of TOP. In case of Japanese, due to the nominal ellipsis, the speaker can omit SUBJ. While it is TL-like, utterances without SUBJ do not qualify as data for this part of the analysis. Even when OBJ is suffixed with -wa (TOP) in such an utterance, it does not show evidence of OBJ topicalisation. This is because one cannot eliminate the possibility that the suffixation of -wa (TOP) occurred to the OBJ because it was the first or only noun in the sentence. In Haru’s main clauses with lexical verbs, there were a few utterances where the OBJ was marked with -wa (TOP). All of them occurred after she was 4;7. However, these instances were either imperative or in sentences without the overt SUBJ. Imperative sentences naturally do not need to have SUBJ; therefore, there is no competition between OBJ and SUBJ. Hence, these utterances did not qualify as data for this analysis. There was, however, one utterance that may tell us something about Haru’s development. Haru produced the utterance shown in (y) at the age of 4;10. This utterance occurred when Haru was practising writing Japanese letters (hiragana) and she made a comment that she was good at writing the letter for the sound te. (y)

H

te

te-wa

watashi-I

joozu

da

Development of discourse functions in Japanese and English bilingual first language acquisition

te te-TOP I good at ‘The letter ‘te’, I am good at (writing).’

59

COP 4;10,26 (38J, 386)

The utterance (y) does not contain a lexical verb, but has an adjective predicate. The SUBJ of the predicate is watsahi (= I). In (y), the word te which is topicalised by -wa (TOP) is not strictly the OBJ of the predicate. The unmarked topic structure (i.e., canonical) of the same sentence in TL would look like (z). The adjective predicate joozu da (= be good at) is placed in the final position. The name or the thing one is good at, te (= the letter te) in this example, takes the -ga (NOM) marking, and the person who is good at the thing described in the sentence, watashi (= I) in this example, takes either -wa (TOP) or -ga (NOM). (z)

watashi-wa/ga te-ga I-TOP/NOM te-NOM ‘I am good at (writing) the letter te.’

joozu good

da COP

In other words, in the unmarked topic sentence, watashi is the only argument that can take the -wa (TOP) marking. As mentioned above, (y) is not an example of OBJ topicalisation in the strict sense. Nonetheless, by topicalising the word te by suffixing the -wa marker, the speaker is using a marked topic structure, i.e., topicalisation of an element that is not SUBJ or ADJ. Summary of Haru’s Japanese development Haru’s Japanese sentences formed in the canonical word order of SOV from the time she began producing utterances with one verb and at least one noun. The first such instance of a one-argument utterance occurred at the age of 2;2. Her first instances of two-argument SOV occurred at the age of 2;6. Haru began to use ADJ in front of SOV at the age of 3;0. Although Haru placed the SUBJ and ADJ in the sentence initial position at 2;2 and 3;0 respectively they were not marked with the topic marker -wa. Although Haru began to use -wa (TOP) from 2;2, she did not use this particle in a clause with a lexical verb until she was 4;7. During the intervening period, Haru encoded the topic by placing it in the sentence initial position, i.e., by using word order. This developmental sequence of initially relying on position, i.e., word order, prior to using morphology to encode TOP is similar to that found in JL1 studies with regard to encoding SUBJ and OBJ (Clancy 1985).

Yuki Itani-Adams

60

The morphological marking of SUBJ and ADJ as TOP both occurred when Haru was 4;7. Then at 4;10, she assigned an argument other than SUBJ as TOP. Using Topic Hypothesis terminology, the results of the analysis showed that Haru reached both the TOP=SUBJ and TOP=ADJ steps at age 4;7. Subsequently she reached the TOP=OBJ step at 4;10. The relative timing of her arrival in each step is shown in Figure 2. It is clear to see that Haru developed her syntax in the sequence predicted by the Topic Hypothesis.

TOP=OBJ

NP+wa

TOP=ADJ

NP+ĭ

NP+wa

TOP=SUBJ

NP+wa

NP+ĭ

1;0

2;0

3;0

4;0

5;0

Age (years;months)

Figure 2 Assignment of TOP in Haru’s Japanese

5

Comparison between Japanese and English

One of the questions I asked in this study was the relationship between Haru’s two languages. In order to investigate such relationships, I compared the timing of Haru’s arrival at each step defined by the Topic Hypothesis in two languages. If the differentiation of SUBJ from TOP, or OBJ topicalisation, happen at the same time in both languages, we cannot conclude whether or not the two languages are developing separately. However, if the developmental steps happen at different times in each language, then we can conclude that the two languages develop separately. As discussed above, Haru developed both Japanese and English following the developmental path predicted by the Topic Hypothesis. However, the comparison of the development of the two languages revealed that she did not develop them at the same rate.

Development of discourse functions in Japanese and English bilingual first language acquisition

61

Figure 3 shows the relevant timing of acquisition of each step in both languages. From this figure, it is clear that Haru’s Japanese and English reached the same step at different times. Haru acquired the canonical word order in English around the same time as her Japanese canonical word order without morphological marking. However, she acquired the subsequent steps at different times in each language. This indicates that Haru’s two languages developed separately.

Japanese

English

TOP=OBJ NP+wa TOP=AD

NP+ĭ

NP+wa

TOP=SUBJ NP+ĭ

1;0

NP+wa

2;0

3;0

4;0

5;0

Age (years;months)

Figure 3 Assignment of TOP in Haru’s Japanese and English

6

Summary

In this paper, I showed how Haru, a Japanese-English bilingual child, developed the assignment of TOP in the two languages. The results from the analysis showed that she assigned TOP in each language in a language specific manner. In English, she initially put SUBJ in the sentence initial position as TOP using the English canonical word order of SVO. Then, she developed to place ADJ and WH-word in the sentence initial position followed by an SVO sentence. Subsequently, she topicalised an OBJ by placing it in the sentence initial position followed by SV (i.e., OBJ = TOP). This developmental sequence follows the sequence predicted by the Topic Hypothesis. As for Japanese, the notion of TOP was first encoded by position, i.e., word order, with the V in the sentence final position (SOV). Later, TOP was grammatically encoded by the use of the topic marker -wa (TOP).

Yuki Itani-Adams

62

When Haru began to encode the TOP grammatically, she encoded both SUBJ and ADV as TOP at once. She suffixed them with -wa (TOP) and placed them in the sentence initial position respectively. After that, she produced an utterance in which non-SUBJ element was topicalised by the suffixation of -wa (TOP) and this was also placed in the sentence initial position. Although a clear step between the TOP=defaultSUBJ stage and the TOP=ADJ stage in Haru’s Japanese corpus was not detected, her developmental sequence does not counteract the sequence predicted by the Topic Hypothesis. The application of the Topic Hypothesis enabled the direct comparison of the development of Haru’s Japanese and English, two typologically different languages. The comparison showed that each developmental step happened at different times in Japanese and English. This study showed that, while the two languages both developed in the sequence predicted by the Topic Hypothesis, they developed separately. Hence, the results of this study indicated that De Houwer’s (1990) Separate Development Hypothesis also holds for the development of TOP in Japanese and English by a bilingual child.

7

References

Bresnan, J. (2001): Lexical-functional syntax. Oxford: Blackwell Publishers. Clancy, P. (1985): “The acquisition of Japanese.” In The Crosslinguistic Study of Language Acquisition,Vol. 1, edited by D. I. Slobin. Hillsdale: Lawrence Erlbaum: 373-524 De Houwer, A. (1990): The Acquisition of Two Languages from Birth: A Case Study. Cambridge: Cambridge University Press. Di Biase, B. (2005, 26-28 September): “The topic hypothesis in Processability Theory: the syntax-pragmatics interface in L2 development.” Paper presented at the The 5th international symposium on processability, second language acquisition and bilingualism, Deakin University, Australia. Döpke, S. (2000): “The interplay between language-specific development and crosslinguistic influence.” In Cross-linguistic structures in simultaneous bilingualism edited by S. Döpke. Amsterdam/Philadelphia: John Benjamins: 79-104 Genesee, F. (1989): “Early bilingual development: one language or two?” Journal of Child Language, Vol.16 (1989): 161-179. Genesee, F. (2000): “Introduction.” Bilingualism: Language and Cognition, Vol.3/3 (2000): 167-172.

Development of discourse functions in Japanese and English bilingual first language acquisition

63

Genesee, F. (2005): “The capacity of the language faculty: Contributions from studies of simultaneous bilingual acquisition.” In ISB4: Proceedings of the 4th International Symposium on Bilingualism edited by J. Cohen, K. T. McAlister, K. Rolstad & J. MacSwain. Somerville, MA: Cascadilla Press: 890-901. Hakuta, K. (1982): “Interaction between particles and word order in the comprehension and production of simple sentences in Japanese children.” Developmental psychology, Vol.18 (1982): 62-75. Hayashibe, H. (1975): “Word order and particles: a developmental study in Japanese.” Descriptive and Applied Linguistics, Vol.8 (1975): 1-18. Hulk, A. (2000). “Non-selective access and activation in child bilingualism: The syntax.” In Cross-linguistic structures in simultaneous bilingualism edited by S. Döpke. Amesterdam/Philadelphia: John Benjamins: 57-78. Hulk, A., & Müller, N. (2000): “Bilingual first language acquisition at the interface between syntax and pragmatics.” Bilingualism: Language and Cognition, Vol.3/3 (2000): 227-244. Itani-Adams, Y. (2005): “Exploring the interface between lexical and morphosyntactic development in Japanese–English bilingual first language acquisition.” In (Ed.), CAESS Conference: Scholarship & Community edited by M. Atherton. Sydney: University of Western Sydney: 1-15. Kawaguchi, S. (2005, 26-28 September): “Lexical mapping theory in Processability Theory.” Paper presented at the The 5th international symposium on processability, second language acquisition and bilingualism, Deakin University, Australia. Meisel, J. M. (Ed.). (1990): Two first languages: Early grammatical development in bilingual children. Dordrecht: Foris. Meisel, J. M. (Ed.). (1994): Bilingual first language acquisition: French and German grammatical development. Amsterdam: John Benjamins. Mishina, S. (1997): Language separation in early bilingual development: A longitudinal study of Japanese/English bilingual children. PhD dissertation. University of California Los Angeles. Miyahara, K. (1974): “The acquisition of Japanese particles.” Journal of Child Language, Vol.1 (1974): 283-286. Müller, N., & Hulk, A. (2001): “Crosslinguistic influence in bilingual language acquisition: Italian and French as recipient languages.” Bilingualism: Language and Cognition, Vol.4/1 (2001): 1-21. Paradis, J., & Genesee, F. (1996): “Syntactic acquisition in bilingual children.” Studies in Second Language Acquisition, Vol.18 (1996): 1-25.

64

Yuki Itani-Adams

Pienemann , M. (1998): Language Processing and Second Language Development : Processability Theory. Amsterdam: John Benjamins. Pienemann, M., Di Biase, B., & Kawaguchi, S. (2005): “Extending processability theory.” In Cross-linguistic aspects of processability theory edited by M. Pienemann. Amsterdam: John Benjamins: 199-253. Pinker, S. (1984): Language Learnability and Language Development. Cambridge, MA: Harvard University Press. Shibatani, M. (1990): The Languages of Japan. Cambridge: Cambridge University Press. Taeschner, T. (1983): The sun is feminine. Berlin: Spinger-Verlag. Volterra, V., & Taeschner, T. (1978): “The acquisition and development of language by bilingual children.” Journal of Child Language, Vol.5 (1978): 311-326. Yip, V., & Matthews, S. (2000): “Syntactic transfer in a CantoneseEnglish bilingual child.” Bilingualism: Language and cognition, Vol.3/3 (2000): 193–208. Zhang, Y. (2005, 26-28 September): “Topic hypothesis and L2 Chinese syntax.” Paper presented at the The 5th international symposium on processability, second language acquisition and bilingualism, Deakin University, Australia.

Development of discourse functions in Japanese and English bilingual first language acquisition

65

Appendix A: Description of verbal morphemes and their abbreviation Verbal morpheme -te -ta -u -oo -tai -teru

Description request past nonpast Cohortative Desiderative progressive

Gloss abbreviation REQ PAST NONPAST COH DES PROG

Yuki Itani-Adams

66

Appendix B: Sessions Selected for Analyses—Codes and Haru's Age Japanese Haru's age English Haru's age session (year;months,day) session (year;months,day) 1J

1;11,12

1E

1;11,12

2J

1;11,27

2E

1;11,29

3J

2;1,12

3E

2;1,9

4J

2;2,13

4E

2;2,17

5J

2;3,16

5E

2;3,16

6J

2;4,14

6E

2;4,15

7J

2;5,11

7E

2;5,10

8J

2;6,18

8E

2;6,19

9J

2;6,26

9E

2;6,26

10J

2;7,6

10E

2;7,6

11J

2;7,24

11E

2;7,28

12J

2;9,15

12E

2;9,9

13J

2;10,5

13E

2;10,5

16J 19J

3;0,22 & 3;0,24 3;3,3

16E 19E

3;0,21 3;3,0

22J

3;6,14 & 3;6,15

22E

3;6,15 & 3;6,17

25J

3;9,15

25E

3;9,12

28J

4;0,21

28E

4;0,15 & 4;0,20

31J

4;3,23

31E

4;3,23

34J

4;7,26

34E

4;7,25

38J

4;10,26

38E

4;10,25 & 4;10,27

THE ACQUISITION OF THE PASSIVE VOICE IN L2 ENGLISH: PERCEPTION AND PRODUCTION DAGMAR KEATINGE JÖRG-U. KEßLER

1 Introduction In this paper we explore the development of L2 English passive voice constructions within a Processability Theory (PT) framework (Pienemann 1998 and 2008) in various settings. Our study allows us to find interlanguage variants for the passive voice. Thus, we hypothesize that the spontaneous oral production develops gradually from what we henceforth refer to as a Pseudo Passive 1 to the target like production. This gradual development undergoes several stages which can be related to existing PT stages for L2 English. In order to test the gradual development of the passive voice in L2 English we examine the interlanguage development both in perception and production.

 1

We are aware that the term Pseudo Passive is not yet established in the literature. However, as the development of the passive voice undergoes various non-target like stages before the L2 language processor is finally able to handle this highly complex structure in a target like manner we felt the necessity to account for the L2 learners’ increasingly processing power before the actual target like production. As this study has been conducted within the PT framework we do not discuss the acquisition of the passive voice from a mastery point of view. In order to decide whether a passive voice structure has been acquired by a learner we use the emergence criterion (Pienemann 1998; for a discussion of the emergence criterion see Keßler 2006 and Pallotti, 2007). It is important to note that the acquisition of the target like passive does not imply that the learner will always apply this structure. The reader who is familiar with PT will recall that the acquisition of a certain structure cannot be measured in terms of accuracy.

The Acquisition of the Passive Voice in L2 English 

68

Our overall research questions are: 1. 2. 3.

How does the passive voice develop from the understanding of the semantic concept (perception) via Pseudo Passive forms to a target like production? How can these different stages be related to existing PT stages for L2 English? Which are – in PT terms – the prerequisites for the perception, the production of Pseudo Passive structures and for a target like production of the passive voice by L2 English learners?

Being in line with PT’s universal applicability to any language we hypothesize that in the acquisition of the passive voice the gradual development from perception to production would be the same for any learner regardless of the learner’s L1 and the setting in which the acquisition takes place. On the other hand, obviously individual learners do not all produce exactly the same Pseudo Passives. We claim that the different Pseudo Passive constructions are not caused by the learners’ L1 but rather due to individually different states of the learners’ interlanguage development. Until today there have only been very few studies on the acquisition of the passive voice in L2 English (e.g. Tomlin 1995, 1997). Within the PT framework, Kawaguchi (2005, 2008) has worked on the acquisition of the passive in Japanese as L2. Apart from our study the only study on the acquisition of the passive voice in English within the PT framework is the one by Wang also published in this volume 2 . In contrast to Wang (this volume), who exclusively studied the online production of L2 English passive construction by Mandarin speakers we apply a larger range of different communicative tasks both for the perception as well as the production of the passive voice by L2 English speakers. As can be seen in chapter 4 of our paper we tested both EFL learners in Germany as well as ESL learners in England. Additionally we tested the perception and production of the passive with child learners of L2 English in different settings.

 2

In a replication study of Wang’s unpublished MA thesis Zhang (2008) collected additional data on the acquisition of the passive voice by Mandarin speakers and reported about her findings at the 7th International Symposium of Processability, Second Language Acquisition and Bilingualism held at the University of Newcastle upon Tyne in September 2008. However her findings have not yet been published.

Dagmar Keatinge & Jörg-U. Keßler

69

Comparing the two studies on the acquisition of the passive voice in L2 English in this volume we present a complimentary rather than a competing study. However, some findings of our study can directly be compared with the data discussed in Wang (this volume) as in both studies the subjects had to undergo Tomlin’s (1995, 1997) Fish Film experiment. In the following sections we will first provide a very basic summary of the underlying ideas of PT its extension to Lexical Mapping Theory and the Topic Hypothesis (Pienemann, Di Biase & Kawaguchi 2005). In chapter 2 we sketch out our hypothesis in detail. The empirical study and the corpora used to test our hypotheses are presented in chapter 3, followed by a brief discussion of the overall research design. Our tasks are introduced in chapter 4. This is done in some detail as we believe that our study has found some interesting Pseudo Passive structures which might be substantiated by replication studies. In chapter 5 we introduce and discuss the results gained from a detailed data analysis. The discussion of the results shed light on our hypothesis and explains our concept of the Pseudo Passive in more detail. Finally in our conclusion we discuss the implication of our findings as a contribution to theory development within the PT framework as well as the application of our results to the EFL classroom. Additionally, implications for further research will also be discussed in this concluding part of our paper.

2. Extended Processability Theory Before we take a look at the acquisition of the passive voice in the developmental path of L2 English in detail, we will provide a short overview of the theoretical framework in which our study is positioned, namely Processability Theory and extended Processability Theory 3 . “It is the aim of Processability Theory [PT] to hypothesise, on the basis of the general architecture of the language processor, a universal hierarchy of processing resources which can be related to the requirements of the specific procedural skills needed for theTL [target language]” (Pienemann 2005: 3)

 3

For a more detailed introduction to PT see Pienemann (2008). As part of PT is modelled in Lexical-Functional Grammar (Bresnan 2001) the reader might want to read up on LFG, too. An excellent summary of LFG is provided by Fabri (2008).

70

The Acquisition of the Passive Voice in L2 English 

PT provides a theoretical framework which addresses two main problems in explaining language acquisition: the developmental problem and the logical problem (cf. Pienemann 2005: 73f). In order to account for the psycholinguistic dimension of language processing and acquisition, PT incorporates Levelt’s (1989) and Kempen and Hoenkamp’s (1987) model of language processing. The key assumption is that language processing is autonomous due to the high speed at which it takes place. The second language learner cannot utilise internal language processing mechanism in the same way a mature native speaker can and is therefore constrained in his ability to process language. In order to process language at high speed, the speaker is required to store grammatical information about parts of the sentence s/he produces in short term memory. Depending on the availability of particular processing procedures, the learner can unify this grammatical information at different levels of constituents. This process of unification is necessary because the output in language production is linear but the mental processes of language production are not. Now if a processing procedure is not available to the language learner, the whole system will be shut off and the utterance will be produced in a linear fashion (cf. Pienemann 1998: 54 ff). In total, 6 procedures (=6 stages) which have to be gradually acquired by the speaker, have been identified so far (see table 3.2). If, for instance the S’- procedure (stage 5) is not available to the learner, s/he is not able to produce a question with the structure WH-Aux- S – V – (X) (=Why does the dog bark). Instead, the speaker would produce a sentence with the structure WH – S – V – (X) (=Why the dog bark) , violating the insertion of an Auxiliary. The hierarchy of processing procedures is implicational (cf. Pienemann 1998 and 2005). This means, procedure 4 can only be acquired after 1, 2 and 3 and not after 1 and 2, leaving out 3. For an overview of the processing hierarchy and its application to English as a Second Language (ESL), see figure 1 below. The second important feature of PT is the incorporation of Lexical Functional Grammar (LFG; Bresnan and Kaplan 1982, Bresnan 2001) as a grammatical formalism. The diagram below illustrates 3 parallel structures in LFG: a(rgument)-structure is mapped onto f(unctional)-structure. The semantic argument roles like ‘agent’, ‘beneficiary’, ‘experiencer’, etc. are mapped onto grammatical functions. This is the process of lexical mapping. C(onstituent)- structure is mapped onto f(unctional)- structure via the process of feature unification. Both processes play a significant role in the explanation of SLA within the PT framework.

Dagmar Keatinge & Jörg-U. Keßler

71

Figure 1. Three paralles structures in LFG (Pienemann et al. 2005: 200) To account for the acquisition of the passive voice, in particularly the Lexical Mapping Theory is of significance. “Lexical Mapping Theory accounts for the mapping of argument structure onto functional structure” (Pienemann 2007: 145). The predicate of the example sentence ‘Peter sees a dog’ is ‘see’ with the core argument roles . Each verb is annotated in the lexicon with its core argument structure. The universal argument roles are then mapped onto the grammatical functions: SUBJ=experiencer and OBJ=theme. The argument roles are organised hierarchically: agent> beneficiary> experiencer/goal> instrument> patient/theme>locative (Bresnan 2001: 307) with the most prominent argument role to the left. Each argument role can, however, be expressed in different grammatical forms. In the example sentence, the most prominent argument role ‘experiencer’ is the subject of the sentence. If the example sentence is changed into the passive sentence ‘A dog is seen by Peter’, the argument roles are mapped onto different grammatical functions and the less prominent argument role (theme) takes the most prominent grammatical form (subject), causing a greater degree of nonlinearisation. “The relationship between argument structure and the two other levels of structure is variable in a specific language, and it also varies between languages. The variable relation between what is intended to be said (argument structure) and the way it is expressed using grammatical forms creates expressiveness in language but it also creates what Levelt (1981) calls “the linearisation problem”. (Pienemann 2007: 144)

72

The Acquisition of the Passive Voice in L2 English 

This means the respective structure is more difficult to process for the language learner and is hence acquired at a later stage than sentences with linear mapping (cf. Pienemann et al 2005: 206 ff). In early stages of the L2 acquisition process, L2 learners are not able to map the less prominent semantic role onto subject function. The initial state in L2 acquisition is the unmarked alignment, which results in canonical word order. L2 learners rely on a one- to- one mapping principle of argument roles onto grammatical functions, The Unmarked Alignment Hypothesis predicts a default mapping of the most prominent semantic role onto SUBJect function, which is the most prominent grammatical role (cf. Pienemann 2007: 145 and 2005: 229 f). This explains why, at this stage, learners are not able to alter this unmarked alignment. Figure 1: Lexical Mapping Theory- One-to-one-correspondence (Pieneman et al 2007: 230)

a-structure

f-structure

agent

patient/ theme locative

default

default

SUBJ

OBJ, OBJ

OBL,

default

default

default

c-structure

default

S NPsubj

NPobj

[…]



3. Hypotheses The inclusion of the Lexical-Mapping-Theory and the Unmarked Alignment Hypothesis into PT now enables us to propose a number of hypotheses regarding the perception and production of the passive voice, which we are going to test empirically. In section 2, we have shown that other syntactic structures included in the developmental sequence for

Dagmar Keatinge & Jörg-U. Keßler

73

English as L2 develop gradually. When learners have not acquired stage 5, which is a prerequisite for the target- like production of a WH- Question, they instead use a number of interlanguage variants (cf. Pienemann 2008: 11) He live here? Where he live? Where does he live

SVO-Question WH-SVO WH-Aux-second

Stage 2 Stage 3 Stage 5

Hypothesis 1: The passive voice develops gradually We predict that the passive voice also develops gradually and learners produce a number of interlanguage variants at different stages of the acquisition process, the pseudo- passive. These interlanguage variants correspond to the increasing processing complexity of the non- linear mapping as proposed in the Topic Hypothesis. As our informants have a number of different first languages and are learning English in a formal environment and as a second language, we also have to investigate if the interlanguage structures produced by the learners differ or are the same regardless of their L1 and acquisition context. Due to the fact that the development of L2 processing procedures is universal and does not depend on the L1 (cf. Pienemann et al. 2005: 85), we predict that the interlanguage variants will follow the same developmental pattern despite the L1 background. Hypothesis 2: Perception of the passive voice is constrained by processing procedures; however, perception does not require productive processing skills of this structure In addition, the prerequisites for the comprehension of the passive voice are also investigated in our study. Pienemann (2007:137) states that “(...) at any stage of development the learner can produce and comprehend only those L2 linguistic forms which the current state of the language processor can handle.” This key proposition of PT allows us to claim that the production and comprehension of the passive voice are constrained by the same processing procedures. However, the comprehension of any structure included in the developmental sequence does not require productive processing skills; the comprehension of the passive voice will therefore occur at an earlier stage of the acquisition process and will occur before or at the same time as the learner starts to produce one of the interlanguage variants of the passive.

74

The Acquisition of the Passive Voice in L2 English 

Hypothesis 3: Perception and Production of the passive voice is independent of input frequencies and teaching We further hypothesize in this context that both, the comprehension and production of the passive are independent of input frequencies and teaching. A number of studies (Pienemann 1984, 1985 and 1989 have empirically tested to see if the developmental pattern can be changed through formal instruction. These studies have shown that it is not possible to change the developmental pattern, as each stage requires the processing procedures of the underlying stages. As the same processing procedures are required to produce and comprehend the passive voice, the acquisition of the passive voice will obviously follow the same mechanisms. Hypothesis 4: Priming will lead to an increase of production Hypothesis 3 will also be checked against possible priming effects. We propose that priming will lead to an increase in production of interlanguage structures but not necessarily to a target like production. Hypothesis 5: Stage 6 is the prerequisite for “target-like” production of the passive voice As we have claimed that the development of the passive voice is gradual (cf. Hypothesis 1) the question arises as to when learners are developmentally ready (cf. Mansouri & Duffy 2005) for a target-like acquisition of the production of the passive voice in L2 English. Due to the processing complexity of the passive voice (cf. Kawaguchi 2005: 269 ff, Pienemann 2007: 145 ff), the target-like production of the passive voice is only possible when the learner has acquired stage 6 of the developmental hierarchy of English as L2. At earlier points in the acquisition process, the learners attempt to produce the passive when there is a communicative need and will produce a non- target like structure.

4. Empirical Study In this section, we are going to present the research design for our studies and describe the methods of data elicitation and analysis for the production and comprehension of the passive voice.

Dagmar Keatinge & Jörg-U. Keßler

75

4.1. The informants In order to investigate the research questions and to test our hypotheses we looked at interlanguage data from 62 learners of English at different stages of interlanguage development. Additionally – mainly in order to test the naturalness of the tasks – we applied our tasks to 10 native speakers of English. To study the acquisition of the passive voice in various settings of English L2 acquisition we collected data in two sets of environments. Table 1 provides an overview of all 72 informants Table 1: Overview of the informants Total # 72 17 12 10

PTStages

L1

6 3-6* ---

Various Various English

11 3-5 10 3-6* 12 4-6 Total: 72

German German German

Age Adults Adults Adults and adolescents 10-11 13-14 14-15

SubCorpus E E N K D C

L2 English ESL ESL --(control) EFL EFL EFL

The 39 informants of the E and N corpora were interviewed in Newcastle (UK). The E-corpus was subdivided into two sets of non-native speakers: The first subset was made up of 17 overseas students who attended MA courses at Newcastle University. They had to pass an entry test and their interlanguage was highly developed; all these learners’ interlanguage development had reached stage 6 of the PT hierarchy (cf. Pienemann 1998, 2008). Their L1s were typologically rather different and included among others German, Finish, Chinese, Greek, Hungarian, and French. The second subset of the E-corpus consists of 12 overseas students who attended a preparatory language course at the University of Northumbria. By the time of the interviews their command of L2 English was not yet good enough to be accepted in one of the university’s MA-programmes. According to the PT-hierarchy their interlanguage ranged from stage 3 to a very early stage 6. Most of those informants spoke a variety of Chinese (Cantonese, Mandarin or Taiwanese) or Arabic as their L1.

76

The Acquisition of the Passive Voice in L2 English 

We also applied our passive tasks to ten L1 English speakers in order to test the naturalness of our tasks. In our N-corpus, all the native speakers of English produced a large number of passives with our tasks. However, a clear priming effect in favour of a larger number of passive constructions was to be noticed with those five L1 English speakers who had first been asked to work with Tomlin’s (1995 and 1997) Fish Film (see section 4.3 – the tasks for further details). One aim of our study was to test our hypotheses with learners from different learning environments. This is why we did not only collect data from adults in an ESL setting but also from children and adolescents who learn English in an instructed setting outside the country of the target language. Therefore, we interviewed 11 grammar school children in year 5 in Germany. In our K-corpus we analysed the interlanguage of those learners aged 10 and 11. Their interlanguage ranged from stages 3 to 5. All learners spoke German as their L1. Additional data was gathered in a secondary school in Germany where we looked at learners from year 7 (D-corpus; age 13-14) and year 8 (Ccorpus; age 14-15). Again the learners’ L1 was German. The D-corpus consists of 10 learners and the C-corpus consists of 12 learners. By choosing learners from rather different learning environments (ESF vs. EFL) as well as from different L1 backgrounds and age groups we catered for a data basis which allows us to draw highly generalised conclusions. In addition with our native speaker control group the selection of different L1 backgrounds, age groups and learning environments provides a broad basis to test our hypotheses and make claims about the acquisition of the passive voice both in perception and production by the different learners. In the following section we will discuss the research design of our study in more detail.

4.2. Research Design Our study examines the acquisition of the passive voice in L2 English. We hypothesize that the passive voice develops gradually in the learners’ interlanguage development and that learners are able to correctly perceive passive constructions at an earlier developmental stage than would be required for spontaneous production of the passive voice (cf. section 2 to have a closer look at our hypotheses).

Dagmar Keatinge & Jörg-U. Keßler

77

As we hypothezise a different pace for the comprehension and production of the passive voice in L2 English we designed our study as two sub studies: In a first step we elicited spontaneous oral speech data to establish the current state of interlanguage development within the PT hierarchy (Pienemann 1998). In order to elicit this oral data we applied a range of elicitation tasks as specified in Pienemann (1998:280f). Table 2 provides an overview of all the tasks we used at this stage of our study. The data produced by the learners was then used to create linguistic profiles using Rapid Profile (cf. Keßler 2007 and 2008; Keßler & Keatinge 2008). Table 2: Overview of the elicitation tasks (linguistic profiles) task type Habitual Action Story Completion Picture Sequencing Spot the difference

Elicited structures 3-sg-s (third person singular ‘s’) Wh-questions questions (general) questions (general); negations and SVO

Knowing about the individual learners current levels of Interlanguage development we asked the learners to work on carefully designed tasks to test both the perception as well as the production of the passive voice. For testing the perception of the passive voice we restricted ourselves to investigate the learners from the EFL corpora. All but the learners from the K-corpus had been taught the passive voice in their English lessons. We wanted to prove that the perception and the spontaneous production of the passive voice were not to be taught as a variational feature – suggesting that it could be taught and acquired at any stage of interlanguage development – but rather a developmental feature. This is why we chose learners in the EFL corpora who had only partly been taught this grammatical structure in their English lessons before we elicited our data with them. According to the split-half-method half the EFL learners were asked to work on the perception tasks first and then continue with the production tasks. The other half of the EFL learners worked on the tasks in the reversed order. All informants worked on three production tasks. Additionally the EFL learners worked on one further task to elicit the perception data. Table 3 provides an overview of all the tasks.

78

The Acquisition of the Passive Voice in L2 English 

Table 3: Overview of the elicitation tasks (passive voice) task type Sentence Comprehension Fish Film (Tomlin 1995/1997) Sentence Completion Story Telling

Perception/Production Perception (only EFL learners) Production and Priming Production Production

As can be seen from table 3 we used Tomlin’s Fish Film. Before we discuss all the tasks in section 4.3 let us briefly explain the underlying idea of syntactic priming as set out in our study. The Fish Film was shown to all informants including the native speakers (N-corpus). We used this task to test whether informants who were primed by “being forced” to produce a lot of passive constructions in the Fish Film actually produced more passive constructions in the other production tasks. Therefore half of the informants of each corpus worked on the fish film as the first task in the second sub study. In Tomlin’s Fish Film two fish swim towards each other. When they meet one fish eats (respectively is eaten) by the other one. When seeing the film the informants have to create sentences using a fish either as the agent or the patient of the sentence, depending on which of two fish is depicted by a little arrow to be the agent of the sentence. The setup of the film creates some time pressure for the production of each sentence – this guarantees spontaneous speech production by the informants. Comparing the two groups of informants, those informants (both native as well as non-native speakers of English) who did the Fish Film tasks first produced a significantly higher amount of passive voice constructions with the other production tasks than the ones who worked on the other tasks before doing the Fish Film. We will come back to this observation when discussing the results of our study and suggest some consequences for instructed L2 English acquisition. Let us now take a closer look at the different tasks specified in tables 2 and 3 before the results are introduced and discussed in section 5.

Dagmar Keatinge & Jörg-U. Keßler

79

4.3. The Tasks In the first part of the data collection communicative tasks were administered to elicit the subjects’ current state of interlanguage development in L2 English. Numerous examples for these profiling tasks have been published (e.g. Keßler & Kohli 2006, Keßler 2008, Keßler & Keatinge 2008). Therefore, in this paper we restrict ourselves to introduce the tasks for the elicitation of the passive voice (cf. table 3). The naturalness of our tasks was tested with native speakers of English and proved to provide a natural context for the perception and production of the passive voice. However, a striking priming effect could be observed with the L1 English speakers in their production of the passive voice when they first worked on Tomlin’s Fish Film before turning to our own passive tasks (cf. section 4.3.2). We will come back to this in more detail in section 5. 4.3.1 The Perception Task In order to test the perception of the passive voice a semantic decision task that is similar to the Fish Film was designed (cf. Keatinge, in prep.). A list of pictures (cf. figure 2) was given to the subjects. Then a sentence (either passive or active voice) was read twice to the subjects who had to point at the correct picture. The test included 20 sentences in total. Let us look at line 16 in figure 2: In the left column, a blue fish eats a green fish or, respectively, the green fish is eaten by the blue one. The right column shows exactly the opposite picture. For this line of pictures the following four sentences were possible: a) The blue fish eats the green fish. [left column, active voice] b) The green fish is eaten by the blue fish. [left column, passive voice] c) The green fish eats the blue fish. [right column, active voice] d) The blue fish is eaten by the green fish. [right column, passive voice]

80

The Acquisition of the Passive Voice in L2 English 

Figure 2: Example for the Perception Task (Passive Voice)

Each subject was asked to listen to twenty sentences (10 in the active and 10 in the passive voice) and to point at one of the pictures for each sentence. The correct matches were counted; for each correct match one point was given. The subjects could reach a result between o and 20 points. In the data analysis in section 5 we will provide an overview of the overall results for the L2 learners 4 . 4.3.2 The Production tasks For the elicitation of the production of the passive voice we used four different tasks with increasing complexity. This increasing complexity was created by the growing number of lexical items needed to complete the tasks as well as by the increasing openness of each task. First we used Tomlin’s Fish Film, followed by two sentence completion tasks. The data elicitation was completed by a story telling task in which a set of pictures was used to tell a story about an accident. In this task the learners were free to choose between active or passive voice to complete their story. In

 4

All native speakers scored 100% in this task.

Dagmar Keatinge & Jörg-U. Keßler

81

the following section we will describe the tasks in more detail before we can then introduce and discuss the results of our study. 4.3.2.1 The Fish Film (Tomlin 1995 and 1997) As Wang (this volume) we also used Tomlin’s Fish Film in our study to elicit part of the production data. We used the Fish Film for a number of reasons: First, it is well established in SLA research and its validity has been tested and documented. Second, in order to study possible differences between the perception and the production of the passive voice and a possible matching of perception and production according to the PT hierarchy (Pienemann 1998 and 2005) we needed a production task as close to our perception task as possible. Third, in order to compare our data to Wang’s study (this volume) we decided to administer at least one task that was also administered to Wang’s informants. Figure 3 provides an example of the Fish Film and includes some of the subjects’ productions:

The red fish is eaten by the blue fish. The blue fish eats the red fish. (avoidance) (The red fish is the snack from the blue fish. (avoidance) *The red fish be eat by the blue fish. (violation)

Figure 3: Fish Film: Interlanguage structures As the validity of the Fish Film for the production of the passive voice has already been well established only 50% of the native speakers were asked to do this task. Of course all native speakers who worked on this task scored 100% 5 .

 5

We asked only half of the native speakers because we wanted to test whether a person who has been primed to use the passive voice by doing the Fish Film task would produce a higher amount of passive voice structures with our other production tasks (cf. Hypothesis 4). Actually, this priming effect could be

The Acquisition of the Passive Voice in L2 English 

82

4.3.2.2 Sentence Completion We designed two sentence completion tasks. The fist one was the less complex one and administered to the subjects from the K, D and C Corpus (cf. table 1). The second one was designed as a more complex one – this one was administered to the subjects from the E Corpus (cf. Keßler 2006). Figure 4 shows some examples for both tasks. Figure 4: Examples for the sentence completion tasks (left: K,D,C corpus; right: E corpus)

In the task for the K, D, and C corpus the learners had to produce passive voice constructions such as: (1) (2) (3) (4)

The book is/was written by Cornelia Funke 6 /by the lady. The tree is/was planted by the farmer. The mouse is/was eaten by the cat. The postman is/was bitten by the dog.

 observed with all subjects who had first worked on the Fish Film. However, the priming effect was more successful with the native speakers. See section 5 (results) for more details. 6

Cornelia Funke is an extremely popular author in Germany with a growing international reputation. Her major works include the Ink Heart Trilogy and Thief Lord – both these books have actually been released as Hollywood film productions.

Dagmar Keatinge & Jörg-U. Keßler

83

As the subjects of the E corpus were young adults the task for this group was more complex. They were asked to have a look at the sentence heads and use the picture as a cue to complete the sentence. Here the subjects could choose to produce either active or passive voice constructions. A range of answers was possible. Typical sentences were: (5) Peter has a car. The car is filled at the petrol station. (6) Peter has a car. The car is white/nice. (7) We have a dog. The dog is small. (8) We have a dog. The dog is taken for a walk. (9) They invited some friends for dinner. The soup was cooked by my wife. (10) They invited some friends for dinner. The soup was hot/delicious. 4.3.2.3 Story Telling The Story Telling task was designed to serve two basic purposes: first it was another production task to elicit passive voice structures. As the passive voice is a non-obligatory structure the subjects were free to either produce the passive or not. The task was set up as a set of pictures 7 about a bike accident. Second, the task further tested our priming hypothesis. First we administered this task to the native speakers in order to test its naturalness for the elicitation of passive voice structures. Actually, the same priming effects as discussed in section 3 were to be observed here again: Those L1 English subjects who had completed the Fish Film task produced a significantly higher number of passive voice structures as compared to those native speakers who had not worked on the Fish Film. In the next step this task was given to the L2 English group. This task was given to the subjects at least seven days after the session when they

 7

All pictures were taken from various internet sources. Due to copy right issues we cannot print them in this paper. In the pictures there was a mountain biker who had an accident. The next picture showed an ambulance from the outside, then from the inside. In the following picture a person was rushed into a hospital on a stretcher. Then there was a picture of an x-ray of a torso with some broken ribs. The last three pictures showed a person in a hospital bed who is visited by somebody; the next picture showed a police car and a police officer, and the last picture showed a news reporter who interviewed the biker.

84

The Acquisition of the Passive Voice in L2 English 

had to work on the sentence completion tasks discussed in section 4.3.2.2. This was done in order to test whether the priming effect would continue to work even after a seven day break. Here as well some priming effect could be traced, however, the priming proved to be less significant as compared to the native speakers group (Keßler 2006).

5. Results and Discussion Before we look at the analyses for each hypothesis made in detail, let us look again at the overall research questions we aim to investigate: (1) At which stage of the acquisition process are learners able to produce a target like passive and what are the interlanguage variants and (2) what are the processing prerequisites for the comprehension of the passive voice. Table X provides a first overview of the overall results of our study. The left column indicates the stage of acquisition, the column perception indicates whether the syntactic parsing of the passive voice is possible at the respective stage and the column production shows, at which stage learners begin to produce interlanguage structures of the passive voice and are thus able to use non-linear mapping principles, which is a prerequisite for the production of the passive voice. The right column gives an example for each interlanguage structure. Looking at the results (table 4) it is quite evident that at the early stages of language acquisition, learners are neither able to parse the passive nor can they produce a syntactic structure. This provides a first answer to our overall research questions. Stage 6 is a prerequisite for the production of a target like passive structure. Interlanguage variants, which we refer to as pseudo- passives are produced by learners who have at least acquired stage 4. At this point, learners are also able to parse the passive. Stage 4 therefore seems to be the starting point for the production and for the syntactic parsing of passive structures in the acquisition process. In order to formulate a generalized hypothesis regarding the acquisition of the passive voice, we will now look the data in greater detail and discuss each hypothesis formulated in section 3 in greater detail.

Dagmar Keatinge & Jörg-U. Keßler

85

Table 4: Overview Results Stage 1 2 3 4

Perception / / / -

Production / / /

Example -------------

is + INF

The castle is break the dog (D04) The familiy is visit the castle (C08) Banana eat by monkey (E20) The car be bought by Peter (E25) The red fish is eating by the blue fish (E29) The mouse was eating from the cat. (D05) The postman was bite from the dog (D04) The shopping be carry by her hand (E29) Leftovers was clean (E30) Breakfast was be take out by my mother (E29)

INF + by Be+PAST+by Is+Inf+ by

5

-

6

-

was+Part.+from was+Past+from Be+INF+by was+INF was+be+INF+by target – irr. verb target-like

The kite was flyed by her father (E26) One tree is hitten by a car (E24) The dog was catched by the young man (E30) The book was written by Kornelia Funke (D09) The girl was helped by her mother (E30) One tree was hit by a car (E23)

Hypotheses 1 and 2: The passive voice develops gradually and stage 6 is the prerequisite for target-like production As Table 4 shows, the acquisition of the passive voice follows a developmental sequence. When there is a communicative need to produce a passive structure and learners are not developmentally ready to produce a target-like passive, they will produce an interlanguage structure instead. This can be observed in learners who have at least acquired stage 4. Learners who have acquired stages 1 to 3 are not able to alter from unmarked alignment- the most prominent semantic role is mapped onto subject function (cf. Pienemann et al. 2005: 239). Consider examples 11 and 12 below. Both informants have acquired stage 3 and in both examples, the syntactic structure does not represent the semantic concept of the passive voice. 8

 8

We claim that learners only violate the semantic concept of the passive in PRODUCTION at stages 1 – 3. This claim does not imply that the learners cannot PROCESS this semantic concept. Due to the constraint nature of their L2 processor, they are, however, not able to express this semantic concept by means

86

The Acquisition of the Passive Voice in L2 English 

(11) The book //write the woman. (D08; L1: German; Stage 3) (12)The car // repair the man. (D07; L1: German; Stage 3) The context created by the task design allowed no other option than the production of a passive structure; the informants have not differentiated between TOP and SUBJ. At stages 4 and 5, the semantic concept of the passive is represented in the syntactic structure. This syntactic structure is nonetheless not yet target like and thus gives evidence for the development of the pseudo- passive. Now, we can observe a number of different interlanguage variants learners use in order to express this semantic concept. Examples 13 to 18 (13) The noticebook is signed the man. (C11; L1: German; Stage 4) (14) The castle is visit the family (C08; L1 German; Stage 4) (15) Blue fish eat by red fish. (E20; L1: Mandarin; Stage 4) (16) The blue fish eated by the green fish (E27; L1 Arabic; Stage 5) (17) The mouse was eating from the cat. (D05; L1: German; Stage 5) (18) The postman was bite from the dog (D04; L1: German; Stage 5) There is, however some variation in the structure of the pseudo-passive. In examples (13) and (14), the informants construct the passive by using a form of “to-be” and either an infinitive or the participle of the main verb. The preposition “by” is omitted. In Examples (15) and (16), the informants construct the pseudo-passive by using a main verb plus the preposition “by”. In these examples, the form of to be is omitted. To sum up, the interlanguage structures in examples (13) – (16) are characterized by the omission of one obligatory component of the syntactic structure of the passive, making the VP less complex. First of all it is necessary at this point to consider the restricted communicative situation of the sentence completion task. All predicates only take two arguments. This task has enabled learners to simply exchange the two noun phrases of the sentence in order to place the patient of the sentence in initial position. This form of interlanguage development (strategies

 of PRODUCING the necessary syntactic form. In this study, we limit our analysis to the SYNTACTIC development of pseudo-passives. The nature of the L2 development of argument structure and semantic structure is the subject of further investigation (cf. Keatinge, in prep.).

Dagmar Keatinge & Jörg-U. Keßler

87

approach) has been documented by Clahsen (1984; reported in Pienemann 1998: 46). In this case, only processing at phrasal level would take place However, this falls in line with PT, as processing at phrasal level occurs at stage 3 at the earliest. Let us now consider examples (17) and (18), which include more syntactic elements of the passive voice. Examples (17) and (18) differ from the previous examples as far as they already include all obligatory structural parts of a target like passive. The structure itself is not target like, because the informant chooses an incorrect form of the participle of the main verb or of the copula to be. 9 However, the crucial difference between examples (11) and (12) (informants on stage 3) and (13) – (18) (informants on stage 4 and 5) is that the structures in (13) – (18) all require an alteration from unmarked alignment. In all examples, the less prominent semantic role is mapped onto SUBJ Function. Only speakers who have acquired stage 6 are able to produce target-like passives. However, not all speakers who have acquired stage 6 produce a target-like passive. Some learners produce structures like example (17) and use non- target like forms of the participle or to be. Our observations regarding the development of the passive voice are so far based on the notion of feature unification and the 1998 version of PT. In order to explain the observed interlanguage patterns from a point of view which also considers the a-structure, it is necessary to include predicates with more than two arguments in the data collection and to restrict the communicative nature of the task. However, we are able to propose a developmental sequence for the acquisition of the passive within the PT framework, based on these empirical findings:

 9

Please note that the use of a preposition other than „by“ is not included in this analysis, as this part of the passive is lexically marked

88

Stage Stage 1-3 Stage 4–5 (1)

Stage 4–5 (2) Stage 6

The Acquisition of the Passive Voice in L2 English 

IL- features Unmarked Alignment Mapping of agent onto SUBJ function, Active form of predicate (main verb only). Mapping of patient onto SUBJ function Use of finalization/initialization strategies Use of passive form of predicate, participle OR preposition Mapping of patient onto SUBJ function Feature unification at phrasal level Inclusion of passive form of predicate and preposition Mapping of patient onto SUBJ function

Violation of semantic concept and syntactic structure Semantic concept is represented in syntax Violation (omission) of target- like syntax and/or morphology Violation of target like morphology

Violation of target like morphology Target-like production

Hypothesis 3: Perception of the passive voice is constrained by processing procedures and Hypothesis 4: Perception does not require productive processing skills of this structure In table 4, we have already shown that stage 4 seems to be a prerequisite for the syntactic parsing of the passive. Table 5 now gives a detailed overview of the results of the comprehension task, displaying the stage of acquisition and the correct answers achieved in the comprehension task (in percent).

Dagmar Keatinge & Jörg-U. Keßler

89

Table 5 Informant Comprehension Stage

D01 0.7 4

D02 0.8 4

D03 0.8 4

D04 0.8 5

D05 0.9 5

Informant Comprehension Stage

D06 0.0 3

D07 0.0 3

D08 0.3 4

D09 0.8 6

D10 0.1 3

The results confirm both hypotheses. It is evident that the acquisition of comprehension skills is one or two stages further developed than the productive skills. With the exception of informant D08, all informants who have acquired stage 4 or 5 were able to comprehend passive structures. The perception of passive structures is only possible, once learners have acquired at least stage 4 - stage 6 is NOT a prerequisite for the PERCEPTION of the structure. Informants D06, D07 and D10, who have acquired stage 3, are not able to comprehend the passive voice. We have observed that learners are able to understand passive structures even though they are not able to produce even a pseudo-passive and not all learners who have acquired stage 4 are automatically able to comprehend passive structures. However, these observations cannot be fully explained at the moment. It is obvious that comprehension of a structure leads to the production of it. Further research needs to be carried out in order to explain the exact mechanisms of how a given structure is parsed by the L2- learner and how this knowledge is utilized in order to create a corresponding language structure within the PT framework (cf. Keatinge, in prep). Hypothesis 5: Priming leads to an increase of production of (pseudo) passive constructions As mentioned above, a priming effect was noticeable not only within the native speaker control group but also in all L2 corpora for the production. However, in the L2 corpora, only those subjects who had reached at least stage 4 were able to produce at least some pseudo passive constructions. Before a learner has reached stage 4 priming does not have any effect as those learners do not produce any passive constructions what so ever. Priming effects for the production of the passive voice worked independently from any kind of formal instruction on the passive voice. Those learners who had not reached stage 4 but had been taught to produce the

90

The Acquisition of the Passive Voice in L2 English 

passive voice did not produce any passive constructions at all; neither in the Fish Film nor in the other production tasks. However, those informants who had reached at least stage 4 did produce some (pseudo) passive constructions no matter whether they had been taught the passive voice in a classroom setting or not. Priming effects, on the other hand, do not last across data collection sessions: All subjects who did produce passive constructions right after having worked on the Fish Film did not necessarily produce a significantly high amount of passive constructions in other production tasks 7 – 10 days after the first data collection.

6. Conclusion Our study has been the first approach to investigate the acquisition of the passive voice with respect to the perception and production of this structure in L2 English. Our results have led to a number of implications for further research, which we will briefly describe below. It is evident that some interlanguage structures of the passive voice show a higher degree of simplification than others. It will be the subject of further research to investigate the extend to which highly simplified structures enhance stabilization. It is also necessary to investigate the perception and production of the passive voice with a greater range of tasks that include predicates with 3 or more predicates in order to make predictions regarding the Topic Hypothesis (Pienemann et al. 2005: 232 ff) and the acquisition of the passive voice. Regarding the perception of the passive voice, further research is required which includes speakers from more than one L1 background to investigate if typological differences are crucial in the perception (and production) of the passive voice. Priming effects were beneficial when priming took place in the beginning of the data collection session. However, long-term effects of priming were not visible during the delayed data collection. Further research needs to look at whether priming can be successfully applied in the EFL classroom in order to support the acquisition of the passive voice once the learner is developmentally ready to acquire the structure beyond a pseudo passive form. A follow-up classroom-based study ought to investigate this potential of SLA research for the fields of Applied Linguistics and teaching methodology.

Dagmar Keatinge & Jörg-U. Keßler

91

Bibliography Bresnan, J. (2001). Lexical- Functional Syntax. Malden, MA: Blackwell. Clahsen, H. (1984): “The acquisition of German word order: A test case for cognitive approaches to second language acquisition”. In: Andersen, R. (ed.): Second Languages. Mass.: Rowley. Fabri, R. (2008): “Lexical Functional Grammar.” In: Keßler, J.-U. (ed.): Processability Approaches to Second Language Development and Second Language Learning. Newcastle upon Tyne: Cambridge Scholars Publishing; pp. 31-65 Kawaguchi, S. (2005): “Argument structure and syntactic development.” In: Pienemann, M. (ed.): Crosslinguistic Aspects of Processability Theory, Amsterdam, John Benjamins, pp. 253 -298. Keatinge, D. (in prep.) Working title: The interface between comprehension and production in second language acquisition- a processability perspective. PhD Thesis, Paderborn University. Keßler, J.-U. (2008): “Communicative Tasks and Second Language Profiling: Linguistic and Pedagogical Implications.“ In: Eckerth, J. & S. Siepmann (eds): Research on task-based language learning and teaching. Theoretical, methodological and pedagogical perspectives. Frankfurt und New York: Peter Lang; pp. 291-310. Keßler, J.-U. (2007): “Assessing EFL-development online: A feasibility study of Rapid Profile.” In: Mansouri, F. (ed.): Second Language Acquisition Research. Theory-Construction and Testing. Newcastle upon Tyne: Cambridge Scholars Press; pp. 119-143. Keßler, J.-U. (2006): “Syntactic Priming and the Production of the Passive Voice in L2 English”. Paper read at Newcastle University. November 2006. Keßler, J.-U. & D. Keatinge (2008): “Profiling Oral Second Language Development”. In: Keßler, J.-U. (ed.): Processability Approaches to Second Language Development and Second Language Learning. Newcastle upon Tyne: Cambridge Scholars Publishing; pp. 167-197. Levelt... Mansouri, F. & L. Duffy (2005): Pallotti, G. (2007): “An operational definition of the emergence criterion.” Applied Linguistics, Vol . 28/3:361-382. Pienemann, M. (1998): Language Processing and Second Language Development. Processability Theory. Amsterdam: John Benjamins. Pienemann, M. (ed.) (2005): Crosslinguistic Aspects of Processabilty Theory. Amsterdam: John Benjamins.

92

The Acquisition of the Passive Voice in L2 English 

Pienemann, M. (2008): “A Brief Introduction to Processability Theory“. In: Keßler, J.-U. (ed.): Processability Approaches to Second Language Development and Second Language Learning. Newcastle upon Tyne: Cambridge Scholars Publishing; pp. 9-29. Tomlin, R. (1995): "Focal Attention, voice, and word order: An experimental, cross-linguistic study." In Word order in discourse edited by P. Downing & M. Noonan. Amsterdam: John Benjamins; pp. XXX Tomlin, R. (1997): "Mapping conceptual representations into linguistic representations: the role of attention in grammar." In Language and conceptualization edited by J. Nuyts & E. Pederson. Cambridge: Cambridge University Press; pp. 162-189.



ACQUIRING THE PASSIVE VOICE: ONLINE PRODUCTION OF THE ENGLISH PASSIVE CONSTRUCTION BY MANDARIN SPEAKERS 1 KENNY WANG Abstract The study utilises Processability Theory (PT; Pienemann 1998) and its extension, the Unmarked Alignment Hypothesis (Pienemann, Di Biase, & Kawaguchi 2005) and the Lexical Mapping Hypothesis (Pienemann et al. 2005; Kawaguchi & Di Biase 2005) as the language development framework to investigate the acquisition of the English passive construction by Mandarin learners of English as a second language (ESL): The two hypotheses predict that second language (L2) learners initially map the most prominent argument role (the agent) onto the grammatical subject to produce L2 in the active voice; as the learners’ L2 develops, they then learn to attribute prominence to the semantic patient to produce the passive construction. Six native Mandarin speakers of three different ESL proficiency levels (early intermediate, late intermediate, advanced) participated in a cross-sectional study where informants were shown a computer animation clip, the Fish Film (Tomlin 1995, 1997), which has been demonstrated to successfully elicit the English passive construction from native English speakers. The study found that the lower level learners consistently could not implement the instructional and contextual cues and persisted in using the active construction throughout the entire task; the late intermediate learner in the study was able to employ an alternative strategy which although was constrained by the processing capacity of the learner neverthelss preserved and conveyed the discoursepragmatic focus of the event. The advanced learners of the study performed comparable to the native speaker control. The results are interpreted as supporting the two PT-based hypotheses, which predict that early learners will cling on to the canonical mapping of Agent-ActionPatient to SVO and that they will not have the necessary processing

1

I wish to thank Bruno Di Biase, Satomi Kawaguchi and Michael Tyler for their helpful comments and advice on an earlier draft of this chapter.

94

Acquiring the Passive Voice: Online Production of the English Passive

resources to handle structures which deviate from the canonical semantic role to grammatical function alignment.

1.

Introduction

In order for L2 learners to communicate effectively in L2, they must not only comply with the L2 grammatical rules, they must also be able to accurately capture and convey the focus and emphasis of a proposition in L2. The study of cross-linguistic pragmatics concerns itself with the aspect of L2 pragmatic competence by learners (Wierzbicka 2003, 2006). Syntax has traditionally been studied extensively by L2 learners through conventional methods such as grammar translation. There is, however, an overlap of these two traditions which has not been given the attention it deserves in second language acquisition (SLA), namely the discoursepragmatic competence in L2. Discourse-pragmatics refers to “the knowledge that speakers bring to bear in deciding whether a syntactic structure can be fit to a developing discourse” (Cowan 1995:29). When native or proficient speakers of English wish to describe an event, there are a number of different ways for the speakers to express the same proposition. For example, in describing an event where a policeman arrested a shoplifter, native English speakers may choose from, among others, the following truth-conditionally equivalent alternatives: (1) (2)

The policeman arrested the shoplifter. The shoplifter was arrested by the policeman.

Stylistic considerations aside, both structures express essentially the same proposition. In deciding which voice to express the proposition, native English speakers would take into consideration the discourse-pragmatic needs of the particular communicative setting and choose the the most appropriate construction to convey the content while at the same time fulfill such discourse-pragmatic needs. Syntax is one of the means which speakers explore in attempting to “focus, direct and tantalise the listener’s attention” (Slobin 1979:64). Likewise, Labov (1990:45) also pointed out from a functionalist perspective that “on the whole grammar is not a tool of logical analysis; grammar is busy with emphasis, focus, down-shifting, and up-grading; it is a way of organizing information and taking alternative points of view.” However, only proficient speakers of a language may fully utilise syntax in such a functional capacity. The function of the passive voice in English

Kenny Wang

95

is to defocus the agent and attribute prominence to the patient (Givón 1979, 1984; Shibatani 1985). It has been observed in first language acquisition (FLA) that English speaking children acquire and master the passive voice relatively late, particularly the full passive construction including the agential by clause (Pienemann et al. 2005). In other words, English speaking children initally rely on the default word order and the active voice to comprehend and express propositions, giving little regard to discourse-pragmatics. The situation is less clear with adult L2 learners where learners by definition are discourse-pragmatically competent speakers of their L1 already, and particularly so where the passive construction exists and functions comparatively in their L1 as in the L2. The successful acquisition of the passive constructions in an L2 like English where the passive construction is a common and regular discourse feature is an important milestone in learners’ L2 development. This aspect of pragmatics, that is, the interactions between learners’ choice of L2 syntax and the intended pragmatic thrust of their proposition (cf. Green 2005) is an area in SLA that deserves further investigation. The objective of this study is to investigate the acquisition of the English passive construction by Mandarin L1 learners of English L2. This study will anwer the following research question: Question: When subjected to compelling pragmatic contexts that incline native speakers to promote the patient to the grammatical subject to produce passive constructions, do ESL learners likewise produce passive constructions as favoured by the pragmatic contexts? This paper will present prelimiary data showing that as hypothesised by the Processability Theory extension (Pienemann 1998; Pienemann et al. 2005), lower level L2 learners do not have the ability to produce English passive constructions spontaneously because they lack the necessary procesing resources to handle the passive construction. The rest of the paper is organised as follows: section 2 presents the development of the English passive construction in First Language Acquisition. Section 3 considers factors that would trigger the use of passives. Section 4 introduces the Processability Theory and its new hypotheses (Pienemann et al. 2005). Section 5 and 6 introduce an empirical study investigating the Mandarin L1 learners of ESL’s acquisition of the English passive construction in their L2 development and outline the research hypotheses and research design respectively.

96

Acquiring the Passive Voice: Online Production of the English Passive

Finally, the results of the study are given in section 7, followed by the conclusion and the implications of the study in section 8.

2.

Development of Passives in FLA

In terms of comprehension, it was found that English speaking children prior to age 4 consistently interpret strings of Noun-Verb-Noun sentences as Actor-Action-Undergoer (Bever 1970). Bever (1970) found that when children were asked to act out, for instance, the sentence “the cow is kissed by the horse”, 4-year-old children tended to interpret the first noun, ie. the cow, as the kisser, hence reversing the meaning of the sentence. Bever’s (1970) experiment was replicated by Maratsos (1974), whose finding confirmed Bever’s original observation. Moreover, de Villiers and de Villiers (1973) also modeled after Bever’s (1970) study but used the informants’ mean length of utterance (MLU) instead of the chronological age as the performance predictor. The results of their study, again, echoed those of Bever’s. In terms of production, when English speaking children do produce passives, they tend to produce truncated passives; full passives are rare in children’s speech production (Beilin & Sack 1975; Horgan 1976; Borer & Wexler 1987, 1992). For example, Harwood (1959) found no full passives in his corpus of approximately 12,700 utterances consisting of approximately 99,000 running words by children between the ages of 4;11 to 5;8. More recently, in a corpus study examining the development of passive participles in seven corpora of children’s spontaneous speech, Israel, Johnson and Brooks (2000) found that children first develop the ambiguous truncated passives, which may have a stative (adjectival) reading or a passive (verbal) reading (see Cook 1990). Furthermore, not only are full passives rare in children’s spontaneous speech, the scarcity continues into adulthood. Xiao, McEnery and Qian’s (2006) corpus-based analysis of both spoken and written British English by adults revealed that truncated be-passives make up 89.2% of the total written be-passives and 94.9% of the total spoken be-passives, while truncated get-passives make up 91.5% of the total written get-passives and 97.4% of the total spoken get-passives. The scarcity of full passives in adult speech consequently means that children are exposed to a very limited and infrequent linguistic input for this construction (Brown 1973). Marchman, Bates, Burkardt and Good (1991) conducted two experiments with children between 3 to 11 years old in an attempt to find out what children would do instead when they do not respond to probes to the patient of an event. In their study, the patient in an event was probed

Kenny Wang

97

by the experimenter in the form of “Tell me about the (undergoer)” or in other similar forms. It was found that when children did not produce passives in these circumstances, they employed the following two response types (Marchman et al. 1991:80-81): 1.

Actor Actives: Actives where the actor of the action is in subject position. e.g., ‘the bear is licking the tiger.’

2.

Non-Passives: Event or Non-event descriptions in which the probed character is topicalised. This category is comprised of four types of responses: (a) Non-Event Descriptions with the undergoer/location in subject position: e.g., ‘the tiger is just sitting there.’ (b) Event Descriptions with the undergoer/location in subject position: e.g., ‘the tiger let the bear lick him.’ (c) Two Clause constructions: e.g., ‘the tiger is just sitting there and the bear licks him.’ (d) Cleft constructions: e.g., ‘it was the tiger that the bear licked.’

In those instances where children did not choose to use passives, they would often utilise alternative constructions that fulfiled the same general discourse function as the passive. In other words, these alternative constructions “provide children with a ‘legal’ way to use easier and previously mastered constructions (i.e. the active voice) and still solve the discourse problem to topicalize the undergoer in the scene.” (Marchman et al. 1991:88-89). Interestingly, Qi (1996) also reported a similar observation of alternative strategies in Mandarin among some of the second graders in her study of monolingual Mandarin speaking children. In summary, the literature in FLA has shown that the English passive constructions are comprehended more poorly and acquired much later than their active counterpart by native English speaking children. It has been suggested that English speaking children do not have full control of the full passives until 7 years of age (Horgan 1978). When children do not produce passives, they tend to utilise alternative strategies that preserve the same discourse-pragmatic function as passives.

98

Acquiring the Passive Voice: Online Production of the English Passive

3.

Motivations for Passives

Speakers are always faced with different choices of syntactic forms, and they are free to make their choice in terms of how to express their intention. The factors that influence the speakers’ choice to favour the passive over the active construction are varied. One of such factors is the establishment of discourse topicality. A general definition of the notion of discourse topic is “the proposition (or a set of propositions) about which the speaker is either providing or requesting new information” (OchsKeenan & Schieffelin 1983:68). Patient-focused questions such as “What is being done to (the patient)?” and “What is happening to (the patient)?” have been found to be effective in eliciting responses in the passive voice from primary and secondary school aged children (Carroll 1958; Turner & Rommetveit 1967). The patient may also be framed as the discourse topic through non-linguistic means such as physically pointing something out with one’s finger (Ochs-Keenan & Schieffelin 1983). The pragmatic presupposition also affects the syntactic structures employed. Bock (1977:723) hypothesised that “alternative surface structures are used differentially in order to array the information in sentences with given information preceding new information.” Bock found that when the semantic patient was the given information or the pragmatically presupposed information, it was more likely to precede the semantic agent, resulting in the interlocuter producing the passive construction. Moreover, the conceptual accessibility of the nouns has been claimed to exert influence on the syntactic array of the nouns in a sentence (Bock & Warren 1985). Bock and Warren found that the conceptually more accessible entity, which in their study was meansured by the imaginability of the noun, was more likely to appear early in a sentence. For example, “the doctor” in “The doctor administered a shock” was more imaginable than “the shock”, it was therefore more conceptually accessibile and more likely to appear first in a sentence, resulting in the active construction. If the conceptual accessbility of the semantic patient was higher than the agent, the passive construction was more likely to occur. The relative thematicity has also been hypothesised as the predictor of grammatical voice in narratives (Tomlin 1983). It was found that the more thematic an entity of an event is, the more likley it would occupy the subject position in a sentence. Tomlin drew support for his hypothesis based on the analyses of corpus data consisting of ice hockey game commentaries. It was concluded that the grammatical subject in

Kenny Wang

99

English is used primarily to encode thematic information, and secondarily to encode agency. Finally, while the core notions of topicality and thematicity have been offered as motivations for the passive construction, Tomlin in his later works (Tomlin 1995, 1997) pointed out that such notions are vague and subjective. In order to pin the vague notions of topicality and thematicity down to something more concrete and empirically manipulable, Tomlin contended replacing the pragmatic notions of “thematicity” and “topicality” with the cognitive counterpart of “visual attention”. Using a computer animation called Fish Film, Tomlin found that the entity that entered the speaker’s visual attention at the moment of speech generation was realised as the sentential subject. In other words, if the semantic patient was the visually attended entity as speakers produced the sentence spontaneously, that entity would occupy the grammatical subject postion, resulting in a passive sentence. Tomlin was able to achieve over 90% accuracy rate with his native English speaking informants, which was far greater than previous experiments involving the manipulations of less tangible variables such as theme, topicality, conceptual accessibility, saliency, animacy. In summary, psycholinguistic studies on the triggers for the passive alternation in English have indicated that the privilege of sentential primacy may be reserved for the most prominent entity, where prominence may be attributed cognitively due to various motivations, including the relative topicality, thematicity, presuppositionality, saliency and animacy of the entities, as well as the speaker’s perspective-taking, empathy, visual attention, etc.

4.

Processability Theory Extension and Its Hypotheses

Processability Theory (Pienemann 1998) explains L2 morphosyntactic development in terms of the architecture of the human language processor and other human psychological constraints such as how lexical access occurs and how working memory constraints L2 production. It offers a universal framework that can be used to predict the developmental paths of L2s of any typology. PT is based on Levelt’s (1989) speech production model, particularly the conceptualiser and the formulator components of the model. It strived to provide an explanation to account for the L2 developmental patterns based on a psychologically plausible speech processing model. The main thrust of PT is that “at any state of development, the learner can produce and comprehend only those L2 linguistic forms that the current state of the language processor can

100 Acquiring the Passive Voice: Online Production of the English Passive

handle.” (Pienemann 2007:137). PT hypothesises that in second language acquisition, learners acquire L2 morphological and syntactical constructions in a predictable and verifiable order. PT hypothesises that the L2 learner’s language processing procedures form a hierarchical set of strata in which the acquisition of the lower level procedure is the prerequisite for acquiring the next procedure up the strata (Pienemann 1998). PT is a universal theory of SLA that has been tested against several typologically different L2s such as Swedish (Pienemann & Håkansson 1999), Italian (Di Biase 2002; Di Biase & Kawaguchi 2002; Kawaguchi & Di Biase 2005), Arabic (Mansouri 1997, 2002, 2005), Japanese (Kawaguchi 2005), and Chinese (Gao 2005, Zhang 2004). PT adopts Lexical-Functional Grammar (LFG; Bresnan 2001) as its grammatical formalism (Pienemann 1998). LFG is a lexical featureunification based theory of grammar. It offers a lexically driven grammatical process of explaining speakers’ linguistic knowledge that is psychologically compatible to the architecture of the human language processor. It is for this reason that PT chooses LFG as its grammatical formalism. The underlying assumption of PT is the psychological constraints on the L2 learners’ ability to process language output. PT hypothesises a universal hierarchy of L2 development. The implicational developmental hierarchy consists of the following procedures in the respective order: lemma acces > category procedure > intra-phrasal procedure > interphrasal procedure > inter-clausal procedure. As L2 learners become more proficient with a language processing procedure, tasks involving it become less psychologically demanding for learners and more automatised, thus freeing up the learners’ processing concentration allowing them to possibly move on to acquiring the next procedure up the hierarchy. The original PT proposed in 1998 was to explain learners’ L2 morphosyntactic development. Three PT-based hypotheses have recently been put forth to extend PT beyond its scope of developmental stages of morphological-syntactical constructions. The PT extension seeks to account for learners’ structural choices as reflections of their L2 development of the syntactactic-pragmatic interface (Pienemann et al. 2005). We shall consider only two of these hypotheses in this study, namely the Unmarked Alignment Hypothesis (UAH) and the Lexical Mapping Hypothesis (LMH), as only these two are relevant to the activepassive alternation.

Kenny Wang

101

Many researchers (e.g. Jackendoff 1972; Foley & Van Valin 1984; Givón 1984; Bresnan 2001) have suggested a universal hierarchy of semantic roles as in (3): (3) Thematic hierarchy (Bresnan 2001:307) Agent > Beneficiary > Experiencer/Goal > Instrument > Patient/Theme > Locative Moreover, grammatical functions also have a hierarchical relationship according to their prominence, as in (4): All core functions are more prominent than noncore functions: (4)

Relational hierarchy (Keenan & Comrie 1977; Bresnan 2001:96) core noncore SUBJ > OBJ > OBJT > OBLT > COMPL > ADJUNCT (OBJT refers to secondary object)

UAH explains how learners by default integrate the top element from each of these two hierarchies as the point of departure of their L2 syntax. The Unmarked Aligned Hypothesis states that: “In second language acquisition learners will initially organise syntax by mapping the most prominent semantic role available onto the subject (i.e. the most prominent grammatical role): The structural expression of the subject, in turn, will occupy the most prominent linear position in cstructure, namely the initial position.” (Pienemann et al. 2005).

Essentailly, UAH claims that learners’ structural choices are initially constrained to one that involves unmarked or canonical mapping due to the limited amount of processing resources available to the learners. The unmarked alignment is illustrated in (5) for the sentence “Peter pats a dog” (from Kawaguchi 2007).

102 Acquiring the Passive Voice: Online Production of the English Passive

(5)

pat < x, y > agent Patient

a-structure (argument roles)

subject

Object

f-structure (grammatical functions)

Peter

a dog

c-structure (word order)

The one-to-one linear correspondence of semantic roles, grammatical functions and c-structure positions produces an optimal alighment in line with the Optimal Syntax (OT-LFG) notion of “harmonic alignment” (Prince & Smolensky 2004). On the basis of this optimal alignment, LMH narrows the scope down to the mappings of semantic roles onto grammatical functions. The Lexical Mapping Hypothesis states that: “L2 learners initially map the most prominent onto SUBJ and gradually learn how to attribute prominence to a particular thematic role, e.g., promoting the patient (rather than the agent) role to SUBJ, first in single clauses such as in Passive constructions and later in complex predicates such as Causative constructions.” (Kawaguchi & Di Biase 2005)

The canonical mapping of argument roles to grammatical functions requires the least amount of processing resources (Pinker 1984). Therefore, when learners have accumulated a number of L2 lexical items and are ready to string words together to produce multi-word utterances, they will initially cling on to the canonical mapping of the target language. In the case of English, UAH predicts that English L2 learners will, regardless of their L1, produce SVO, where the subject position is occupied by the semantic agent. LMH predicts that learners will only be able to differentiate the notions of agent and subject and manipulate the mapping of non-agent roles to subject much later in their L2 development, since doing so requires learners to detour from the canonical and the least computationally costly mapping.

5.

Hypotheses

Based on the theoretical background and theoretical framework presented thus far, to answer our research question, the following hypotheses are

Kenny Wang

103

proposed concerning the acquisition of the English passive construction by Chinese learners: Hypothesis A: Lower level Mandarin L1 learners of English L2 will encode the patient as the subject when the patient is indicatd by the pragmatic context as the more prominent entity. Tomlin’s (1995, 1997) cross-linguistic study showed that native English speakers consistently realised the patient as the subject when it was contextually more prominent than the agent. The native Mandarin speakers of Taiwanese origin in Tomlin’s study also consistently narrated the patient-cued trails using the passive b΁ construction. Hypothesis A therefore predicts that early Mandarin L1 learners will mirror the Mandarin passive voice and use the simple English passive construction similar to native English speakers. Hypothesis B: Lower level Mandarin L1 learners of English L2 do not have the necessary processing resournces to produce the passive construction. Therefore they will encode the agent as the subject despite the contextual and instructional cues presented. Advanced learners on the other hand will be able to produce English passives comparable to native English speakers. The PT extension (Pienemann et al. 2005) predicts that since the canonical mapping requires the least amount of processing resources, lower level learners will consistently produce sentences by employing the canonical mapping, even when the pragmatic context favours a non-canonical mapping.

6.

Research Design Informants

In order to test Hypotheses A and B, a cross-sectional experiment with seven informants and a control was conducted using a quantitative approach, which is commonly used in second language acquisition studies (Doughty & Long 2003). Seven native Mandarin speakers were successfully recruited as informants in the study: one at the beginner level, three at the early intermediate level, one at the late intermediate level and two at the advanced level. The informants were choosen based on their highest level

104 Acquiring the Passive Voice: Online Production of the English Passive

of formal adult ESL instruction received in Australia. The three early intermediate and one late intermediate learners were students at the Adult Migrant English Services (AMES), an adult ESL education provider in Australia. The advanced learners were students enroled in the Bachelor of Arts program at the University of Western Sydney. In addition, a native English speaker was engaged in the study both as control and to ensure the appropriateness of the task in eliciting the target structure. All of the informants participated as volunteers. The ESL abilities of the informants were determined by external education providers. The beginner level learner had just arrived in Australia two months earlier and did not possess the necessary L2 lexical means to cope with the task given and was therefore excluded from the study. In order to preserve the privacy of the informants, a fictitious name is assigned to each of the informants for identification purposes. The following table gives an overview of the informants in our cross-sectional study: ESL Level

Pseudonym

Country of Origin

No. of Years in Australia

Early Intermediate

Suan

China

2

Early Intermediate

Ming

Taiwan

2

Early Intermediate

Mei

Vietnam

25

Late Intermediate

Cindy

China

8

Advanced

Zu

China

2

Advanced

Chen

China

2

Native Control

Linda

Australia

N/A

Table 1: An overview of the informants.

Other Information Has been studying English for 1.5 years Has been studying English for 2 years Has been studying English for 2 years Completed 510 hours of ESL at AMES BA student at UWS BA student at UWS

Kenny Wang

105

Data Collection An output elicitation task was used in the session with the intention of eliciting naturalistic spoken production of the English passive construction from the informants. Since my aim was to subject the learners to relatively coercive contextual conditions that would cause learners to produce passives, I have to this end employed Tomlin’s (1995) Fish Film 2 to create such conditions. The Fish Film is a 4 minute 40 second cartoon animation. The animation clip visually manipulated informants’ visual attention to create a dynamic context, which has been demonstrated to reliably incline native English speakers to generate patient-fronting SVO (passive voice) utterances. It was subsequently used to investigate the relationship between structural outcomes and attention direction in native speakers of typologically different languages, including Chinese (Mandarin), Indonesian and Russian (Tomlin 1995, 1997). The clip contained 32 trails. Each trail of the clip depicted two fish of different colours. The two fish would swim towards each other from the two sides of the screen. As the two fish of each trail appeared on the screen, either of the two would be visually cued by a flashing arrow appearing above the selected fish. As the two fish reached each other at the centre of the screen, one would swallow the other. The fish that was cued with the arrow could have been either the agent fish or the patient fish, the informant only found out at the moment the one swallowed the other (see Figure 1). The film ran at 12 frames per second. The direction of the agent fish (entering from the left or the right) was counterbalanced; the colours of the two fish in each episode were randomly selected. In half the trails the agent was cued, in the other half the patient.

2

Copyrighted © 2002-04 Russell S. Tomlin. Expressed consent to use the clip for research purposes is given on Tomlin’s website, which states: “The Fish Film is copyrighted: © 2002-04 Russell S. Tomlin, though I am pleased for anyone to use the film in support of basic research in linguistics, psychology, and related disciplines.”

106 Acquiring the Passive Voice: Online Production of the English Passive

Figure 1: The Fish Film. (Tomlin 1995, 1997) The informants were instructed to keep up with the unfolding of the episodes and were asked to keep their eyes on the arrow-flashed

Kenny Wang

107

referent, hence ensuring that their visual attention would be allocated to the target fish at the crucial moment (swallowing): The informants were to give an online narration of the events as they unfolded. The entire session was tape recorded and transcribed for analysis.

7.

Results

Of the 32 trails in the Fish Film, 16 were patient-cuing trails. This means that each participant was subjected to 16 passive construction friendly contexts. Linda the control in this study produced 15 passives out of the 16 passive-construction friendly contexts, which corroborated Tomlin’s claim that the Fish Film task was a very coersive context to producing the English passive construction. The overall performance and the number of each type of narrations by the cross-sectional informants is organised in Table 2 below:

Name

English Level

Passive

Frequency AgentActive

PatientActive

Linda

Native Speaker

15

1

0

0

16

0

0

16

0

0

15

0

0

11

5

Susan Ming Mei Cindy

Early Intermediate Early Intermediate Early Intermediate Late Intermediate

Zu

Advanced

12

4

0

Chen

Advanced

15

1

0

Table 2: Frequency table of the types of narrations by informants. The empirical data showed that all three early intermediate level learners consistently encoded the agent as the subject regardless of which entity was cued, producing the active construction in the canonical SVO word order and mapping in all trails.

108 Acquiring the Passive Voice: Online Production of the English Passive

In the case of our late intermediate learner Cindy, of the 16 trails where the patient was visually cued, Cindy quite consistently, in 11 out of 16 instances, encoded the agent as the subject despite the visual cuing of the patient. For the remaining 5 instances, she adopted an alterantive strategy where she narrated the event from the patient’s perspective, in essence converting the patient to an agent of the same event followed by an accompanying active voice predicate (e.g. the dog chased the cat/the cat ran away from the dog). Cindy’s performance suggested that while she still relied quite heavily on the canonical mapping, she had slightly more processing resources at her disposal. The two advanced learners performance was much more similar to the native control, suggesting that they had the necessary processing resources to select the most pragmatically felicitous syntactic structure in their on-line construction of descriptive utterances. Examples of Responses Let us now turn to some examples of each participant’s production for the passive-cuing trails. Starting with our control, Linda produced 15 passives out of 16 passive friendly trails, achieving a 94% hit rate. This is very close to Tomlin’s 100% individuated prediction. A typical response by Linda’s to a patient-cuing trail is given below: (6) Trail No. 12

Visually Cued Patient Green fish

Linda’s Narration now the green one got swallowed by the pink one. it had the arrow on it. the green one

The early intermediate learner, Mei, is from Vietnam, but Mandarin is her first language. Although she had been in Australia for 25 years, she had been working in a non-English speaking workplace for most of that time, during which her contact with the English speaking world and her use of English had been extremely limited. She had just been retired for two years when the data collection session took place and had just taken up full-time English study 1.5 years prior to the time of data collection. Mei, produced 0 out of 16 patient-cuing trails. She used agentactives despite the cuing arrows in 15 of the 16 trails and missed one trail due to long hesitance:

Kenny Wang

109

(7) Trail No.

Visually Cued Patient

Mei’s Narration

11

Blue fish

the green. eat the blue

17

Red fish

the grey eat the red

Ming had been studying English for 2 years at AMES. She produced 0 out of the 16 patient-cuing trails. Ming consistently used the default canonical order, assigning the agent the subject despite the cuing arrows: (8) Trail No.

Visually Cued Patient

11

Blue fish

25

Black fish

Ming’s Narration and the green and the blue. the. green eat the blue the black and the white.. and the white eat the black

Suan had been studying English at AMES for 1.5 years. She produced 0 out of the 16 patient-cued trails. Like Ming, Suan also consistently used the default canonical order, assigning the agent the subject despite the cuing arrows: (9) Trail No.

Visually Cued Patient

Suan’s Narration

6

Pink fish

the blue fish eat pink fish

17

Red fish

the grey fish eat red fish

The performance of Cindy the late intermediate learner in the study is of particular interest. Cindy had been in Australia for 8 years. She had completed 510 hours of adult ESL course at AMES and had reached AMES’ late intermediate level. Cindy produced 0 passive out of the 16 patient-cued trails. Of the 16 patient-cued trails, Cindy produced 11 agent actives despite the presence of the arrows cuing the patient fish. All of the 11 responses are listed in (10) below:

110 Acquiring the Passive Voice: Online Production of the English Passive

(10) Trail No.

Visually Cued Patient

Cindy’s Narration

2

Red fish

Blue fish eat red fish

3

Blue fish

Green fish eat blue. fish

8

Black fish

Red eat black

10

Red fish

Blue eat red

11

Blue fish

Green fish eat blue. fish

12

Green fish

Pink eat blue. green

16

Pink fish

Red eat pink

20

Blue fish

Pink eat blue

21

Pink fish

White white eat pink

22

White fish

Black eat white

27

Black fish

White eat black

However, for the remaining 5 patient-cued trails, Cindy employed a compensatory strategy which I shall call the Patient-Active Strategy. The Patient-Active Strategy essentially means that the learner maps the patient of an event to the subject, therefore organises the sentence in a syntactically canonical order. The learner at the same time selects a verb that takes a perspective that corresponds to that of the patient, thus effectively converting the patient of the event into the agent of the same eventuality but as seen from an alternative perspective. This strategy then is akin to the principles of frame semantics purported by Fillmore and Petruck (Fillmore 1985, 2006; Petruck 1996). The following are the 5 instances where Cindy employed the Patient-Active Strategy: (11) Trail No. 4 6

Visually Cued Patient Green fish Pink fish

Cindy’s Narration The green fish in pink fish The pink fish eat eat black (revised upon trail replay): pink fish come in black fish

Kenny Wang

111

17

Red fish

Red. goes white

18

White fish

The white goes go go to blue

29

Grey fish

Grey… go to white

For the advanced learners, Zu was a first year student in the Bachelor of Arts program at the University of Western Sydney. He produced 12 passives out of 16 patient-cued trails, achieving a 75% hit rate. Zu’s use of the English passive construction is much more native-like than early learners’ active voice. Two examples of his sentences in the passive voice are given below: (12) Trail No.

Visually Cued Patient

Zu’s Narration

11

Blue fish

a blue fish with arrows were eaten by green fish from right

17

Red fish

a red fish with arrows come from left right ah eaten by a grey fish

Chen is also a first year student in the Bachelor of Arts program at the University of Western Sydney. He achieved a 94% hit rate or 15 out of 16 passives. With the exception of one trail, Chen otherwise quite consistently mapped the cued patient onto the subject: (13) Trail No.

Visually Cued Patient

11

Blue fish

12

Green fish

Chen’s Narration the blue one is swallowed by the green one and the green one swallowed by the pink one

Overall, the performance of the six informants and the native control is shown in Fig. 1:

112 Acquiring the Passive Voice: Online Production of the English Passive

Ming

20

Susan

10

15

Mei Cindy

Li n da

Ch en

Zu

Me i

15

0 dy

0 Cin

0 n

0

Mi ng

0

12

Su sa

Number of Passives

Number of Passives Produced by Each Informant

Zu Chen Linda

Figure 1: Number of passives produced by each informant. The results of the study overwhelmingly rejects Hypothesis A and supports Hypothesis B.

8.

Conclusion and Implications of the Study

Our empirical data has demonstrated that, as hypothesised by UAH and LMH, lower level Mandarin L1 learners of English L2 do not possess the necessary procedural skills to assign prominence to the semantic patient role by means of promoting it to the subject function. The early intermediate learners in the study fared as hypothesised by LMH, they consistently failed to incorporate any prominence attribution in their L2 syntax; only the two advanced learners of English L2 were able to cope with the patient-cuing pragmatic contexts by producing sentences using the pragmatically appropriate yet non-canonical syntactic mapping. I interpret the absence of prominence attribution through syntax on the part of the lower level learners as due to their lack of the necessary processing resources to handle the non-canonical mapping of the patient onto the subject. Furthermore, the late intermediate learner in the study, Cindy, was fairly reliant on the canonical order. However, when visually inclined by the context to produce the passive construction, in the absence of the necessary processing skills to produce the passive construction, she was nonetheless able to map the relatively more prominent patient to the subject on five occasions. Cindy’s compensatory Patient-Active Strategy showed that while her limited processing resources still confined her to reserving the subject for the agent, she was nevertheless able to assign the

Kenny Wang

113

more prominent participant to the sentence initial position at times, thereby preserving the same pragmatic focus within the boundaries of her processability as warranted by the context. In conclusion, the results of the study indicate that learners’ sensitivity to discourse-pragmatic contexts does interface with their L2 syntax; however such interface is developmentally moderated; that is, learners’ ability to manipulate the semantic role to grammatical function alignment is subject to their having the necessary processing resources for the specific L2 constructions in question. The empirical evidence of this study has a number of implications for language teaching, syllabus design and language assessment. The Teachability Hypothesis (Pienemann 1989) suggests that formal language instruction should work with and facilitate the natural acquisition sequence. This study highlights the importance for TESOL educators to adopt an ESL learner processability oriented paedegogical approach and syllabus design to teaching the grammatical voices. In terms of language assessment, this study may contribute to the inclusion of non-canonical syntax as a new dimension in the future development of linguistic profiling using Rapid Profile (Keßler & Keatinge 2008), thus further enhancing the robustness of this language profiling and online assessment procedure.

References Beilin, H., & Sack, H. G. (1975): "Experiments on the passive." In Studies in the cognitive basis of language development edited by H. Beilin. New York: Academic Press. Bever, T. G. (1970): "The cognitive basis for linguistic structures." In Cognition and the development of language edited by J. Hayes. New York: Wiley. Bock, J. K. (1977): "The effect of a pragmatic presupposition on syntactic structure in question answering". Journal of Verbal Learning and Verbal Behavior, Vol. 16/6 (1977): 723-734. Bock, J. K., & Warren, R. K. (1985): "Conceptual accessibility and syntactic structure in sentence formulation." Cognition, Vol. 21/1 (1985): 47-67. Borer, H., & Wexler, K. (1987): "The maturation of syntax." In Parameter setting edited by T. Roeper & E. Williams. Dordrecht: Reidel.

114 Acquiring the Passive Voice: Online Production of the English Passive

Borer, H., & Wexler, K. (1992): "Bi-unique relations and the maturation of grammatical principles." Natural Language & Linguistic Theory, Vol. 10 (1992): 147-189. Bresnan, J. (2001): Lexical-functional syntax. Malden: Blackwell. Brown, R. (1973): A first language: The early stages. London: Allen and Unwin. Carroll, J. B. (1958): "Process and content in psycholinguistics." In Current trends in the description and analysis of behavior edited by R. Glaser. The Hague: Mouton: 175-200 Cook, W. A. (1990): "Passive semantics: Ambiguity of the short passive." Georgetown Journal of Languages and Linguistics, Vol. 1/1 (1990): 2530. Cowan, R. (1995): "What are discourse principles made of?" In Word order in discourse edited by P. Downing & M. Noonan. Amsterdam: John Benjamins: 29-49 de Villiers, P. A., & de Villiers, J. G. (1973): "Development of the use of word order in comprehension." Journal of Psycholinguistic Research, Vol. 2/4 (1973): 331-341. Di Biase, B. (2002): "Focusing strategies in second language development: A classroom-based study of Italian L2 in primary school." In Developing a second language: Acquisition, processing and pedagogy issues in Arabic, Chinese, English, Italian, Japanese and Swedish edited by B. Di Biase. Melbourne: Language Australia. Di Biase, B., & Kawaguchi, S. (2002): "Exploring the typological plausibility of processability theory: Language development in Italian second language and Japanese second language." Second Language Research, Vol. 18/3 (2002): 274-302. Doughty, C., & Long, M. H. (2003): The handbook of second language acquisition. Oxford: Blackwell. Foley, W. A., & Van Valin, R. D. (1984): Functional syntax and universal grammar. Cambridge: Cambridge University Press. Fillmore, C. J. (1985): "Frames and the semantics of understanding." Quaderni di Semantica, Vol. 6/2 (1985): 222-254. Fillmore, C. J. (2006): "Frame semantics." In Cognitive linguistics: Basic readings edited by D. Geeraerts. Berlin: Mouton de Gruyter. Gao, X. (2005): Noun phrase morphemes and topic development in L2 Mandarin Chinese: A processability perspective. Unpublished PhD thesis, Victoria University of Wellington, Wellington. Givón, T. (1979): On understanding grammar. New York: Academic Press.

Kenny Wang

115

Givón, T. (1984): Syntax: A functional-typological introduction. Amsterdam: John Benjamins. Green, G. M. (2005). "Some interactions of pragmatics and grammar." In The handbook of pragmatics edited by L. R. Horn & G. Ward. Oxford: Blackwell Publishing. Harwood, F. W. (1959): "Quantitative study of the speech of Australian children." Language and Speech, Vol. 2 (1959): 236-270. Horgan, D. (1978): "The development of the full passive." Journal of Child Language, Vol. 5 (1978): 65-80. Israel, M., Johnson, C., & Brooks, P. J. (2000): "From states to events: The acquisition of English passive participles." Cognitive Linguistics, Vol. 11/2 (2000): 103-129. Jackendoff, R. (1972): Semantic interpretation in generative grammar. Cambridge: The MIT Press. Kawaguchi, S. (2005): Acquisition of Japanese as a second language: A Processability Perspective. Unpublished PhD dissertation, University of Western Sydney. Kawaguchi, S. (2007): "Lexical mapping in Processability Theory: A case study in Japanese." In Second language acquisition research: Theoryconstruction and testing edited by F. Mansouri. Newcastle upon Tyne: Cambridge Scholars Publishing. Kawaguchi, S., & Di Biase, B. (2005): Second language development at the syntax-pragmatics interface. Paper presented at the the 7th Annual International Conference of the Japanese Society of Language Sciences, Sophia University, Tokyo. Keßler, J.-U., & Keatinge, D. (2008). "Profiling oral second language development". In Processability approaches to second language development and second language learning edited by J.-U. Keßler. Newcastle upon Tyne: Cambridge Scholars Publishing. Keenan, E. O., & Comrie, B. (1977): "Noun phrase accessibility and Universal Grammar." Linguistic Inquiry, Vol. 8/1 (1977): 63-99. Labov, W. (1990): "On the adequacy of natural languages: I. The developoment of tense." In Pidgin and creole tense-mood-aspect systems edited by J. V. Singer. Amsterdam: John Benjamins. Levelt, W. J. M. (1989): Speaking: From intention to articulation. Cambridge, Mass.: The MIT Press. Mansouri, F. (1997): "From emergence to acquisition: Developmental issues in Arabic interlanguage morphology." The Australian Review of Applied Linguistics, Vol. 20/1 (1997): 83-104. Mansouri, F. (2002): "Exploring the interface between syntax and morphology in second language development." In Developing a second

116 Acquiring the Passive Voice: Online Production of the English Passive

language: Acquisition, processing and pedagogy of Arabic, Chinese, English, Italian, Japanese, Swedish edited by B. Di Biase. Melbourne: Language Australia. Mansouri, F. (2005): "Agreement morphology in Arabic as a second language: Typological features and their processing implications." In Cross-linguistic aspects of Processability Theory edited by M. Pienemann. Amsterdam: John Benjamins. Maratsos, M. (1974): "Children who get worse at understanding the passive: A replication of Bever." Journal of Psycholinguistic Research, Vol. 3 (1974): 65-74. Marchman, V. A., Bates, E., Burkardt, A., & Good, A. B. (1991): "Functional constraints of the acquisition of the passive: toward a model of the competence to perform." First Language, Vol. 11 (1991): 65-92. Ochs-Keenan, E., & Schieffelin, B. (1983): "Topic as a discourse notion: A study of topic in the conversations of children and adults." In Acquiring conversational competence edited by E. Ochs-Keenan & B. Schieffelin. London: Routledge & Kegan Paul. Petruck, M. R. L. (1996): "Frame semantics." In Handbook of Pragmatics edited by J. Verschueren, J.-O. Östman, J. Blommaert & C. Bulcaen: Amsterdam: John Benjamins. Pienemann, M. (1989). "Is language teachable? Psycholinguistics experiments and hypotheses." Applied linguisticsm, Vol. 10/1 (1989): 52-70. Pienemann, M. (1998): Language processing and language development: Processability Theory. Amsterdam: John Benjamins. Pienemann, M. (2007): "Processability Theory." In Theories in second language acquisition: An introduction edited by B. VanPatten & J. Williams. London: Lawrence Erlbaum. Pienemann, M., Di Biase, B., & Kawaguchi, S. (2005): "Extending processability theory." In Cross-linguistic aspects of processability theory edited by M. Pienemann. Amsterdam: John Benjamins. Pienemann, M., & Håkansson, G. (1999): "A unified approach towards the development of Swedish as L2: A processability account." Studies in Second Language Acquisition, Vol. 21/3 (1999): 383-420. Pinker, S. (1984): Language learnability and language development. Cambridge: Harvard University Press. Prince, A. S., & Smolensky, P. (2004): Optimality Theory: Constraint interaction in generative grammar. Oxford: Blackwell Publishing. Qi, Y. (1996): "A survey of school children's use of passive sentences." Applied Linguistics, Vol. 19/3 (1996): 41-46. (In Chinese)

Kenny Wang

117

Shibatani, M. (1985): "Passive and related constructions: A prototype analysis." Language, Vol. 61/4 (1985): 821-848. Slobin, D. I. (1979): Psycholinguistics (2nd ed.): Glenview: Scott Foresman. Tomlin, R. (1983): "On the interaction of syntactic subject, thematic information and agent in English." Journal of Pragmatics, Vol. 7/4 (1983): 411-432. Tomlin, R. (1995): "Focal Attention, voice, and word order: An experimental, cross-linguistic study." In Word order in discourse edited by P. Downing & M. Noonan. Amsterdam: John Benjamins. Tomlin, R. (1997): "Mapping conceptual representations into linguistic representations: the role of attention in grammar." In Language and conceptualization edited by J. Nuyts & E. Pederson. Cambridge: Cambridge University Press: 162-189. Turner, E. A., & Rommetveit, R. (1967): "Experimental manipulation of the production of active and passive voice in children." Language and Speech, Vol. 10 (1967): 169-180. Wierzbicka, A. (2003): Cross-cultural pragmatics: The semantics of human interaction (Expanded second ed.): Berlin: Mouton de Gruyter. Wierzbicka, A. (2006): "Intercultural pragmatics and communication." In The Encyclopedia of Language and Linguistics edited by K. Brown. London: Elsevier. Xiao, R., McEnery, T., & Qian, Y. (2006): "Passive constructions in English and Chinese: A corpus-based contrastive study." Languages in Contrast, Vol. 6/1 (2006): 109-149. Zhang, Y. (2004): "Processing constraints, categorial analysis, and the second language acquisition of the Chinese adjective suffix -de(ADJ):" Language Learning, Vol. 54/3 (2004): 437-468.

PART II SECOND LANGUAGE GRAMMARS ACROSS LANGUAGES





MORPHOLOGICAL DEVELOPMENT IN WRITTEN L2 FRENCH: A PROCESSABILITY PERSPECTIVE MALIN ÅGREN 1. Introduction During the last decade, Processability Theory (PT) (Pienemann 1998) has been successfully applied to the second language (L2) development of various morpho-syntactic phenomena in many typologically different languages, e.g. Arabic, Chinese, English, Italian, Japanese, Swedish and other Scandinavian languages (Mansouri 2005; Zhang 2005; Pienemann 1998; Di Biase and Kawaguchi 2002; Pienemann and Håkansson 1999; Glahn et al. 2001). A few attempts have also been made to apply the PT framework to the L2 acquisition of French. Bartning (1999, 2000) as well as Dewaele and Véronique (2000, 2001) used PT in the analysis of gender assignment and agreement in groups of advanced L2 learners of French. In general, the results have not been interpreted in favour of the PT hierarchy. Pienemann (2005: 62) discusses these findings and calls attention to the fact that certain methodological aspects in Dewaele and Véronique (2001) are problematic from a PT perspective. First of all, a lexical feature like gender is not a good candidate for a PT analysis of morphological development. This is due to the fact that gender assignment has to be established separately for every item in the learner’s lexicon before one can tell how well this information is exchanged between constituents at different processing levels. Secondly, previous studies on L2 French seem to focus on learners at advanced levels whereas studies in the PT framework emphasize the emergence of grammatical structures at an initial level of L2. Finally, PT does not make any predictions on levels of accuracy, which is the basic acquisition criterion used in the above mentioned studies of L2 French. Pienemann (2005: 62) states that “accuracy rates do not relate to levels of acquisition in a linear manner and are not a valid measure of development”. The goal of the present paper is to reintroduce the question of PT and L2 French in a study adapted to the PT methodology. In what follows, I will investigate the development of written number marking and

122 Morphological Development in L2 French: A Processability Perspective

agreement in Swedish L2 learners of French. Number agreement is a grammatical procedure that has already been examined in languages other than French using the PT framework (see Pienemann and Håkansson, 1999; DiBiase and Kawaguchi, 2002; Håkansson and Norrby, 2007). However, since number morphemes in spoken French are, to a very large extent, indistinguishable (e.g. silent), this paper is concerned with the written production of number agreement. This is in line with recent studies showing that PT can account for morphological development in both spoken and written L2 production (Håkansson and Norrby, 2007; Rahkonen and Håkansson, in press). In opposition to Bartning’s and Dewaele and Véronique’s studies, this paper focuses on the development of morphology at lower levels of acquisition, i.e. initial, post-initial, intermediate and low advanced. Following the tradition of the PT framework, the analysis is based on the emergence criterion as an indicator of morphological development. In this paper, I present longitudinal data from fifteen Swedish L2 learners of French in the instructional setting. The outline of the study is as follows: After a very brief look at PT and its adaptation to French in section 2, the number system of written French is presented in section 3. These two sections give rise to a certain number of predictions, a hypothesis, and research questions, for the L2 development of written French. These are presented in section 4. Section 5 introduces the corpus on which the linguistic analysis is based. Furthermore, the methodology and acquisition criteria used are discussed in detail. Results of the longitudinal study are presented in section 6 that includes an individual and an inter-individual analysis of the L2 data. Finally, the results are summarised and discussed in section 7.

2. The processability perspective on L2 development PT predicts the sequence in which language processing routines will develop in the L2 learner. Grammatical structures will only be produced by the learner if the necessary processing procedures needed to carry out the production of the structure in question are available. The options available to the learner are thus constrained at a given time in the L2 acquisition process. Put simply, learners cannot produce linguistic structures which they are not able to process. PT is based on Levelt’s (1989) model of oral language production and is interpreted through a theory of grammar, Lexical Functional Grammar (Kaplan and Bresnan 1982; Bresnan 2001), with which it shares the idea of the unification of grammatical features. The hierarchy of processing complexity proposed by PT in its original version identifies five stages of development based on

Malin Ågren

123

different levels of information exchange between constituents, as shown in table 1 which is adapted to French.

PT level

Processing procedures

Structural outcome

Examples in French

5

Subordinate clause procedure

Info exchange between main and subordinate clause

-

4

S-procedure

Interphrasal information exchange

les filles chantent, les filles sont belles

3

Phrasal procedure

Phrasal information exchange

les (belles) filles, X a chanté, X va chanter

2

Category procedure

Lexical morphemes

garçons, filles, X chante(r), X chanté

1

Word or lemma access

Words, formulae

Bonjour, cinq fille, J’ai X, je m’appelle X

Table 1: Processing procedures, their structural outcomes and examples in French (adapted from Pienemann and Håkansson, 1999: 392) A key aspect of the developing interlanguage (IL) is the exchange of grammatical information at different structural and developmental levels. According to the PT model, learners identify and acquire individual words and invariant forms of the target language at level 1. At this level words and formulae can be combined but there is no exchange of grammatical information between them and no productive use of morphology. The next step is to categorize the lexicon and to list the diacritic features of the word,

124 Morphological Development in L2 French: A Processability Perspective

i.e. features such as definiteness, number, case or tense. Thus, level 2 is the level of category procedures which implies the use of lexical morphemes that do not ask for exchange of information between constituents. Lexical morphology is a necessary condition for phrasal morphology (level 3), which asks for exchange of grammatical information between constituents within a phrase (NP or VP). The processing of phrasal morphology permits an exchange of diacritic features between the head element and its modifiers, as in les belles filles (‘the beautiful girls’). The next step in the processing hierarchy is the automatization of inter-phrasal morphology (level 4). An example is SVagreement, les filles chantent (‘the girls sing’) or agreement between a noun and its predicative adjective in VP, les filles sont belles (‘the girls are beautiful’). Finally, at level 5, the learner is able to process the exchange of information between a main and subordinate clause and, if necessary, apply different rules to the two types of clauses. However, this stage is not applicable in the case of number morphology in French. The different stages in the processing hierarchy are acquired in an implicational order. This means that a particular processing level is a prerequisite for the following level. A recent study (Håkansson and Norrby 2007) addresses the question of whether the developmental hierarchy proposed by PT can be found in both spoken and written L2 production. According to the “steadiness hypothesis” (see Pienemann 1998: 273) “the basic nature of the grammatical system of an IL does not change in different communicative tasks as long as those are based on the same skill type in language production”. Håkansson and Norrby test this claim in the written and spoken language production of twenty L2 learners of Swedish in different communicative tasks. The results seem to indicate that processing constraints guide both oral and written L2 production and can be used to predict language development in both modalities, if the tasks are focused on communication. In addition, resemblances between morphosyntactic phenomena in both modes of production have recently been observed in Granfeldt’s (2007) study of oral and written production in L2 French by Swedish learners. .

3. Number agreement in (written) French compared to other languages A characteristic of written French is the so-called “silent morphology” with regard to verbal homophones, as well as gender and number

Malin Ågren

125

agreement in both NP and VP. The discrepancy between the oral and the written mode leads to a difficulty among learners of L1 French (and, for that matter, even experienced writers) to acquire and correctly produce these language items in writing (Jaffré and David 1999; Brissaud and Jaffré 2003). Unlike French, many other languages, e.g. English, German, Spanish, and Italian, are characterized by homology between the spoken and written expression of number. In French, however, the writer cannot take advantage of the oral language when expressing number morphology. The exceptions are liaison contexts (les_enfants, ils_entendent) and the determiner which have audible plural forms. The silent morphology can be seen in the comparison between (1) and (2) where the sole audible difference between the singular and the plural contexts is the pronunciation of the determiner (le [lΩ] > les [le]). (1) Le petit garçon parle avec la grande fille TheSG smallSG boySG talkSG to theSG bigSG girlSG’ ‘The small boy talks to the big girl’ (2)

Le|s petit|s garçon|s parle|nt avec le|s grande|s ThePL smallPL boyPL talkPL to thePL bigPL fille|s girlPL ‘The small boys talk to the big girls’

In the case of written French, both lexical and phrasal number agreements are facilitated by the comprehensive regularity of the plural morphemes: -s in NP and –nt in VP 1 . In the majority of cases, plural is expressed with the addition of –s to the noun and to the elements in NP that agree with it. Furthermore, a majority of French verbs agree in number (3pl) with the morpheme –nt added to the verb stem. In fact, this morpheme is present in both regular verbs (e.g. parler, ‘talk’) which constitutes 90 % of all French verbs, and in irregular verbs like prendre (‘take’) and dire (‘say’), even though the latter also include a stem alternation in 3rd person plural. Four very frequent verbs, être (‘be’), avoir (‘have’), aller (‘go’) and faire (‘do’), agree in 3rd person plural with a suppletive form (see below). Considering the interrelation of the two plural morphemes –s and -nt, Fayol (2003) 1

In accordance with Dubois (1967:35), only the alternation between singular and plural in the 3rd person will be considered in this paper: il/elle (‘he/she’) versus ils/elles (‘they’). First and second person plural, nous (‘we’) and vous (‘you’), should not be considered as the plural equivalents of je (‘I’) and tu (‘you’), but as different persons.

126 Morphological Development in L2 French: A Processability Perspective

states that –s in NP is much more frequent than –nt in VP. Firstly, the morpheme -s occurs on several constituents (N, Adj, Det, Pro) while –nt only occurs on verbs. Secondly, nouns in the plural are more frequent than verbs in the 3rd person plural. The learners of the present study are not unfamiliar with number agreement since their L1, Swedish, is a language that expresses number within NP on nouns, determiners and adjectives as is shown in (3). Even so, it is important to mention the fact that Swedish verbs do not agree in number or person, as seen in (4) and (5). (3)

röd|a boll|ar|na De The:PL:DEF red|PL ball|PL|DEF ’The red balls’

(4)

spela|r fotboll pojk|en boy|SG.DEF play|PRES football ‘the boy plays fotball’

(5)

fotboll pojk|ar|na spela|r boy|PL|DEF play|PRES football ‘the boys play football’

Furthermore, English is the “first” L2 to be learnt by children in Swedish schools from the age of eight/nine years. It is noteworthy that in English the same morpheme –s (audible) is used to mark plural on nouns (deux voitures – two cars). In English however, neither adjectives nor verbs agree in the plural (6). In fact, in English, the morpheme –s on the verb stem indicates 3rd person singular (!). (6)

ENG (sg): FR (sg):

play|s football The young boy Le jeune garçon joue au foot

ENG (pl): FR (pl):

The young boy|s play football. Le|s jeune|s garçon|s joue|nt au foot.

A small group of French nouns and adjectives (2 %) mark plural with an irregular –x : le bateau – les bateaux (‘the boat’ – ‘the boats’) (see New, 2004). However, in the context of the present study, the difference between regular and irregular plural morphemes is factored out. We consider the presence of a plural morpheme in a plural context as an instance of agreement, whether it is regular (les ballons, ‘the balls’),

Malin Ågren

127

irregular (les bateaux, ‘the boats’) or non-target-like (les *bateaus, ‘the boats’). As far as the determiner is concerned, gender is only expressed in the singular. In the present analysis, all forms of plural determiners followed by a noun in the plural are counted as instances of plural agreement (les, des, ses, leurs, mes, ces filles; ‘the,- , his/her, their, my, these girls’). Thus, a non-agreement between these two elements could be illustrated in two ways: either by the absence of a plural determiner (*le ballons) or by the absence of a plural morpheme on the noun (les *ballon). However, numeric quantifiers (deux ballons) and other quantifying elements (beaucoup de ballons, plusieurs ballons, ‘many/several balls’), characterized by a root that lexically expresses plural, are analysed separately as proposed by Charters and Jansen (2007). While determiners always precede the head noun in French, the attributive adjective in NP can be placed in the pre-nominal position, as in les petits ballons (‘the small balls’) or in the post-nominal position, as in les ballons rouges (‘the red balls’). Only a small group of short and frequent adjectives can be placed in the pre-nominal position. In this study the position of attributive adjectives is factored out. Number agreement in attributive adjectives is compared to the agreement in predicative adjectives found in VP, as in les ballons sont rouges (‘the balls are red’). In VP, the analysis is focused on the regular agreement pattern expressed by the morpheme –nt added to a verb stem, whether it is a verb from the first conjugation (cf. les garçons parlent, ‘the boys talk’) or from the second or third conjugation (cf. les filles choisissent, ‘the girls choose’, or les filles prennent, ‘the girls take’). The audible stem alternation in plural that characterizes many irregular French verbs is factored out in this study and we consider the presence of the number morpheme –nt on a verb stem as a sign of number agreement (cf. Pallotti, 2007). The group of four very short and frequent verbs that express 3rd person plural with an audible suppletive form have been separated from the “regular group” (cf. il est / ils sont (être, ’be’), il a / ils ont (avoir, ’have’), il fait / ils font (faire, ’do’) and il va / ils vont (aller, ’go’). Note however that these suppletive forms have the same ending (nt) as the regular plural morpheme.

4. Hypothesis, predictions and research questions The hypothesis of the present study is that the morphological development in L2 French should follow the hierarchy proposed by PT, as shown in table 2. Our focus is on stages one to four of the hierarchy, from the

128 Morphological Development in L2 French: A Processability Perspective

expression of individual items (level 1) to the level of interphrasal agreement (level 4). Procedures

Examples

Subordinate clause procedure Inter-phrasal (S-) procedure Phrasal procedure Category procedure Word or lemma access

les filles (elles) chantent les belles filles Filles Fille

Order of development 1 2 3 4 -

5 +

-

-

-

+

+

+

+ +

+ + +

+ + +

+ + +

Table 2: Implicational sequence of processing procedures and examples from French Following the PT hierarchy, we hypothesize that number agreement in French should be produced in the following order: 1) Bare words / invariant forms. 2) Plural on nouns (lexical agreement). 3) Plural agreement between nouns, determiners and attributive adjectives (phrasal agreement). 4) Plural SV-agreement (inter-phrasal), agreement on predicative adjectives. The use of level 5 structures is not relevant to the present study of number agreement in L2 French and will not be considered in this paper. In the following section, the proposed order of production of number morphology will be empirically tested on the written production of L2 French in the CEFLE Corpus (see method). More specifically, the following research questions will be addressed: - Does the development of number marking and agreement in the written production of French among Swedish L2 learners follow the predictions of PT? If not, what other factors can account for the findings? - Are there characteristics in the L2 French agreement system which influence the interpretation in terms of PT? If there are, what are these characteristics and how should they be analysed?

Malin Ågren

129

5. Corpus and method The texts analysed in this study come from the CEFLE Corpus, Corpus Écrit de Français Langue Étrangère (‘Written Corpus of L2 French’), which contains approximately 400 written texts from Swedish high school L2 learners of French and from a control group of French adolescents. This corpus was created at Lund University, Sweden, to follow the morphosyntactic development of written L2 French (Ågren, 2005). In this paper, a longitudinal study is presented, which follows the individual development of fifteen learners over a year in school.

5.1 Informants The participants in the present study are high school L2 learners of French at different levels of proficiency. Their L1 is Swedish and, at the beginning of the study, they have been studying French at school for between four and five years (age 16-18 years). All subjects filled in a small background questionnaire concerning their mother tongue, other languages studied, years of studying French, age of onset, experiences of the French language and visits in France, etc. Furthermore, all subjects did a test which permits us to estimate their general level of grammar and vocabulary. All of the participants in the CEFLE corpus have been grouped into four different levels of acquisition of L2 French, i.e. initial, post-initial, intermediate and low-advanced, based on independent linguistic criteria in several key domains for Swedish L2 learners of French, such as sentence structure, finiteness, mode and expression of time and aspect (see Bartning and Schlyter 2004).

130 Morphological Development in L2 French: A Processability Perspective

Learner

Sex

Age

Andrea Anders Amie

Develop. Stage (B&S, 2004) Initial Initial Initial

Years of French

Other L2s

Female Male Female

16/17 16/17 16/17

4 3 4

Male Female

16/17 16/17

4 4

Post-initial Post-initial Intermediate Intermediate

Female Male Male Male

16/17 16/17 16/17 16/17

4 4 4 4

Carolina Cecilia Deborah Dagmar Daniella

Intermediate Intermediate Low advanc. Low advanc. Low advanc.

Female Female Female Female Female

16/17 16/17 17/18 17/18 17/18

4 4 5 5 5

Daisy

Low advanc.

Female

17/18

5

English English English Danish English English Serbian English English English English Spanish English English English English English Spanish English Spanish

Bror Bibbi

Post-initial Post-initial

Barbara Billy Caesar Conny

Test result (max. 60) 10 8,5 13 23 20 23,5 41 41 42,5 45 50 56 58 57,5

Table 3: The participants in the longitudinal study In addition, the corpus includes texts from a control group of age-matched French adolescents from two different high schools in France (for details about the control group, see Ågren, 2008, in press). It is important to mention the fact that all participants have studied English since the age of 8-9. Thus, English should be considered their “first” second language. In addition, some of them have experiences from other foreign languages 2 . Relevant information on the subjects of the study is presented in table 3.

2

Three learners are beginners in Spanish which is to be considered as their fourth L2. According to the background questionnaire, they are much more experienced learners of French than of Spanish.

Malin Ågren

131

5.2 Data collection The participants were followed during the period of one school year, from September 2003 to May 2004. Data were collected with an interval of approximately two months. At each data collection point, the learners were told to write a text in French on a computer. The texts are narratives based on personal memories or picture stories, where the learners’ attention was focused on the contents and the transmission of the message rather than on form. Two picture stories were created to elicit number and gender agreement: L’homme sur l’île (‘The man on the island’) and Le voyage en Italie (‘The trip to Italy’)) whereas two texts had a more general design (Un souvenir de voyage (‘Memories from a journey’) and Moi, ma famille et mes amis (‘Me, my family and my friends’). For the purpose of the present study, the text Un souvenir de voyage, ‘Memory from a journey’, is excluded due to specific problems caused by the elicitation of past tense. The three texts referred to in this study and the time of their collection is indicated in table 4. Text1 Text2 Text3

Titel of the task L’homme sur l’île ‘The man on the island’ Le voyage en Italie ‘The trip to Italy’ Moi, ma famille et mes amis ‘Me, my familly and my friends’

Type of task Picture story

Data collection September 2003

Picture story

January/February 2004 May/June 2004

Picture story + Personal narrative

Table 4: Description of the tasks in the longitudinal study

132 Morphological Development in L2 French: A Processability Perspective

In the picture stories, differences in nationality, size, form and colour were used to stimulate the use of adjectives that are often very rare in spontaneous language production. The learners were instructed to tell their story in as much detail as possible to someone who did not know them and who could not see the picture story. They had a maximum of 50 minutes at their disposal but many of the learners did not use the whole time span to accomplish the task. The texts in L2 French vary in length from approximately 100 to 600 words. The plural contexts have been studied in detail with focus on nouns, pronouns, determiners, adjectives and verbs. In the present study, we do not differentiate regular from irregular morphology. Thus, all forms of plural agreement are taken into account. However, the suppletive verb forms in the plural SV-agreement are treated separately. Note that our focus is on morphology and therefore spelling problems are not accounted for.

5.3 Acquisition criteria and method As is well known, acquisition can be, and is, defined in many ways. According to Pallotti (2007) this is a problem for the field of SLA since it is difficult to compare results and test theories. Acquisition criteria are formulated because researchers need a standardized way of recognizing the presence or absence of structures, as well as distinguishing one level from another. One criterion that is applied in many studies is that of 90 % of correct use in obligatory contexts (Brown 1973; Dulay and Burt 1974). Another criterion is TLU (target-like use) which is regarded to be a more sensitive developmental index (Pica 1984) that accounts for both omissions and overuse. The common denominator of these criteria is their emphasis on learner production (errors) compared to the target language. Other scholars claim that the distance between non-native and native grammars is an irrelevant criterion and emphasize that the IL should be studied in its own right. These scholars have instead based the acquisition criteria on the first uses of a structure: “onset” (Hammarberg 1996) or “emergence” (Pienemann, 1998). Pienemann (1998: 138) defines the emergence criterion as follows: “the point in time at which certain skills have, in principle, been attained or at which certain operations can, in principle, be carried out”. This criterion focuses the beginning of the acquisition process. Pienemann’s definition of the emergence criterion is used in the PT framework to account for developmental sequences in IL. However, it seems that the emergence criterion is used in slightly different ways by different scholars.

Malin Ågren

133

According to Pienemann and others, the learner has to produce more than one case of a structure in order to ensure that the structure in question is really processed by the learner (see Pienemann and Håkansson, 1999; DiBiase and Kawaguchi, 2002 or Håkansson and Norrby, 2007, among others). Zhang (2005: 166) uses “a minimum of three tokens in lexically varied contexts” which reduces the chance of mistaking morphemic chunks for productive use of specific morphemes. Pallotti (2007) argues in favour of a homogenization of the emergence criterion and proposes three different sources of evidence for emergence to be stated: At least two minimal pairs, or two creative constructions, or a certain amount of lexical variety in the application of a grammatical morpheme on three different lexical items (tokens). The present longitudinal study adopts the emergence criterion of Pallotti (2007: 375). A distributional analysis of number marking and agreement in singular and plural contexts was carried out on three texts of all fifteen learners. Then, the findings were summarized in individual tables. When the emergence criterion is met, this is indicated with the sign +. The plus sign within brackets (+) indicates that emergence is close at hand, often two occurrences out of two contexts. If the emergence criterion is not met in a text that presents a sufficient number of contexts, the sign - is used. A slash (/) indicates that the learner does not create a sufficient number of contexts for a certain structure in the text. The slash indicates that it is impossible to tell if a structure has emerged or not. When analyzing the L2 texts, special attention has been paid to minimal pairs (MP) and to creative constructions (CC) as exemplified in (7) and (8). According to many authors, e.g. Pallotti (2007), creative constructions are very interesting signs of productive use of a structure in the developing IL system. (7)

Nominal MP: garcon – garcons Adjectival MP: la belle fille – deux belles filles Verbal MP : le garçon parle – les garçons parlent

(8)

Nominal CC: bateaus (instead of bateaux) Adjectival CC: les nouveaus vélos (instead of nouveaux) Verbal CC : les garçons allent (instead of vont) les garçons prendent (instead of prennent)

Finally, as part of a qualitative analysis of the material, examples of learner production will be presented and discussed. If the example contains an omission of number agreement, this is indicated with (*) in front of the item in question (ex. les *voiture) and if the plural agreement

134 Morphological Development in L2 French: A Processability Perspective

is produced in a non-native way, this is indicated with the abbreviation NNPL (non-native plural) in the English translation.

6. Results from the longitudinal study In the following section, the production of number marking and agreement will be examined over time in individuals and between subjects. Our goal is to discuss the interpretations of the data in terms of PT and to raise possible issues that may prove difficult when analysing L2 French in the proposed theoretical framework. Suggestions for further deepening of the analysis are presented in the discussion at the end of this paper.

6.1 Inter-individual development Table 5 below is presented in order to provide a global view of the different linguistic levels of the subjects. This table illustrates the production of number morphology at an individual level in the second text Le voyage en Italie (‘The trip to Italy’) as exemplified through lexical agreement on the noun (level 2), phrasal agreement between the noun and the determiner (level 3) and SV-agreement at the inter-phrasal level (4). As illustrated in table 5, and in line with the predictions of PT, the emergence of number morphology on nouns takes place at a very early stage. Anders (Ans), who is at a very low level of L2 French, is the only learner that does not produce a sufficient number of contexts in any of the agreement domains analyzed. According to the emergence criterion adopted in this study, at least two of the subjects, Andrea (And) and Billy (Bil), are located at level 2, since the emergence of number morphology is established exclusively at a lexical level. Furthermore, Bror (Bro), Barbara (Bar) and possibly Amie (Ami) produce the phrasal Det-N agreement, therefore they are at level 3, but no emergence is stated at the inter-phrasal level between the subject and the verb (level 4) among these three learners. The remaining nine learners show emergence of number agreement at all three levels of processing. The pattern observed when analysing the emergence of plural morphology on the above mentioned constituents resembles the stepwise development predicted by PT.

Malin Ågren

Agreement in obligatory contexts Level 2 Level 3 Level 4 Lexical IntraInter(noun) ph. ph. (Det-N) (S-Vreg) Dai

18/18

13/13

Dan Dag

18/18 28/28

13/13 27/27

Deb Car Cec Con Cae Bib Bro Bar Ami Bil And Ans

18/20 9/13 15/17 6/7 8/8 8/14 6/7 5/5 10/12 7/8 4/4 0/1

16/16 9/12 10/11 5/6 1/1 6/10 4/5 3/3 2/2 0/6 0/1 0/1

1/1 (Im10/1 0) 5/5 2/2 (Im 0/9) 9/12 16/16 11/13 9/14 5/12 10/10 2/7 0/8 2/7 1/ 4 0/4 0/1

135

Emergence Level 4 Interph. (S-Vreg)

+

Level 3 Intraph. (DetN) +

+ +

+ +

+ (+/-)

+ + + + + + + + + + + /

+ + + + / + + + (+) / /

+ + + + + + /

Level 2 Lexical (noun)

+

Table 5: Number agreement in text 2 “Le voyage en Italie” in fifteen subjects: agreement in obligatory contexts (to the left) and emergence (to the right). Another important observation in table 5 is that a ceiling effect, due to the emergence criterion, seems to appear already at rather low levels. In the data of nine of the learners, number agreement has already emerged at all three levels. One can easily get the false impression that for example Bibbi (Bib) and Daniella (Dan) are at approximately the same level of producing number agreement. However, if we broaden the scope of our investigation and examine different constituents within each level, important differences between individuals can be observed. Table 6 gives a more revealing picture of the emergence of number morphology between different constituents at level 2, 3 and 4. This table discriminates determiners, quantifiers and attributive adjectives at phrasal level (3) and

136 Morphological Development in L2 French: A Processability Perspective

the regular (–nt) versus the irregular (sont/ont) SV-agreement at the interphrasal level (4).

Dai Dan Dag Deb Car Cec Con Cae Bib Bro Bar Bil Ami And Ans

Level 2 Lexical (N-s) + + + + + + + + + + + + + + /

Level 3 Intra-phrasal (Q+N) (Det+N) (N+adj) + + + + + + + + + + + + (+) + + + / + (-) + / / (+) + (+) + (+) + (-) + + (+) + / / / / /

Level 4 Inter-phrasal (S+Virr) (S+Vreg) + + + + + (+/-) + + + + + + + + + + + + + + + (-) / / /

Table 6: Emergence of lexical agreement on nouns (level 2), phrasal agreement on quantifiers, determiners, attributive adjectives (level 3), inter-phrasal SV-agreement (level 4) in 15 individual learners, text 2. As observed in table 6, number agreement between the noun and a quantifier contrasts with that of the noun and a determiner, on the one hand, and even more so with the agreement between the noun and an attributive adjective, on the other hand; all three of them are part of a phrasal agreement at level 3 according to PT. The production of a quantifier in relation to a noun in the plural is a preferred and correctly used structure from early on whereas the adjective agreement only seems to emerge in the most advanced learners. This observation actualizes the question of potential differences in processing load when different structures at the same level in the PT hierarchy are involved in the agreement. It has been argued that the combination of a quantifier and a noun in the plural is more of a co-activation of two lexically marked elements than a unification process with the exchange of grammatical information as in the case of Det-N agreement (Charters and Jansen 2007).

Malin Ågren

137

Following this interpretation, the co-activation process should emerge before the unification process. The data of for example Billy could indeed be interpreted in this direction. However, even when exclusively considering the differences in production of the Det-N versus the N-Adj agreement, the time of emergence of these two phenomena is strikingly different. Four of the subjects show an emergence of number agreement on the determiner whereas the attributive adjectives, on the other hand, clearly do not agree in number (Bror, Bibbi, Cecilia and Carolina). According to PT, these two agreements are constrained by the same processing procedure at phrasal level. Thus, there must be other factors that can explain why these two agreements emerge at different times in the IL development. We will return to this in the discussion section of this paper. Another interesting observation in table 6 is that among learners at post-initial or intermediate levels, number morphemes emerge on the verb (inter-phrasal morphology) even though they are not yet present on attributive adjectives. The same result has been found in spoken L2 French (Prodeau 2005). The SV-agreement in written French seems to appear relatively early, albeit not always according to the norm of the target language. In the following individual analysis we will have a closer look at this SV-agreement.

6.2 Individual development After having a brief look into the texts of subjects located at the initial level (Amie) versus the low-advanced level (Daisy), the production of number morphology in three subjects at post-initial and intermediate levels will be emphasized (Billy, Bibbi and Cecilia). When using the emergence criterion as a measure of morphological development, it seems that the most interesting observations are found in subjects at the intermediate levels of proficiency. Focusing on the more detailed level, tables 7 to 12 discriminate quantifiers, determiners and adjectives at the phrasal level and regular verbs from irregular verbs at the inter-phrasal level. The developmental pattern in Amie, table 7, and Daisy, table 8, (or rather, lack thereof) is representative for several subjects at initial and low advanced levels respectively. The common denominator of these groups is that they do not show any marked tendency to develop over the period of observation. At her very low level of proficiency, Amie has difficulty in producing any text at all. Her texts are very short and they have a strong nominal organisation in which verbs and adjectives are rare. As can be read out of the many slashes in table 7, she creates very few plural

138 Morphological Development in L2 French: A Processability Perspective

contexts. However, in her second and third texts, plural contexts are sufficient to tell that plural agreement at phrasal and inter-phrasal levels have not yet emerged in her IL system, as exemplified in (9) and (10). (9) Amie, txt2

Karin et Josefin *boire rouge vin… ‘Karin and Josefin drinkINF red wine’

(10) Amie, txt3 Elle a *petit *petit mains ‘She has smallSG smallSG handPL’

PT level

Morphological structure

4

Subj+Vreg Subj+Virr

3 2

N+V+ADJpred ADJattr+N DET+N Quant+N N-s

Occurrences T1 / 1/1 1mp / 0/1 2/2 2/2 4/4

Emergence

T2 2/7 0/2

T3 0/1 /

T1 / /

T2 (-)

T3 / /

/ 0/1 2/2 8/10 10/12 1mp

/ 0/3 1/2 4/4 6/7 3mp

/ / (+) (+) +

/ / (+) + +

/ / + +

Table 7: Production of number morphology in AMIE, initial level, texts 1, 2 and 3 3 Interestingly, the only two occurrences of SV-agreement in Amy’s data appear with a pronominal subject (ils parlent et ils dancent) whereas all omissions of SV-agreement are preceded by an NP-subject. This is an interesting observation which might be an indication of an earlier emergence of SV-agreement with a pronominal subject (weak) than with an NP-subject (strong). This phenomenon could point in the direction of a level 3 interpretation of the Pronoun-Verb agreement (see discussion). Daisy, on the other hand, has a well developed L2 French which she demonstrates in long and elaborate texts. In her case we can observe a ceiling effect due to the emergence criterion. 3 Table caption: 2/7: Agreement is produced in two out of seven contexts; mp: minimal pair.

Malin Ågren

PT level

Morphological structure

4

Subj+Vreg

Occurrences T1 5/7 3mp 1cc

Subj+Virr

3/3 1mp

N+V+ADJpred

3/4 3mp 5/6

3

ADJattr+N

2

DET+N Quant+N N

139

26/26 6/6 31/31 3mp

T2 1/1 Imp 8/8 2mp, 2cc Aux 20/20 Imp 6/7 Pqp 3/3 3/3 9/9 1mp 13/13 7/7 18/18 4mp

Emergence T3 7/8 3mp

T1 +

T2 +

T3 +

3/3 2mp

+

+

+

3/3 1mp 6/6 2mp 20/20 13/13 33/33 6mp

+

+

+

+

+

+

+ + +

+ + +

+ + +

Table 8: Production of number morphology in DAISY, low advanced level, texts 1, 2 and 3 4 Number agreement is highly automatized in Daisy’s IL system, as illustrated in table 8. A brief look at the minimal pairs in all three texts reveals that Daisy’s expression of number is at a very high level of correctness on all constituents in the sentence. This is examplified in (11). (11) Daisy, txt3:

4

un chien / deux chiens ma sœur / mes sœurs ma meilleure amie / mes meilleures amies elle aime se baigner / ils aiment se promener

Table caption: mp: minimal pair; cc: Creative construction; Aux: The irregular verb is used as an auxiliary verb; Imp: The verb is used in the imperfect; Pqp: The verb form is in the pluperfect.

140 Morphological Development in L2 French: A Processability Perspective

Let us now take a look at the developing IL system of a few subjects at the post-initial and intermediate levels. Table 9 and 10 illustrate the morphological development in Billy and Bibbi respectively. PT Lev el

Morphological Structure

4

Subj+Vreg

T1 0/4

Subj+Virr

/

3 2

N+V+ADJpred ADJattr + N DET + N Quant + N N

Occurrences

/ 0/3 0/7 4/5 10/1 2 3mp 1cc

T2 0/3 1cc 7/8 1cc 0/1 0/3 0/6 6/6 7/8 1mp

T3 0/0 1cc 1/1 1mp 0/2 / 2/3 1/1 3/3

Emergence T1 -

T2 -

T3 /

/

+

/

/ + +

/ + +

(-) / (+) / +

Table 9: Production of number morphology in BILLY, post-initial level, texts 1, 2 and 3 5 In Billy’s three texts, one observes the lexical agreement on nouns and in combination with quantifiers but neither determiners nor adjectives agree in number. However, it seems that the phrasal Det-N agreement is on its way to emerge in the last text 6 . In his first text, Billy produces nouns in the plural with a determiner in the singular, as in (12), while a determiner in the plural can be found in the third text on two out of three contexts, as in (13). (12) Billy, txt1: *son amies ‘hisSG friendPL’ *le banque-hommes ‘theSG bankmanPL’

5

Table caption: see note 4 at table 8. This development is also observed in another learner at the post-initial level: Barbara. In her first text, she still uses determiners in the singular with a noun in the plural (*le monsieurs, * le vélos…) while these forms disappear in favour of the correct Det-N agreement (les vêtements, les plats…) during the observation period. 6

Malin Ågren

141

(13) Billy, txt3: mes argents ‘myPL moneyPL’ mes deux sœurs ‘myPL two sisterPL’ Adjective agreement has clearly not emerged in Billy’s IL, neither at a phrasal nor at an inter-phrasal level, as illustrated in (14) and (15). (14) Billy, txt2: deux *italien *cèlebite playboys ‘two ItalianSG famousSG playboyPL’ (15) Billy, txt3 : …la deux [sœurs] sont *térrible et *malin TheSG two (sisters) bePL terribleSG and trickySG’ Concerning SV-agreement, Billy’s second text shows the emergence of plural morphemes in the group of irregular verbs (Virr). Interestingly, Billy is elaborating with several plural forms (vont, avons etc.) which sometimes results in the production of non-native forms that nevertheless agree in number (even if not in person) with a plural subject. A highly significant sign of productivity is that of “creative constructions” that do not correspond to the norm of the target language. Such creative constructions are present in texts 2 and 3 on both types of verbs, as shown in (16) and (17) (underlined). (16) Billy, txt2: Ils prendons ses valise and allons de la voiture ‘theyPL takeNNPL theirPL suitecaseSG and goPL from the car’ mais ils avont trop beaucoupe valises ‘but theyPL haveNNPL to many suitecasePL’ (17) Billy, txt3: les deux (sœurs) …prendent mes argents ‘thePL two (sisters) takeNNPL myPL moneyPL’ Even if the number of plural forms in VP is small, their presence in the texts seems to indicate that the emergence of number agreement at the inter-phrasal level is close in Billy’s case. Considering the nature of the subject in the SV-agreement, it seems that the first time that a plural morpheme appears in Billy’s texts it is accompanied by a pronominal subject (ils prendons) whereas it is preceded by an NP subject in his third text (les deux prendent…) 7 . In the case of Billy, it seems that the 7

Even though there is not enough space to include a detailed analysis of Caesar’s written production in this paper, it is worth mentioning that the SV-agreement in Caesar’s texts is similar to Billy’s in the sense that the SV-agreement first appears

142 Morphological Development in L2 French: A Processability Perspective

emergence of SV-agreement at level 4 is accompanied rather than preceded by the plural Det-N agreement at level 3. The production of non-native SV-agreement can be observed in Bibbi’s texts as well (table 10). Bibbi produces the same kind of non-target-like SV-agreement as Billy, used with both groups of verbs (Virr and Vreg), as exemplified in (18) (underlined). (18) Bibbi, txt2 : Les filles, Monique et Jeanne, dessaient au table. elles boisent de café au lait et parlent si le voyage en Italie cette été. ‘ThePL girlPL, M and J, VERB.PL at the table. ShePL drinkNNPL café au lait and talkPL about the trip to Italy this summer’ elles parlons et riront beaucoup ‘shePL talkNNPL and laughNNPL a lot’ In two instances, this emerging SV-agreement is preceded by an NP subject whereas the majority of occurrences (8/10) are produced including a pronominal subject. Even if the preferred subject in Bibbi’s IL is pronominal we cannot conclude that the SV-agreement is emerging first with this kind of subject.

with pronominal subjects. Text 2: Pronominal subject with SV-agreement (5/8), NP subject with SV-agreement (0/4).

Malin Ågren

PT level

Morphological structure

4

Subj+Vreg

Occurrences T1 1/1

Subj+Virr

3/3 1mp

N + V + ADJpred ADJattr + N DET + N Quant + N N

T2 8/8 1mp,1cc Imp 1/1 4/4 Aux 1/1 1mp,1cc 0/2 1/5 6/10 2/3 8/14

1/2 / 1/1 2/2 2/2 2 1mp 1cc Table 10: Production of number morphology level, texts 1, 2 and 3 8 3

143

Emergence T3 /

T1 /

T2 +

T3 /

1/1 1mp

+

+

/

0/1 / 3/3 2/2 5/5

/ / / (+) +

(-) + (+) +

/ / + (+) +

in BIBBI, post-initial

However, adjective agreement at phrasal level has not yet emerged in Bibbi’s IL. Only in one out of five contexts does the adjective agree with a noun in a plural context. As illustrated in (19), most adjectives do not agree in number. (19) Bibbi, txt2:

les *long, *rouge rideaux ‘thePL longSG redSG curtainPL’

Table 11 presents the development of number morphology in Cecilia. It is obvious from the first text onwards that SV-agreement has emerged in both groups of verbs (and with both types of subjects), as seen in the number of minimal pairs (see 20 and 21) and the level of correct agreement in plural contexts.

8

Table caption: see note 3 at table 8.

144 Morphological Development in L2 French: A Processability Perspective

PT level

Morphological structure

4

S + Vreg S + Virr

3

N+ V + ADJpred ADJattr + N

2

DET + N Quant + N N

Occurrences T1 6/6 2mp,1cc ¾ 3mp Aux 2/2 ½ 1mp 1/1 1mp 12/13 4/4 7/8 3mp,7cc

Emergence

T2 11/13 2mp,2cc 10/10 2mp Aux 3/4 1/1

T3 3/3 2mp 3/3 2mp

T1 +

T2 +

T3 +

+

+

+

0/2

/

/

(-)

0/5

2inv

/

-

/

10/11 6/7 15/17 2mp

7/8 7/7 12/13 2mp

+ + +

+ + +

+ + +

Table 11: Production of number morphology in CECILIA, intermediate level, texts 1, 2 and 3. (20) Cecilia, txt1: Pierre aime / Les oiseaus n’aiment pas ‘Pierre likeSG’ / ‘thePL birdPL not likePL’ un chien est noir/ deux oiseaus sont sur l’île ‘aSG dog SG beSG blackSG’ / ‘ two birdPL bePL on theSG islandSG’ (21) Cecilia, txt2: Le garcon achete / La femme et le garcon achetent ‘TheSG boySG buySG’ / ‘theSG womanSG and theSG boySG buyPL’ Le soleil tombe / Deux garcons tombent ‘TheSG sunSG fallSG’ / ‘two boyPL fallPL’ Once again, the data show that adjective agreement at phrasal level does not seem to precede SV-agreement at the inter-phrasal level. In her first text, Cecilia produces one adjective that agrees in number (des chapeaus noirs) but according to our emergence criterion, this is not enough to state that this agreement is productive. In her second text, Cecilia does not produce a single adjective that agrees in number, as shown by the examples in (22).

Malin Ågren

145

(22) Cecilia, txt2: des cheveux *blond ‘det.PL hairPL blondSG’ des garçons *italien ‘det.PL boyPL ItalianSG’ These data clearly show that all agreements at level 3 do not emerge before those of level 4. On the contrary, in Cecilia’s texts SV-agreement has emerged whereas attributive adjective agreement has not.

8. Summary and discussion It would seem that the answer to the first research question regarding the emergence order of number agreements in L2 French depends upon which level of analysis it is that is adopted. When looking at the development of number morphology, the main results of the longitudinal study can be summarized as follows: x

x

Plural morphemes appear on nouns very early (fille-s) which corresponds to the predictions of PT. In PT terms, this is a “lexical agreement” that does not call for any exchange of grammatical information between constituents. All learners except one show a clear emergence of nominal plural marking. Further, as predicted by the theory, Det-N agreement (les filles) at phrasal level follows the lexical marking. Moreover, the Det-N agreement precedes the SV-agreement (les filles chantent) at inter-phrasal level. Hence, when considering these constituents, the predictions based on PT could account for the morphological development in individuals. When lexical, phrasal and inter-phrasal procedures were analyzed longitudinally, an individual development was seen in a small number of learners during the observation period. For example, the emergence of Det-N agreement (les filles) or SV-agreement (ils prendent) was observed in some individuals. Furthermore, in other subjects, phrasal and inter-phrasal morphology seem to appear simultaneously rather than one after the other. Generally, the step-by-step development predicted by PT could only to a limited extent be seen in the individual data. This is most probably due to the fact that the study covers a too short time span (9 months) and that the acquisition criterion used does not permit further analysis once the emergence has been stated. This result might in fact suggest a combination of different acquisition criteria at more advanced levels.

146 Morphological Development in L2 French: A Processability Perspective

x

When applying a more fine-grained analysis to the data, interesting differences in agreement were found. The emergence differences within one PT level cannot be explained by the PT hierarchy. It is necessary to take other linguistic factors into account when comparing quantifiers, determiners and adjectives at a phrasal level, or when looking at the SV-agreement at an inter-phrasal level, which can include two kinds of subjects: nominal or pronominal.

These observations call for a discussion of other possible intervening factors alongside the processing constraints accounted for by PT. When comparing determiners with quantifiers, for example, the question arises as to whether it is, in effect, the same procedure that is at work when a quantifier is used with a noun in the plural (deux filles) as when a determiner is used (les filles). Data show that a quantifier in combination with a noun seems to pattern more with the lexical level (2) than with the other morphemes at the phrasal level (3). Indeed, some scholars (Charters and Jansen, 2007) suggest that the quantifier-noun relation is more of a coactivation process of two plural elements than of a phrasal agreement between a head and its modifier. In the longitudinal data, all learners (frequently) use the combination of a noun in the plural with a quantifier, even those who have not yet acquired the Det-N agreement. Thus, this result suggests a separate analysis of the two phenomena in relation to the PT-levels. If one wants to analyse a true phrasal agreement in French, the Det-N agreement seems the most appropriate. Furthermore, all but three learners express the Det-N agreement in the plural. This result stands in sharp contrast to the lack of adjective agreement on attributive adjectives which, according to PT, is situated at the same level of processing. The divergence is so important that an analysis comparing N-Adj agreement at phrasal level with SV-agreement at inter-phrasal level would come up with results contradicting the PT hierarchy. The same observation has been reported in Dewaele and Véronique (2001) concerning gender agreement in L2 French. These authors suggest that it is the nature of the lemma itself and not the level of the processing hierarchy that affects the accuracy level of gender agreement in determiners and attributive adjectives. Dewaele and Véronique (2001: 290) remind us that determiners and adjectives are lemmata of very different nature. Unlike adjectives, determiners are “syntactic lemmata” with a close relation to the noun. Determiners, like object pronouns and subject pronouns, have a clitic, or at least weak, syntactic status in French, which is acquired gradually in L2 acquisition

Malin Ågren

147

(see Granfeldt and Schlyter, 2003, and references quoted there). It seems that the clitic/weak status of determiners as opposed to the non-clitic status of adjectives could play a role in the acquisition process. In any respect, the determiner has a “tighter” relation to its host than the adjective has to the noun. Furthermore, the frequency effect could play a role here. There are fewer kinds of determiners than there are of adjectives, and determiners are shorter and occur much more frequently. Dewaele and Véronique point out that the adjective is not systematically learned in adjective-noun pairs, which leads to the rarer use of this item in production and perception (see also Hinkel, 2001; Bartning and Schlyter, 2004). Finally, if adjective agreement emerges late in written L2 French, the same cannot be said about SV-agreement. Learners at post-initial or intermediate levels show emergence of verbal plural agreement in various ways. Some of them mix several different plural morphemes (-nt, -ons, ont) that they use in plural contexts on various kind of stems (however, never in a singular context!). Interestingly, the present study shows that number agreement emerges before person agreement. Creative constructions have been observed in many texts as a sign of productive use of plural verbal morphemes. Tendencies in the data of Amy, Caesar and Billy suggest that the number agreement emerges earlier/is more correct if the subject is pronominal rather than nominal. However, in other learners no such observations were found. As in the case of determiners, the weak syntactic status of subject pronouns in French must be considered when analysing the L2 data. In fact, if subject pronouns are analysed as “weak/deficient” in the sense of Cardinaletti and Starke (1999) (or Kayne, 1975), SV-agreement between the subject pronoun and the verb does not involve unification of features between phrases. The weak pronoun is generated directly under V and the agreement should then be considered as a phrasal procedure (level 3). However, if subject pronouns are analysed as non-clitic elements, with the same syntactic status as full NPs, the procedure involves an exchange of grammatical information between two different phrases, as in level 4 of the PT hierarchy (see DiBiase and Kawaguchi 2002 for a theoretical discussion of a LFG perspective on the clitic status of pronouns in L2 Italian). What makes this question tricky is that the syntactic status of subject pronouns might change during the acquisition process of an L2, as proposed by Granfeldt and Schlyter (2003). At initial levels, it is probable that learners treat subject pronouns like full NPs. Later in the acquisition process, however, the weak/clitic status of these elements is acquired. At the moment, we will have to leave this theoretical and methodological issue open. Future studies will treat this question in further detail (Ågren, forthcoming).

148 Morphological Development in L2 French: A Processability Perspective

Another factor that most probably causes the early SV-agreement is the regular morphological system in written French. Even though number agreement is often a so-called “silent agreement”, it presents a major advantage to the oral language: it is extremely regular. As underlined by Goldschneider and DeKeyser (2001), morphological regularity is an abstract form of salience, i.e. salience of the form-meaning relationship. A (near) one-to-one relationship between form (-s in NP, -nt in VP) and meaning (plural) makes this kind of morphology easier to acquire than a morphology with an obscure form-meaning relationship. This can be illustrated by the SV-agreement in written French. In the spoken language, number agreement in 3rd person plural is inaudible in regular verbs and in many irregular verbs the agreement corresponds to an unpredictable alternation of the final consonant (il veut – ils veulent; il sait - ils savent; il finit - ils finisissent, etc.). In writing however, 3rd person plural is always expressed with the same morphological ending: –nt. Thus, what in spoken language is irregular can in written language be very regular, something that clearly is an advantage for L2 writers. It has been observed in this study that the morphological agreement in writing (-nt) emerges before the stem alternation (lexical agreement) in the plural. To summarise, this study shows that PT can account for the morphological development in individual learners in a general way. However, when looking into details, different constituents involved in the agreement at one level can show very different emergence patterns. These differences cannot be understood solely on the basis of the PT hierarchy. We suggest a multi-factor approach as described in Goldschneider and DeKeyser (2001) to account for the morphological details in the analysis of L2 development. The present study raises the question whether the clitic/weak status of determiners and subject pronouns, as compared to non-clitic constituents, influences the emergence of number agreement in L2 French. However, this issue will have to be further investigated in the future. Other factors influencing the development of morphology are inherent to the morphemes in question, for example their high frequency in the written language system and the overwhelmingly regular pattern of number in written, as compared to spoken, French.

Malin Ågren

149

References Ågren, M. (2005) : Développement de la morphologie du nombre en français langue étrangère à l’écrit. Étude transversale, PERLES no 21, Institute for Romance Languages, Lund University, Sweden. Ågren, M. (2008, in press) “The advanced writer of French as a foreign language - A study of number agreement in Swedish learners” in Labeau, E. and Myles, F. (ed.) The advanced learner variety: The case of French. Contemporary Studies in Descriptive Linguistics, Peter Lang. Ågren, M. (forthcoming) “Évolution morphologique en français L2 écrit : étude du marquage et de l’accord en nombre chez des apprenants suédois”, (Titre provisoire), Ph.D Thesis to be defended at Lund University, Sweden. Bartning, I. (1999) “Stratégies d’acquisition dans l’emploi du genre et de l’accord adjectival en français parlé”, in M. Plénat, M. Aurnague, A. Condamine, J-P Maurel, Ch. Molinier and Ch. Muller (ed.) Mélanges de sémantique et de syntaxe offerts à Andrée Borillo par un groupe d’amis, de collègues et de disciples, Amsterdam : Rodopi. Bartning, I. (2000) “Gender agreement in L2 French : Pre-advanced vs advanced learners”, Studia Linguistica 54 (2), pp. 225-237. Bartning, I. and Schlyter, S. (2004) “Itinéraires acquisitionnels et stades de développement en français L2”, Journal of French Language Studies 14 (3): 281-299. Bressnan, J. (2001) Lexical-Functional Syntax, Oxford: Blackwell. Brissaud, C. and Jaffré, J.P. (2003) “Présentation : Regards nouveaux sur la lecture et l’écriture”, Revue Française de Linguistique Appliquée, 2003, Volyme VIII (1): Regards nouveaux sur la lecture et l’écriture. Brown, R. (1973) A First Language : The Early Stages, Cambridge MA : Harvard University Press. Cardinaletti, A. and Starke, M. (1999) “The typology of structural deficiency”, in H. van Riemsdijk (ed.) Clitics in the Languages of Europe, Berlin/New York: Mouton de Gruyter, pp. 145-233. Charters, H. and Jansen, L. (2007) “Do we agree? The SLA of number concepts, features morphs and agreement”, presentation held at the EUROSLA Conference in Newcastle, England, September 2007. Dewaele, J-M et Véronique, D. (2000) “Relating gender errors to morphosyntax and lexicon in advanced French interlanguage”, Studia Linguistica, 54 (2), pp. 212-224. Dewaele, J-M et Véronique, D. (2001) “Gender assignment and gender agreement in advanced French interlanguage : a cross-sectional study”, Bilingualism : Language and Cognition, 4 (3), pp. 275-297.

150 Morphological Development in L2 French: A Processability Perspective

DiBiase, B. (2002) Developing a second language. Acquisition, processing and pedagogy of Arabic, Chinese, English, Italian, Japanese, Swedish (ed.) Melbourne: Language Australia. DiBiase, B. and Kawaguchi, S. (2002) “Exploring the typological plausibility of Processability Theory: language development in Italian second language and Japanese second language”, Second Language Research, 18(3), pp. 274-302. Dulay, H. and Burt, M. (1974) “Natural sequences in child second language acquisition”, Language Learning 24: pp. 37-53. Fayol, M. (2003) “L’acquisition/apprentissage de la morphologie du nombre. Bilan et perspectives” in Perspectives : Rééducation Orthographique – No 213 – mars 2003. Glahn, E., Håkansson, G., Hammarberg, B., Holmen, A., Hvenekilde, A. and Lund, K. (2001) “Processability in Scandinavian Second Language Acquisition”, Studies in Second Language Acquisition 23 : pp 389-416. Goldschneider, J.M. and DeKeyser, R.M. (2001) “Explaining the ‘Natural Order of L2 Morpheme Acquisition’ in English: A Meta-analysis of Multiple Determinants”, Language Learning 51 (1): 1-50. Granfeldt, J. (2007, submitted) "Speaking and Writing in French L2: Exploring Effects on Fluency, Complexity and Accuracy" in Proceedings of the conference on Complexity, Accuracy and Fluency in Second Language Use, Learning and Teaching. Brussels March 2930, 2007. Granfeldt, J. and Schlyter, S. (2003) “Clitisation in the acquisition of French as L1 and L2”, in P. Prévost and J. Paradis (eds) The acquisition of French in different learners: Focus on Functional Categories, Amsterdam/Philadelphia: John Benjamins. Hammarberg, B. (1996) “Examining the Processability Hierarchy: The case of adjective agreement in L2 Swedish”, in E. Kellerman, B. Weltens and T. Bongaerts (eds): Eurosla 6: a selection of papers. Toegepaste Taalwetenschap in artikelen, 55/2: 75-88. Hinkel, E. (2001) Second Language Writers’ Text. Linguistic and Rhetorical Features, Mahwah NJ/London: Lawrence Erlbaum Associates Publishers. Håkansson, G., Pienemann, M. and Sayehli, S. (2002) “Transfer and typological proximity in the context of second language processing”, Second Language Research, 18(3): pp. 250-273. Håkansson, G. and Norrby, C. (2007) “Processability Theory applied to written and oral L2 Swedish”, in F. Mansouri (ed.) Second language acquisition research: Theory-construction and testing, Cambridge: Cambridge Scholars Press UK, pp. 81-94.

Malin Ågren

151

Jaffré, J. P. and David, J. (1999) “Le nombre : Essai d’analyse génétique”, Langue Française (124) : pp. 7-23. Kaplan, R. and Bresnan, J. (1982) “Lexical Functional Grammar: A formal system for grammatical representation”, in J. Bresnan (ed) The Mental Representation of Grammatical Relations, Cambridge MA: The MIT Press, pp. 173-281. Kayne, R.S. (1975) French Syntax. Cambridge, Mass: MIT Press. Levelt, W.J.M (1989) Speaking. From intention to articulation. Cambridge, MA: The MIT Press. Mansouri, F. (2005) “Agreement morphology in Arabic as a second language : Typological features and their processing implications” in M. Pienemann (ed) Cross-Linguistic Aspects of Processability Theory, Amsterdam/Philadelphis: John Benjamins Publishing Company. New, B. (2004) “Le traitement des mots singuliers et pluriels en français et en anglais”, Fondation Fyssen, no 19, pp. 131-138. Pallotti, G. (2007) “An operational definition of the emergence criterion”, Applied Linguistics 28/3, pp. 361-382. Pica, T. (1984) “Methods of morpheme quantification: their effect on the interpretation of second language data”, Studies in Second Language Acquisition 6/1, pp. 69-78. Pienemann, M. (1998) Language Processing and Second Language Development – Processability Theory, John Benjamins Publishing Company: Amsterdam/Philadelphia. Pienemann, M. and Håkansson, G. (1999) “A unified approach towards the development of Swedish as L2: A processability approach”, Studies in Second Language Acquisition, 21, pp. 383-420. Pienemann, M. (2005a) Cross-Linguistic Aspects of Processability Theory (eds) Amsterdam/Philadelphia: John Benjamins. Pienemann, M. (2005b) “An introduction to Processability Theory”, in M. Pienemann (ed.) Cross-Linguistic Aspects of Processability Theory, pp. 1-60. Prodeau, M. (2005) “Gender and Number in French L2: Can we find out more about the constraints on production in L2?” in J.M. Dewaele (ed.) Focus on French as a Foreign Language: Multidisciplinary Approches, Clevedon/Buffalo/Toronto: Multilingual Matters, pp. 135-163. Rahkonen, M. and Håkansson, G. (2007, in press) “Production of written L2 Swedish – Processability or Input Frequencies?” in J. Kessler (ed.) Proceedings of the Zhang, Y. (2005) “Processing and formal instruction in the L2 acquisition of five Chinese grammatical morphemes”, in M. Pienemann (ed.) CrossLinguistic Aspects of Processability Theory, pp 155-177.

POSTVERBAL SUBJECT IN ITALIAN L2 – A PROCESSABILITY THEORY APPROACH CAMILLA BETTONI BRUNO DI BIASE ELENA NUZZO 1. Introduction In any language, for a variety of pragmatic reasons, the same propositional content can be expressed in different perspectives, which then require a range of structural realisations. In most languages, sentences may vary between active and passive, between affirmative and question forms, etc. Speakers may also choose to place constituents in prominent positions by topicalising or focussing them, or they may choose not to do so. Many of these structural choices are devices for directing the hearer’s attention (Levelt 1989), and contribute to the representation of meaning, making communication more effective. However, how and how often these devices are deployed is very much language-specific. The native speaker uses them effortlessly, but for the L2 learner these choices are limited. That is, they are moderated by the current state of their language processor. Processability Theory (PT from now on) offers a principled and parsimonious account of grammatical development in interlanguage (Pienemann 1998; Pienemann, Di Biase and Kawaguchi 2005). As such, it is one of the nine theories of Second Language Acquisition selected for discussion by VanPatten and Williams (2007) in the most recent introduction to this field. Originally accounting for German L2 and English L2, it is now tested typologically on different languages 1 ,

1

Among them, cf. Swedish (e.g., Håkansson 1997), Chinese (e.g., Zhang 2004), Arabic (e.g., Mansouri 2005), and Japanese (e.g., Kawaguchi 2005).

154

Postverbal subject in Italian L2

including Italian 2 . PT’s importance to us here is crucial: while its original version (Pienemann 1998) focuses on the development of obligatory morphosyntactic structures, which are acquired following an implicational pattern, its 2005 extension (Pienemann, Di Biase and Kawaguchi 2005) deals with how learners develop choices at the structural-pragmatic interface. These too are anchored on the structural development of the learner, and are hierarchically ordered: from canonical to noncanonical, from least marked to more marked choices. As a pro-drop language, Italian is syntactically highly sensitive to pragmatic and discourse choices. For instance, the choice between the sequence NP-V and V-NP depends on whether the NP is topical or focal (Van Valin and LaPolla 1997: 418). Hence, Italian is an ideal language to test the new PT hypotheses. In this paper, we analyse postverbal subject realisation in the speech of two advanced learners of Italian L2. Because Italian canonical word order is SVO, VS represents a syntactically marked structure. However, with some verbs, this syntactic markedness may represent the unmarked pragmatic choice. Our hypothesis here is that postverbal subjects will be acquired at different times with different types of verbs and in different structures. We will therefore, first, give an overview of Italian VS order (that is, what the learner needs to know), and illustrate how PT deals with it (that is, how the learner is supposed to learn it). Then, we look at how our two learners actually deploy this structure which is crucially meaningful for efficient communication.

2. VS in Italian As a non-configurational language, Italian enjoys a free word order where a postverbal subject is not a rare occurrence. In principle, any lexical verb can take it, provided there is a specific discourse-pragmatic environment requiring the subject in focus position. In practice, however, some verbs entail it more than others. This fact has received a fair amount of attention, and various explanations have been proposed. One of the first proposals is Burzio’s (1986), who within a generativist approach relates postverbal subjects with unaccusative verbs 3 , maintaining that with these verbs what appears as a subject on the surface is in fact an object in the 2

Cf. Di Biase (2002, 2006); Di Biase and Kawaguchi (2002); Bettoni and Di Biase (2005, 2008); Di Biase and Bettoni (2006); Bettoni, Di Biase and Ferraris (2008). 3 The term refers to the distinction between unaccusatives and unergatives proposed by Perlmutter (1978) and developed by Burzio (1986) with slightly different labelling.

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

155

deep structure, and as such occupies the canonical position of this constituent. Rather than invoking a single factor, other studies (e.g., Sornicola 1994, 1995; Bernini 1995; Cennamo 1995; Benincà, Salvi and Frison 2001 [1988]) maintain that SV and VS alternation can only be accounted for by several factors working together. The various explanations proposed seem to identify two main types of postverbal constructions: one which is motivated by semantic-syntactic factors inherent in the verb, and the other which is triggered by pragmatic choices determined by the context. While we do not intend to elaborate this matter theorically, we try to account for VS in terms of Lexical Functional Grammar (LFG from now on – the grammar theoretical framework for PT), and thus to handle it in terms of lexical features interacting with constituent order. According to LFG, all lexical entries are stored in the mental lexicon with their specific annotations. This position is compatible with Levelt’s language generation model (Levelt 1989; Levelt, Roelofs and Meyer 1999), according to which the features of each lemma are distributed in a three-tier system: the first tier is conceptual-semantic, the second grammatical, and the third phono-morphological. The bundle of grammatical features places the lemma in its syntactic frame. This means that the lexical entry of each verb should contain also information about its distribution in the information packaging of the clause. Now, assuming with Van Valin (2005: 75) that, within the core of the clause 4 , in Italian declaratives the focus 5 is restricted to postverbal position, it follows that postverbal subjects tend to have a focal function 6 (and preverbal subjects tend to be topical). Because some Italian verbs are intrinsically marked as requiring a focal NP, in the case of verbs subcategorising a single argument, this can only be the subject. Thus some Italian verbs have VS as their canonical order: typically, the unaccusatives (1), but also unergatives (2).

4

The core is the unit of the clause which contains the predicate and its arguments (Van Valin 2005: 4). 5 By focus, we mean here both Lambrecht’s (2000: 612) focus (“that element of a pragmatically structured proposition whose occurrence makes it possible for the sentence to express a ‘pragmatic assertion’, i.e. to convey new information to an addressee”), and Bresnan’s (2001), a discoursive nonargumental syntactic function of the sentence. 6 We speak of a tendency here because, as we shall see below with example (11), Italian can use a postverbal subject as topic confirmation.

156

Postverbal subject in Italian L2 (1) [what has happened?] è arrivato l’ambasciatore [has arrived the ambassador] (2) [what has happened?] ha telefonato la nonna [has phoned grandma]

Of course, this canonicity annotated in the lexicon can be violated for pragmatic reasons, so that a verb which canonically does not require a focal NP, like piangere in (3), may have one if the speaker wants to put a constituent in focus (4). (3) [what’s happening?] Pierino piange [Pierino is crying] (4) [are they all crying?] piange solo Pierino [is crying only Pierino]

Likewise, a speaker may, for contrastive discourse or pragmatic choice, use a preverbal subject with a single-argument verb canonically requiring a focus, like telefonare, as we can see comparing (5) with (2) above. (5) [has grandma written?] (ha scritto la zia,) la nonna ha telefonato [(has written auntie,) grandma has phoned]

There seem to be then two conflicting levels of canonicity in Italian, as far as subject position is concerned. On the one hand, we have the general canonical word order SVO in the c-structure. On the other, we have a canonical VS order with those single-argument verbs which require a focal NP. How do learners of Italian L2 deal with this ‘incoherent’ system? And how do they learn to express not only different canonicities, but also the noncanonicity entailed by pragmatic choices? While not all these questions will be answered here, we will try to answer some of them. With transitive verbs, the postverbal position of the focal subject is always a question of pragmatic choice. Compare the difference between (6) and (7), where the SVO sentence in (6) enjoys a much wider usability than the VS one in (7):

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

157

(6) [what has happened?] il vigile ha preso il ladro [the policeman has caught the thief] (7) [who’s caught the thief?] il ladro lo ha preso il vigile [the thief him-ACC.MASC/SING has caught the policeman]

Handling VS with transitive verbs is more difficult for learners. Since there are two arguments within the clausal core, the pragmatic choice of placing the subject in focal position requires not only violating the canonical SV order, but also dealing with the second argument NP. This requires deploying further processing resources, because when speakers want to produce VS, they cannot signal the focal subject by position only. This is most vividly shown by comparing (6) with (8), where both NPs refer semantically to human entities, either of which could plausibly assume the role of agent. If the referential meaning of both sentences were to remain constant, and nothing were to signal the grammatical function of the constituents in the pragmatically marked sentence (8), the listener would understand that it is the thief who caught the policeman: (8) * il ladro ha preso il vigile * [the thief has caught the policeman]

Hence, besides linear position, Italian can – and indeed must – mark the focality of the subject by other means 7 . These are either morpho-syntactic or prosodic devices, or both, or even all three devices interacting in a complex way 8 , according to the function and position of the object. When the object is topical, Italian marks this function of the preverbal NP morphologically by an obligatory coreferential clitic pronoun attached to the verb – as in the left dislocation construction in (7). This extra morphological burden of the co-referential clitic 9 must be present also 7

Across languages, Levelt (1989) identifies three devices for attributing prominence to a participant in an eventuality: syntactic function, linear position, and prosody. 8 Without labouring these complexities, suffice it to say here that Italian prefers to keep the grammatical function rather than use a passive verb, for example, as English would in similar cases. 9 Clitic pronouns are notoriously difficult to acquire, as they are formally complex and similar to articles, as well as unstressed and difficult to discern in the input. Cf., e.g., Berretta (1990: 192, 1992: 139), and Chini and Ferraris (2003: 62) per L2 acquisition, and Caselli et al. (1993), and Leonard (2000) for L1 acquisition.

158

Postverbal subject in Italian L2

when the topic NP is not referentially expressed. Because the argument in postverbal position is focal, the one in preverbal position tends to be topical. As such, it tends to be found easily in the context, linguistic or otherwise, as part of the pragmatic presupposition (Lambrecht 1994: 52), and hence expressed in a weaker form, as a clitic within the sentence, as in (9): (9) [who caught the thief?] lo ha preso il vigile [him-ACC.MASC/SING has caught the policeman]

Thus, when a complement is cliticised, the subject is usually found after the verb because it is pragmatically new, as remarked by Benincà, Salvi and Frison (2001: 139 [1988]). When, as well as the subject, also the object is postverbal, there are two possibilities: VSO and VOS, according to which of the two is focal and which confirmation topic 10 . First, if the focal subject is also contrastive, and the object NP is topical, added at the end to ensure that the listener does understand the participant as topic, this is prosodically weak, and the object clitic is optional, as in (10) 11 : (10) (li) lava MARIO i piatti (non Rosa) [(them-ACC.MASC/PL) washes MARIO the dishes (not Rosa)]

The second possibility is when the object is the contrastive focus, and the subject is not focal (like in all other constructions considered in this paper), but rather a confirmation topic added to prevent a potential ambiguity caused by prodrop. In this case, the object NP is prosodically strong, and the clitic is agrammatical because the object occupies its canonical position, as in (11): (11) (*li) lava I PIATTI Mario (non le pentole) [(*them-ACC.MASC/PL) washes THE DISHES Mario (not the pots and pans)]

Likewise, also intransitive verbs subcategorising a second argument can behave like transitive ones, in the sense of favouring, in pragmatically 10 We will not discuss here the formal grammar-theoretical consequences of this (rather common) ordering of constituents. 11 Notice, however, that if in (9) a topic confirmation were added (lo ha preso IL VIGILE il ladro), the more likely ambiguity between the two argument roles (both referring to human participants) would strongly recommend the use of the clitic.

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

159

marked contexts, clitic or prosodic marking to assign the function to their constituents 12 . Compare the pragmatically marked VS order in (13)-(14) with the more neutral SV in (12). (12) [where does your father’s family live?] mio zio abita a Roma, i nonni a Napoli [my uncle lives in Rome, my grandparents in Naples] (13) [why are you going to Rome?] (a Roma) ci abita mio zio [(in Rome) there lives my uncle] (14) a Roma (ci) abita MIO ZIO (non mio nonno) [in Rome (there) lives in MY UNCLE (not my granpa)]

Finally, in Italian, postverbal subject occurs most frequently with the kind of verbs Pinker (1984) calls ‘exceptional’ – typically among them, psychological verbs such as piacere, divertire, sembrare, interessare. These are intransitive verbs subcategorising two arguments, and as such presenting formal and pragmatic choices similar to the transitive and the two-argument intransitive verbs we have just seen. But they behave ‘exceptionally’ in two crucial ways. First, their pragmatically unmarked structure is not SV, as it is with prendere in (6) or abitare in (12), but rather VS as in (15)-(16): (15) a Maria piace la cioccolata [to Mary is pleasing chocolate = Mary likes chocolate] (16) le interessano le prime tre pagine [to her interest the first three pages = she’s interested in the first three pages]

Secondly, the mapping of their semantic roles onto their grammatical functions is not direct, as the ‘experiencer’ (the highest role available) is not mapped onto the subject (cf. below § 3). Here too, as we have seen in (8) above, should the two NPs be functionally marked only by position, serious misunderstanding may occur. If, for instance, both referents are 12

Strickly speaking, when the second argument is left dislocated, there is a slight difference between transitive verbs and two-argument intransitive verbs, in the sense that, while with the former verbs the clitic is obligatory, with the latter this requirement is less categorical in the higher registers. However, since in the everyday spoken language it is rarely omitted, we do not make this distinction here. Cf. also Benincà, Salvi and Frison (2001: 189 [1988]).

160

Postverbal subject in Italian L2

animate entities, either could plausibly assume the role of the experiencer or the stimulus. This ambiguity is resolved by marking the experiencer as indirect object, either with a preposition if it is expressed referentially by an NP, as in (15), or with a dative clitic pronoun if it is retrievable in the context, as in (16). To sum up, we have identified different uses of postverbal subject in Italian with four types of verbs: x one-argument verbs lexically requiring a focus (type 1 verbs); x one-argument verbs not requiring a focus (type 2 verbs); x two-argument verbs, with either argument in contrastive focus (type 3 verbs); x ‘exceptional’ verbs (type 4 verbs).

3. PT’s predictions According to LFG, the main problem of a syntactic theory is to “characterise the mapping between semantic predicate-argument relationships and surface word- and phrase-configurations by which they are expressed” (Kaplan and Bresnan 1982: 174). Thus grammatical formalism is essentially based on the correspondence with which a sentence maps onto one another their three parallel structures: argumentstructure, functional-structure and constituent-structure. Because each of these structures has its own hierarchy 13 , the mapping between them can be aligned in more than one way (Sells 2001). Pienemann, Di Biase and Kawaguchi (2005) call this alignment unmarked and marked, according to whether, respectively, there is or there is not correspondence among all the highest ranking elements. That is, the alignment is unmarked (or optimal) when the agent is mapped directly onto the subject, which in turn occupies the first position in the linear structure, as in (17): (17) Unmarked alignment: “il portiere prende Charlot” [the janitor catches Charlot]

13 Most briefly, Agent > Beneficiary > Esperiencer/Goal > Instrument > Patient/Theme > Locative for a-structure (Bresnan 2001: 307; Jackendoff 1972; Keenan and Comrie 1977; Hopper and Thompson 1980), Foley and Van Valin (1984), Givón (1984); SUBJ, OBJ, OBLI, for f-structure; and 1st, 2nd, etc. position for c-structure.

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo prendere agent Ļ SUBJ Ĺ il portiere

patient Ļ OBJ Ĺ Charlot

161

a-structure (thematic roles) Ļ f-structure (grammatical functions) Ĺ c-structure (word order)

The alignment is marked when the correspondence fails in either of two ways: between the a- and f-structures; or between the c- and fstructures. The marked alignment of a- to f-structure is typical of the verbs Pinker (1984) calls ‘exceptional’, like piacere (to please) or interessare (to interest), where the experiencer is not the subject, as we have seen in (15)(16). The marked alignment of c- to f-structure occurs when the subject is not in the most prominent (first) position in the linear structure, as in (2), (4), (7), (9)-(11) and (15)-(16) above. Notice too, in (18), that with a verb such as piacere there is markedness on both accounts, as the experiencer is not the subject, and the first position is not occupied by the subject. (18) Marked alignment: a Maria piace la cioccolata [to Mary is pleasing chocolate = Mary likes chocolate] piace experiencer Ļ OBJi Ĺ a Maria

theme Ļ SUBJ Ĺ la cioccolata

a-structure (thematic roles) Ļ f-structure (grammatical functions) Ĺ c-structure (word order)

The sort of correspondence obtaining between the three structures is crucial for acquisition, and PT predicts that the development of L2 syntax depends on the alignment in their mapping (Pienemann, Di Biase and Kawaguchi 2005). More specifically, PT proposes two hypotheses: the Lexical Mapping Hypothesis for learning to map verb arguments onto grammatical functions; and the Topic Hypothesis for learning to map constituents in specific linear positions onto syntactic functions in a pragmatically appropriate way. Like with morphology in Pienemann (1998), also with syntax here, it is a question of cognitive processability. The least marked the alignment, the less costly the processing, the earlier the learning. On the other hand, a more marked alignment entails higher processing cost and later learning.

162

Postverbal subject in Italian L2

Without going into details here, we briefly account for VS acquisition in PT terms. After the initial stage when learners use single words and formulae, like other approaches explaining language acquisition, PT predicts ‘canonical order’ as the unmarked initial syntactic alignment, since learners can map the conceptual structure directly onto linguistic form (e.g., agent in first position). This is computationally the least costly means of organising syntax, as learners do not differentiate between discourse functions such as topic and subject: to them the topic is the subject. After this stage, which PT accounts for with its Unmarked Alignment Hypothesis 14 , now the Topic Hypothesis accounts for learning to differentiate discourse functions from argument functions, most typically the topic from the subject 15 . If at the earlier stage the topic (i.e. the most available entity in the conceptualiser) is the agent in the event, encoded as the subject and placed in initial position, now the addition of something else in first position – typically an adjunct (XP + Canonical order), expressing contextual information (time, place of the event, etc.) – triggers a dislocation of the subject from its canonical first position. When the topic is an adjunct, this addition of an initial element does not affect the canonical order internally. But further on, the processing cost increases, when learners start associating a discourse function (topic) to a grammatical function other than the subject, typically the object; or conversely, associating a focal function to the subject. So, if for pragmatic or discourse reasons, speakers want to direct the listeners’ attention or avoid miscommunication, they can choose the most appropriate word order they need. That is, in Italian they can put the subject in focal position or the object in topical. Specifically in this paper, taking into account what learners need to learn and when they are supposed to do it, our hypothesis is that the postverbal subject position will be acquired at this latter advanced stage in the development identified by the PT extension, and that this will happen gradually. Different types of verbs will favour VS in the following sequence: x first, single-argument verbs which annotate the focus lexically (type 1 verbs). Here, learning to disentangle the focus from the 14

This initial alignment may vary in a language-specific way. For instance, while Italian, like English, is SVO, Japanese is SOV (Kawaguchi 2002) and Catalan VOS (Juan-Garau and Perez-Vidal 2000). 15 In LFG, SUBJ is both a discourse and an argument function (Bresnan 2001).

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

163

object is easiest. In order to apply VS correctly with these verbs, the learner must know that: 1) the subject is not necessarily placed before the verb, 2) some verbs favour VS canonically; x then, single-argument verbs not annotating the focus lexically, whose canonical order is SV with the subject as topic (type 2 verbs). Here the focal subject is a discourse-pragmatic choice. In order to apply VS correctly with these verbs, having learned (1)-(2) as above, the learner must also know that: 3) if the subject is to express new information it must follow the verb; x next, all transitive and some two-argument intransitive verbs, whose canonical order is SVO, with the subject as topic (type 3 verbs). Here too, as with single-argument verbs not annotating the focus lexically, the postverbal subject is a discourse-pragmatic choice. But processing difficulties increase because also the object must be placed in its appropriate position, which can be either before or after the verb: ƒ if the object, as well as the subject, is postverbal, the clitic is either optional in the case of VSO, when the subject is prosodically strong, and the object is a topic confirmation; or agrammatical in the case of VOS, when the object is prosodically strong, and the subject is a topic confirmation (in either case, type 3a structure); ƒ if the second argument is not lexically realised in the clause but retrievable in the context, the clitic is always obligatory; on the other hand, if the second argument is lexically realised and preverbal, the clitic is obligatory when co-referential to a direct object complement, less strictly so but most often found in the spoken language when co-referential to a locative or an indirect object complement (in either case 16 , type 3b structure). In order to apply VS correctly with these two-argument verbs, having learned (1)-(3) as above, the learner must know also that: 4) when non-new information is not supplied by the subject, it cannot be left understood but must be expressed, either by prosody, or most often by a clitic, sometimes by both;

16

Cf. note 11 above.

164

Postverbal subject in Italian L2

x finally, exceptional verbs (type 4 verbs). Here complex processing involves not only marked alignment of c-structure onto f-structure, but also noncanonical mapping of thematic roles onto syntactic functions. In order to apply VS correctly with these verbs, the learner must know, beside (1)-(4) above, also that: 5) the experiencer must be mapped as indirect object and occupy the canonical preverbal position of the subject.

4. The study Our two learners of Italian L2 are Catherine from Ghana, aged 20, and Pandita from India, aged 17 – these names, of course, are pseudonyms. Their L1s are Twi and Punjabi respectively. They both live with their families around Modena in Northern Italy, and attend a technical school, with reasonably good results, so good that they are thinking of going on to university. Both learned Italian untutored, except for some formal help given by the school, especially at the beginning. For comparative reasons, crucial in the case of optional linguistic structures, a native speaker of Italian, Elisa, is also included in the study. She is roughly the same age as the two learners, and attends the same school. The data analysed in this paper belong to a large corpus gathered by Camilla Bettoni and Gabriele Pallotti 17 . They consist of a total of 2,385 clauses, 874 by Catherine, 1,256 by Pandita, and 255 by Elisa, collected by means of similar communicative tasks performed at three different points in time by the learners, and once by the native speaker. The two tasks are a film retelling (episodes from Chaplin’s Modern Times and the Pink Panther series), and an interview during which the girls talk freely about familiar topics, such as their relatives, histories of migration, schooling, friends and hobbies. The time interval between t1, t2 and t3 is one year in each case. The clauses considered here all have a lexical verb as a predicate, and are simple, active, and declarative. In other words, we leave out copulative and presentative clauses, where topic and focus are hardly distinguishable, and in beginning to address this thorny topic, we also bracket out further complications entailed by ‘non-basic forms’ such as interrogative, imperative, and subordinate clauses. 17

Data were gathered with the help of two grants: COFIN 2003, coordinated by C. Bettoni, and co-financed by the Italian Ministry of University and Research and the University of Verona; and COFIN 2006, coordinated by G. Pallotti, and cofinanced by the Ministry and the University of Modena and Reggio Emilia.

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

165

5. Results Table 1 gives a general overview of the learners’ syntactic development, and compares it to the performance by the native speaker. As befits their advanced learning stage, at t1 both Catherine and Pandita can already produce VS structures in numbers that are altogether not unlike those of the native speaker Elisa – given the general freedom of choice for the occurrence of pragmatically motivated structures. Catherine t1

t2

Pandita

t3

t1

t2

Elisa t3

t1

n. of clauses*

299 233 342 328 369 559

255

V SUBJ°

19

4

19

29

20

36

14

ADJ SUBJ V

1

2

3

1

5

9

0

SUBJ V

47

45

55

72

109 169

53

pro drop

233 184 268 227 239 354

187

* Numbers in this row are lower than the sum of all categories in the rows below, because SUBJ V includes ADJ SUBJ V. ° Numbers include some formulas.

Table 1. Syntactic development Let us now look in table 2 at the distribution of VS structures in the period under observation. VS word order is altogether most frequently used with single-argument verbs annotating the focus lexically (53 times by the learners). This structure is already firmly in place with a robust number of occurrences in both of them at t1. The most common such verbs are venire (to come), as in (19), and arrivare (to arrive), but we find also morire (to die), as in (20), or succedere (to happen) among several others. (19) poi viene un polisiotto (Pan, t1) [then comes a policeman] (20) subito è morto mio padre (Cat, t1) [soon died my father]

Postverbal subject in Italian L2

166

Catherine

Total*

Pandita

Elisa

t1

t2

t3

t1

t2

t3

t1

16

3

19

27

19

32

12

0/2

2/2

6/8

0/1

2/6

1/9

2/7 1/10

1

2

4

1

1

1

11

2

type 4 verbs° type 3 verbs, with obligatory clitic type 3 verbs, with optional/no clitic

1

type 2 verbs

3

type 1 verbs

12

1

9

4/4

4

1

18

7

* Total numbers here are slightly lower than in table 3 because formulas are excluded. ° The number after the slash gives the context for the clitic (or alternatively with type 4 verbs for either the clitic or the preposition marking the NP as indirect object). The number before the slash shows the times the clitic is actually supplied.

Table 2. VS structures VS constructions with type 2 verbs, that is, with single-argument verbs not annotating the focus lexically, are far less common (overall 5 occurrences) – not surprising as they are pragmatically marked. Since they seem already in place at t1, there is little comment to add about their development from t1 to t3. Among these verbs, however, the range is relatively wide, as we find mangiare (to eat), as in (21), piangere (to cry), as in (22), sedere (to sit), and twice parlare (to speak). (21) mangiamo tutti in cucina (Cat, t1) [eat=1st PERS.PL. all in the kitchen] (22) piangevano tutti (Pan, t1) [were=3st PERS.PL. crying all]

The same can be said about type 3a constructions, those with twoargument verbs not requiring cliticisation. They too are few (altogether 13), yet diverse, not only lexically (avere, fare, rubare, pensare, andare)

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

167

but also structurally: that is, with postverbal topic confirmation provided both by the object, as in (23), and by the subject, as in (24). Like VS constructions with type 1 and type 2 verbs, these too are handled correctly since t1: (23) ho rubato io il pane (Pan, t1) [have stolen I the bread] (24) non sono andata a Brescia io (Cat, t1) [not have gone to Brescia I]

VS constructions with type 3 verbs in type b structures – that is, those categorically requiring cliticisation, or strongly favouring it in the spoken register – are less rare occurrences (there altogether 33), except that Catherine and Pandita rarely supply the clitic (overall, only 6 times). Here, we may discern some development over the two-year period. At the beginning, Catherine provides no contexts for cliticisation. Then at t2 there is one context, in which however even the ‘easier’ locative clitic ci, requiring no gender and number specification, is missing. Finally in t3, she supplies two clitics correctly out of 6 needed. These are a locative ci, and another most common one: the first person mi, which also requires no gender specification: (25) non è tanto bello ma ci abitano le persone (Cat, t3) [not is very beautiful but there live people] (26) mi aiutava il mio compagno (Cat, t3) [me helped my friend]

Pandita is more adventurous in supplying contexts from the very beginning, but then her clitics are equally few, at great risk for communication. We can see this clearly in (27), where the clitic is missing: (27) [l’]ha già preso il polisiotto (Pan, t1) [[him=Charlot] has already caught the policeman]

In t2 and t3, like Catherine’s, also Pandita’s are first person clitics or a semiformulaic lo used with the verb sapere: (28) hindi, lo sanno tutti (Pan, t2) [hindi, it know all]

168

Postverbal subject in Italian L2

Among the exceptional verbs, piacere is generally by far the most frequently used. Here too occurrences are rare, yet we can discern minimal development. In t1 and t2 Catherine does not use any of them. Then, when she starts using them, she seems to map the experiencer as subject in preverbal position, once with a full referential NP, as in (29) where she omits the preposition, the other with pro-drop, as in (30) where she omits the dative clitic pronoun: (29) invece [a] mia sorella piace l- e:hm tch televisione (Cat, t3) [instead [to] my sister pleases television] (30) non [gli] interessava niente (Cat, t3) [not [to him] interested anything]

More questionable is whether also Pandita starts off by mapping the experiencer onto the subject, because already in t1 her verb correctly agrees with the postverbal subject, so that the experiencer would appear to be an indirect object: (31) non mi piace questa scola (Pan, t1) [not to me pleases this school = I don’t like this school] (32) mi piacciono questi orecchini questi (Pan, t1) [to me please these earrings these = I like these earrings]

However, in t2 all her 6 correct occurrences also involve the ‘easier’, most frequent and perhaps formulaic, first person singular mi, while the two incorrect occurrences suggest direct mapping of the experiencer as agent in first position, especially in (34) where not only the preposition is missing but also the verb agrees with the preverbal NP rather than with the postverbal one: (33) poi, [a] loro serve una ragazza (Pan, t2) [then, [to] ?they/them is needed a girl] (34) [a] lei non piaceva[no]. quei a:lti scoli (Pan, t2) [[to] ?she/her not pleased those high schools]

To summarise our findings, it would seem that there is support for our developmental hypothesis regarding verb and structure types, despite few occurrences of some VS constructions. First, Catherine’s data suggest that the progression is indeed from inaccusative and unergative verbs in t1, to two-argument verbs in t2, and ‘exceptional verbs’ in t3. Secondly, both

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

169

learners’ data clearly indicate that verbs and structures that do not require the clitic object marker on the verb are learned before those that do. For those needing a clitic, first a context for the clitic is created, then some formulaic and less inflected clitics are produced, while more time is needed for learning their full range (e.g., the agreement in gender and number with the topic NP). Thus, a crucial factor in the sequence seems to be the acquisition of the relevant morphology, in the form of the clitics.

6. Conclusion Our data may not be rich in terms of development, partly because the learners we have observed are too advanced, partly because, despite our wide corpus, the occurrences of the observed features are relatively scarce. This is not surprising, since in an SVO language it is the object that normally carries new information. And learners can go a long way with this canonical word order only. However, when the subject is focal it carries the main information load of the sentence, and learners must be able to signal its function unambigously in order to realise a fuller range of expressive perspectives and to avoid miscommunication. Our data nevertheless can highlight some interesting developmental phenomena. At the beginning of the observation period, postverbal subject is already in place with intransitive verbs which subcategorise only one argument and annotate the focus lexically. With this type of predicates both learners seem to master perfectly the SV inversion. This might reflect the fact not only that with such predicates VS occurs most frequently, but also that placing the subject in the focal object position is easier for the learners if they do not need to handle any additional argument(s). Later on, when topical information is not encoded by the subject but by a cliticized complement, learners seem able to realise the subject in focal position, but unable to supply the clitic pronoun. They thus try to create the syntactic environment required by the information structure of the sentence but are yet unable to fully grammaticalise it according to the L2 rules. Then full grammaticalisation happens first with less inflected clitics, such as locatives and first person ones. ‘Exceptional’ verbs create most difficulties for learners, as they entail marked alignment between both c- and fstructures and a- and f-structures. Looking at their realisations, we notice a tendency to formulaic use with the most common first person, and to a least marked correspondence between the three structures, whereby the role of the experiencer is mapped directly onto the subject, which also occupies the most prominent position in the sentence. Finally, we wish to mention two further points. The first is pedagogical:

170

Postverbal subject in Italian L2

let us stress how slow the progress is for the spontaneous acquisition of the proper grammatical expression of marked pragmatic choices, even for bright young women learning Italian as an L2 while regularly attending high school in Italy. It would seem, then, that here some targeted instructional intervention could be of great assistance in speeding up the rate of learning. The second point is methodological: since VS is an optional construction, if we wish to gain a clearer picture of its sequential emergence with verbs and structures favouring it, more data is required, not only longitudinally starting from lower levels of proficiency than those of our two learners, but also gathered by means of specifically devised tasks readily eliciting the targeted verbs and structures. Such tasks would then be useful on both accounts: pedagogically, and methodologically.

7. Bibliography Benincà, P., G. Salvi & L. Frison (2001 [1988]): “L’Ordine degli elementi della frase e le costruzioni marcate”, in: L. Renzi, G. Salvi & A. Cardinaletti [eds.] (2001 [1988]): Grande Grammatica Italiana di Consultazione. Bologna: il Mulino, Vol. III, pp. 129-239. Bernini, G. (1995): “Verb-subject order in Italian: An investigation of short announcements and telecast news”, in: Sprachtypologie und Universalien Forschung, Vol. 48; pp. 44-71. Berretta, M. (1990): “Morfologia in italiano lingua seconda”, in: E. Banfi & P. Cordin [eds.] (1990): Storia dell’Italiano e Forme dell’Italianizzazione. Roma: Bulzoni, pp. 181-202. Berretta, M. (1992): “Marcatezza in morfologia e apprendimento di lingue seconde“, in: Quaderni del Dipartimento di Linguistica e Letterature Comparate dell’Università di Bergamo, Vol. 8; pp. 129-56. Bettoni, C. & B. Di Biase (2005): “Sviluppo obbligato e progresso morfosintattico: un caso di Processabilità in italiano L2”, in: ITALS. Didattica e Linguistica dell’Italiano come Lingua Straniera, Vol. III/7; pp. 27-48. Bettoni, C. & B. Di Biase (2008): “Lessico verbale e questioni di Processabilità in italiano L2”, in: M. Barni, D. Troncarelli & C: Bagna [eds.] (2008): Lessico e Apprendimenti. Perugia: Guerra Edizioni, pp. 260-267. Bettoni, C., B. Di Biase & S. Ferraris (2008): “Sviluppo sintattico e sviluppo morfologico: ipotesi di corrispondenze nella Processabilità dell’italiano L2”, in: G. Bernini, L. Spreafico & A. Valentini [eds.] (2008): Competenze Lessicali e Discorsive nell’Acquisizione di Lingue Seconde. Perugia: Guerra Edizioni, pp. 355-382.

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

171

Bock, J.K. & R.K. Warren (1985): “Conceptual accessibility and syntactic structure in sentence formulation”, in: Cognition, Vol. 21; pp. 47-67. Bresnan, J. (2001): Lexical-Functional Syntax. Oxford: Blackwell. Burzio, L. (1986): Italian Syntax: A Government-Binding Approach. Dordrecht: Reidel. Caselli, C.M., L.B. Leonard, V. Volterra & M.G. Campagnoli (1993): “Toward mastery of Italian morphology: a cross-sectional study”, in: Journal of Child Language, Vol. 20; pp. 377-393. Cennamo, M. (1995): “Transitivity and VS order in Italian reflexives”, in: Sprachtypologie und Universalien Forschung, Vol. 48; pp. 84-105. Chini, M. & S. Ferraris (2003): “Morfologia del nome”, in: A. Giacalone Ramat [ed.] (2003): Verso l’Italiano. Percorsi e Strategie di acquisizione. Roma: Carocci, pp. 37-69. Di Biase, B. & C. Bettoni (2006): “Funzioni discorsive e Processabilità in italiano”, in: M. Chini, A. De Meo, P. Desideri & G. Pallotti [eds.] (2006): Imparare una Lingua. Recenti Sviluppi Teorici e Proposte Applicative. Perugia: Guerra Edizioni, pp. 209-233. Di Biase, B. & S. Kawaguchi (2002): “Exploring the typological plausibility of Processabilty Theory: Language development in Italian second language and Japanese second language”, in: Second Language Research, Vol. 18/3; pp. 272-300. Di Biase, B. (2002): “Focusing Strategies in Second Language Development: A classroom-based study of Italian L2 in primary school”, in: B. Di Biase [ed.] (2002): Developing a Second Language: Acquisition, Processing and Pedagogy of Arabic, Chinese, English, Italian, Japanese, Swedish. Melbourne: Language Australia, pp. 95120. Di Biase, B. (2005): “The Topic Hypothesis in Processability Theory”, paper presented at the 5th International Symposium on Processability, Second Language Acquisition and Bilingualism, Deakin University, Melbourne, 26-28 September 2005. Di Biase, B. (2006): “Developmentally moderated focus-on-form”, paper presented at the 5th Pacific Second Language Forum, University of Queensland, Brisbane, Australia, July 4-6, 2006. Falk, Y.N. (2001): Lexical-Functional Grammar: An Introduction to Parallel Constraint-based Syntax. Stanford, CA: CSLI Publication. Foley, W.A. & R.J. Van Valin (1984): Functional Syntax and Universal Grammar. Cambridge: Cambridge University Press. Givón, T. (1984): Syntax: A Functional-Typological Introduction. Amsterdam: Benjamins. Håkansson, G. (1997): “Modern Times in L2 Swedish. On the acquisition

172

Postverbal subject in Italian L2

of Swedish morphology and syntax in formal and informal settings”, in: L. Diaz & C. Pérez [eds.] (1997): Eurosla 7 Proceedings: Views on the Acquisition and Use of a Second Language. Barcelona: Universitat Pompeu Fabra, pp. 39-50. Hopper, J.P. & S.A. Thompson (1980): “Transitivity in grammar and discourse”, in: Language, Vol. 56; pp. 251-299. Jackendoff, R.S. (1972): Semantic Interpretation in Generative Grammar. Cambridge, MA: MIT Press. Juan-Garau, M. & C. Perez-Vidal, (2000): “Subject realization in the syntactic development of a bilingual child”, Bilingualism: Language and Cognition, Vol. 3; pp. 173-191. Kaplan, R. & J. Bresnan (1982): “Lexical-functional grammar: A formal system for gram-matical representation”, in: J. Bresnan [ed.] (1982): The Mental Representation of Grammatical Relations. Cambridge, MA: MIT Press, pp. 173-281. Kawaguchi, S. (2005): “Argument structure and syntactic development in Japanese as a second language”, in: M. Pienemann [ed.] (2005): CrossLinguistic Aspects of Processability Theory. Amsterdam: Benjamins, pp. 253-299. Keenan, E.O. & B. Comrie (1977): “Noun phrase accessability and Universal Grammar”, in: Linguistic Inquiry, Vol. 8; pp. 63-99. Lambrecht, K. (1994): Information Structure and Sentence Form. Cambridge: Cambridge University Press. Lambrecht, K. (2000): “When subjects behave like objects: A markedness analysis of sentence focus constructions across languages”, in: Studies in Language, Vol. 24; pp. 11-82. Leonard, L.B. (2000): “Specific language impairment across languages”, in: D.V.M. Bishp & B.L. Leonard [eds.] (200): Speech and Language Impairments in Children: Causes, Characteristics, Intervention and Outcome. Hove: Psychology Press, pp. 115-129. Levelt, W. J. M., A. Roelofs & A.S. Meyer (1999): “A theory of lexical access in speech production”, in: Behavioral and Brain Sciences, Vol. 22; pp. 1-75. Levelt, W.M. (1989): Speaking. From Intention to Articulation. Cambridge, MA: MIT Press. Mansouri, F. (2005): “Agreement morphology in Arabic as a second language: Typological features and their processing implications”, in: M. Pienemann [ed.] (2005): Cross-Linguistic Aspects of Processability Theory. Amsterdam: Benjamins, pp. 117-154. Perlmutter, D. (1978): “Impersonal passives and the unaccusative hypothesis”, in: Berkeley Linguistics Society, Vol. 4; pp. 157-89.

Camilla Bettoni / Bruno Di Biase / Elena Nuzzo

173

Pienemann, M. (1998): Language Processing and Second Language Development: Processability Theory. Amsterdam: Benjamins. Pienemann, M., B. Di Biase & S. Kawaguchi (2005): “Extending Processability Theory”, in: M. Pienemann [ed.] (2005): Crosslinguistic Aspects of Processability Theory. Amsterdam: Benjamins, pp. 199-251. Pinker, S. (1984): Language Learnability and Language Development. Cambridge, MA: Harvard University Press. Sells, P. (2001): “Introduction”, in: P. Sells [ed.] (2001): Formal and Empirical Issues in Optimality Theoretical Syntax. Stanford, CA: CSLI Publications, pp. 1-16. Sornicola, R. (1994): “On word order variability: A study from a corpus of Italian”, in: Lingua e Stile, Vol. 29; pp. 25-57. Sornicola, R. (1995): “Theticity, VS order and the interplay of syntax, semantics and pragmatics”, in: Sprachtypologie und Universalien Forschung, Vol. 48; pp. 72-83. Van Valin, R.D. & R.J. La Polla (1997): Syntax, Structure, Meaning and Function. Cambridge: Cambridge University Press. Van Valin, R.D. (2005): Exploring the Syntax-Semantics Interface. Cambridge: Cambridge University Press. VanPatten, B. & J. Williams (2007): “The Nature of Theories”, in: B. VanPatten & J. Williams [eds.] (2007): Theories of Second Language Acquisition. Mahwah, NJ: Lawrence Erlbaum, pp. 1-16. Zhang, Y. (2004): “Processing Constraints, Categorical Analysis, and Second Language Acquisition of the Chinese Adjective Suffix de (ADJ)”, in: Language Learning, Vol. 54; pp. 437-468.

APPLYING PROCESSABILITY THEORY AND ITS EXTENSION TO SERBIAN AS A FAMILY AND COMMUNITY LANGUAGE IN AUSTRALIA

LUCIJA MEDOJEVIû This chapter presents the first application of Processability Theory (PT) (Pienemann 1998a) and its extension (Pienemann, Di Biase and Kawaguchi 2005) as a theoretical framework for the acquisition of Serbian, a highly morphologised Slavonic language, as a family and community language in Australia. A hierarchy of stages for the acquisition of Serbian morphosyntax is firstly hypothesised and formally described using Lexical Functional Grammar (Bresnan 2001). The proposed hierarchy is then used to assess the language processing skills of SerbianAustralian bilingual teenagers through a cross-sectional study involving spoken production by three bilingual informants who acquired Serbian in their home environment. One of the informants, additionally, had formal instruction in the Serbian language. An adult native speaker of Serbian, also residing in Australia, performed the same tasks as a comparison informant. In addition to morphosyntactic development, three hypotheses proposed in the current extension of PT, i.e., the Unmarked Alignment Hypothesis, the Lexical Mapping Hypothesis and the Topic Hypothesis, (Pienemann et al. 2005) are applied to assess the range of pragmaticssyntax choices that teenage bilinguals might exercise given the relatively free word order of Serbian (Hammond 2005). A distributional analysis of the results suggests that all three bilingual informants reached the highest inter-phrasal morphosyntactic stage in the proposed Serbian PT-based hierarchy. However, the acquirer who had additional instruction in Serbian demonstrated higher skills in exercising alternative pragmatics-syntax choices than the other two teenage bilinguals. This study extends PT to Slavonic languages and further supports its typological plausibility. It also shows how the new hypotheses are useful in measuring language

176 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

development under specific circumstances of limited domains of use in Australia 1 .

1. Introduction The objective of this paper is to apply PT and its recent theoretical extension to the investigation of the development of Serbian as a family and community language among Serbian-English bilingual teenagers in Australia. As for other bilingual first language acquisition in Australia, while the context of acquisition and use of English involves practically all linguistically relevant domains, for Serbian the context is limited, to the home and family domain. Serbian is used in communication with family members and visiting family friends and community members. Such limited context for language input and the lack of opportunities for the language output under the influence of the dominant English language affects the ‘rate and ultimate attainment’ (in Doughty and Long 2003 terms) of Serbian language knowledge and skills. The present study thus looks at how far the Serbian language is acquired by Serbian-Australian teenage bilinguals using the predictions of Processability Theory. What makes Serbian an interesting language for PT is its complex morphology and its so called ‘free’ word order (Hammond 2005). In fact, this ‘free’ word order is not at all arbitrary but is motivated by pragmatic factors such as topicality and prominence (Comrie and Corbett 1993). To my knowledge, PT has not been tested on any other Slavonic language previously and it has not been used to investigate special circumstances of language acquisition such as those prevailing in migration. The aim of this paper is to propose a PT-based hierarchy for Serbian, in order to investigate the stage of morphosyntactic development in three Serbian-Australian teenagers. Intuitively, Serbian-Australian teenage bilinguals who use Serbian in communication with their family members should possibly be at an advance stage of language development, so it would be interesting to test for the hypotheses based on the extension of PT proposed by Pienemann, Di Biase and Kawaguchi (2005). Given that these hypotheses address the interface between syntax and discoursepragmatics, they should help assess whether the teenage bilinguals are able to use the so called ‘free’ word order and other non-default 1

I wish to express my special gratitude to Dr Bruno Di Biase for his generous advice in various points of writing this chapter. I also wish to thank Dr Satomi Kawaguchi for her comments and advice. I also express my sincere appreciation to the anonymous reviews who gave me invaluable advice to improve my chapter. Remaining errors are solely my responsibility.

Lucija Medojeviü

177

pragmatically motivated choices, some of which require resources placed at higher levels in the developmental hierarchy. This chapter is organised as follows. In the next section I present a short description of the theoretical framework, i.e., Processability Theory and its current extension. Section 3 offers a brief sketch of Serbian typological characteristics and the grammar relevant to this paper. Section 4 presents the proposed Serbian PT-based hierarchy and the corresponding research question and hypothesis. In section 5 the pragmatic-syntax choices (alternative word order and passive) of Serbian are discussed with the corresponding question and hypothesis. Next (section 6) I will describe the informants, settings and data collection which is followed by the results of the study in conjunction with the hypotheses posed in this study (section 7). The discussion and conclusion of the results is presented in section 8.

2. Theoretical framework Processability Theory (Pienemann 1998a; Pienemann et al. 2005) was chosen as the framework for the present study for several reasons. First, Processability Theory is a universalistic theory of language acquisition, i.e., it is not language-specific, as it focuses on the relation between language processing and language acquisition. The second reason is its typological plausibility tested on distinct languages (Di Biase and Kawaguchi 2002). Thirdly, it has been successfully applied to analyses of first and second language acquisition (e.g., Pienemann 1998b).Finally, the theory has also been used in comparing the development of two first languages of a Japanese-Australian child (Itani-Adams 2007). The study supported de Houwer’s (1990) Separate Development Hypothesis and showed that both languages follow the hierarchy proposed by PT (Pienemann et al. 2005). Processability Theory describes the nature of computational mechanisms in the process of language acquisition and the way in which they are acquired. It hypothesises a universal hierarchy of morphosyntactic development. PT is based on Levelt’s (1989) speech production model and relies on Lexical-Functional Grammar (Kaplan and Bresnan 1982; Bresnan 2001) for the formal description of linguistic structures that are psychologically and typologically plausible. The theory has been widely tested on typologically close languages: Swedish L2 (Håkansson and Pienemann 1999), German L2 (Håkansson, Pienemann and Sayehli 2002)

178 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

and Scandinavian languages (Glahn, Håkansson, Hammarberg, Holmen, Hvenekide and Lund 2001) as well as typologically different languages: Italian (Di Biase 1999, 2002), Japanese (Kawaguchi 2000, 2002, 2005a), Chinese (Zhang 2002, 2004) and Arabic (Mansouri 2002, 2005). This theory assumes a set of key grammatical encoding procedures that are arranged according to their sequence of activation in the language generation process as listed in (1). (1) Clause boundary Word order rules S-procedure Phrase procedures Category procedures Word/ Lemma This hierarchy is implicationally arranged, that is, each lower level procedure in this sequence is a necessary prerequisite for the following procedure. Feature unification explains how speakers acquire morphologically and syntactically correct utterances. Learners will follow this empirical sequence in the process of L2 acquisition. If an element is missing from this sequence the learner grammar will be cut off and replaced by direct mapping of conceptual structures on to surface form (Pienemann 1998a). Work on typologically distant languages such as Japanese and Italian paved the way to the new extension of PT (Pienemann et al.2005). This adds three hypotheses to the PT: the Unmarked Alignment Hypothesis (UAH), Topic Hypothesis (TOP) and the Lexical Mapping Hypothesis (LMH) which deal with the acquisition of discourse-pragmatic functions and so it goes beyond feature unification presented in the original PT hierarchy. These hypotheses are concerned with the L2 acquisition of nondefault relationships between argument structure (a-structure) and grammatical functions (f-structure) and its representation on the constituent structure (c-structure). Default mapping (in English) is illustrated in (2).

179

Lucija Medojeviü (2)

eat < cat, fish> Agent SUBJECT NP subj The cat

Patient

(a-structure)

VERB

OBJECT

(f-structure)

NP obj the fish.

(c-structure)

ate

According to the Unmarked Alignment Hypothesis (UAH) the language acquirer initially preserves canonical word order in the organisation of the syntax in a language. The Topic Hypothesis (TOP) on the other hand captures the beginning of a differentiation of Topic and Subject as discourse functions in SLA. It allows a wider range of syntactic variability and expressiveness predicting the c- to f- structure mapping. For example, the position of the XP adjunct before the canonical word order is the ccorrespondence of the initial differentiation of TOP and SUBJ in English. This structure disrupts the linearity of the canonical order causing the learner not to depend solely on the linear mapping and position of canonical word order. Canonical order is further disrupted at a higher stage when an argument (rather than an adjunct) other than the subject is topicalised, i.e., is given prominence in linear order. The Lexical Mapping Hypothesis predicts how mapping develops from the constraints of the Unmarked Alignment Hypothesis to the nondefault mapping principles of the target language, i.e., a- to f- structure correspondence (Pienemann et al. 2005). This hypothesis assumes that learners gradually attain skills to map less prominent thematic roles (i.e. patient) onto the subject function in structures like passives and causatives. This non-canonical mapping of argument roles onto the grammatical function requires additional processing. The investigation of the acquisition of Serbian in Australian conditions offers a good opportunity to test Serbian’s pragmatically driven “free” word order and its complex and highly irregular case marking system (Hammond 2005) in the language of the teenage bilinguals. The PT extension provides a firm foundation for predictions about the corresponding mechanisms in the development of bilinguals’ language.

180 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

3. Typology of Serbian and a brief sketch of its grammar Serbian is a Slavonic language and can be characterised as a conservative Indo-European language (Comrie and Corbett 1993). A typological characteristic of the Serbian language is its highly fusional morphology. Serbian nouns that fall into masculine, feminine or neuter grammatical gender consist of a stem and inflections that are assigned to it at the lexical level. The inflections indicate the number (singular or plural) and the case on the noun and exhibit form variation. A single inflectional morpheme may be used for several different morphological contrasts which make the case system complex and highly irregular. In English, functional roles are identified through word order while in Serbian, case endings on the nouns identify the subject, object or indirect object (Hammond 2005). There is no one-to-one relationship between form and function. There are seven cases (nominative, accusative, genitive, dative, locative, instrumental and vocative) marked in the language, each expressing a different function for each type of noun, adjective or pronoun (Hammond 2005). The order of major constituents is determined not so much by syntactic factors as in English or German with relatively fixed word order as by pragmatic factors. Constituent order in the sentence is determined largely by topic-comment structure (cf. Levelt 1989). Thus, word order in Serbian is flexible and there are six possible permutations (SVO, SOV, VSO, VOS, OSV and OVS). Orders where the object precedes the subject are rare even though all orders are grammatically acceptable (Uroševiü, Carello, Saviü, Lukatela and Turvey 1986). SVO is the most frequent word order in Serbian and the most dominant. Subject and verb features are systematically related to each other and agree in number and gender. Serbian is richly endowed with regular and mixed agreements (Wechsler and Zlatiü 2000, 2001). Due to the rich inflectional morphology the subject of the sentence can be omitted as the relevant information is retrievable from the verb inflection (Comrie and Corbett 1993). The only element that has a fixed position in the sentence is the clitic. The clitic or clitic group comes in second position in the sentence and displays fixed internal order in the group: interrogative enclitic ‘li’ + pronominal enclitics + verbal enclitics usually AUX + reflexive enclitic ‘se’ (Hammond 2005). Learning the pragmatics behind this free word order and its connection with the clitic group as well as the rich inflectional morphology are the greatest difficulties faced by acquirers of Serbian. The main cue used for sentence interpretation is the casemarking system (Slobin 1973).

181

Lucija Medojeviü

The present study thus focuses on the typological characteristics of Serbian nominal morphology and its syntactic representation. A brief sketch of the nominal morphosyntax is presented below. Nouns in Serbian carry case endings that mark their function in the sentence. Nominal modifiers that precede the head noun obligatorily agree in number, case and gender with the head (Hammond 2005). This is illustrated in Figure 1. The elements of the singular NP ‘this blue dress’ agree in number, gender and case. VP Ĺ=Ļ V

(Ĺ OBJ) = Ļ NP DEM

Peglam

N’ AP

N

A ov-u

plav-u

haljin-u.

(Ĺ PRED) = ‘iron’ < SUBJ, OBJ> (Ĺ TENSE) = PRESENT (Ĺ SUBJ PER) = 1 (Ĺ SUBJ NUM) = SING (Ĺ SPEC) = ‘this’ (Ĺ NUM) = SING (Ĺ GEN) = FEM (Ĺ CASE) = ACC

(Ĺ ADJ PRED) = ‘blue’ (Ĺ NUM) = SING (Ĺ GEN) = FEM (Ĺ CASE) = ACC

(Ĺ PRED) = ‘dress’ (Ĺ NUM) = SING (Ĺ GEN) = FEM (Ĺ CASE) = ACC

Figure1: LFG c-structure representation of singular NP agreement Nouns, adjectives and adjectival pronouns also have a form without case distinction, used when accompanying numerals ‘2’, ‘both’, ‘3’ and ‘4’ (Comrie and Corbett 1993). This form is observed in the quantifier NPs. An example on a quantifier NP is illustrated in Figure 2 below.

182 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

VP Ĺ=Ļ V

(Ĺ OBJ) = Ļ NP DEM

N’ AP

Vidim

N

A tri

velik-a

medved-a

(Ĺ PRED) = ‘see’ < SUBJ, OBJ> (Ĺ TENSE) = PRESEN (Ĺ SUBJ PER) = 1 (Ĺ SUBJ NUM) = SING (Ĺ SPEC) = ‘three’ (Ĺ NUM) = PL

(Ĺ ADJ PRED) = ‘big’ (Ĺ NUM) = PL (Ĺ GEN) = MASC (Ĺ CASE) = GEN

(Ĺ PRED) = ‘bears’ (Ĺ NUM) = PL (Ĺ GEN) = MASC (Ĺ CASE) = GEN

Figure 2: LFG c-structure representation of Quantifier NP agreement Two syntactic structures exist that are expressed by count nouns or mass nouns: adjectival and governmental. Adjectival quantification of NP has case, number and gender of the head noun agreeing with all the elements in the NP (Comrie and Corbett 1993). The governmental quantification however has the quantifier imposing a form on the noun and it is observed with numerals ‘two’, ‘both’, ‘three’ and ‘four’ as illustrated in Figure 2 above. These numerals occur with pluralia tantum nouns and induce genitive case. This type of internal case feature distribution is heterogeneous and is known as genitive of quantification GEN (Q) (Franks 1995). Quantifier NPs in Russian (Franks 1995: 95-97) on the other hand, in an oblique case position, appear entirely in the appropriate case (“homogeneous agreement”). In overt case positions the quantifiers include genitive in the following noun and its modifiers (“heterogeneous agreement”). The Russian quantifier phrase (QP) projection is labelled as

Lucija Medojeviü

183

structural case. Nevertheless, Franks (1995) states that this theory does not extend to Serbian as GEN (Q) is an inherent case in this language. The quantified NPs in Serbian occur in an oblique case position without marking that oblique case overtly (Franks 1995). All modifiers in the quantificational NP must be in the genitive regardless of the position. Thus, this structure is treated as NP in Serbian language. In the NP ‘three big bears’ presented above the numeral three imposes the “234”2 form on to the adjective and noun that agree in number, case and gender. The morphosyntactic structures described above play an important role in the application of PT on Serbian language in this study. Thus, the nominal morphosyntactic stages for Serbian PT-based hierarchy are proposed in the next section.

4. The proposed Serbian PT-based hierarchy and the processability of the structures A Serbian PT-based hierarchy is proposed next to assess SerbianAustralian teenage bilinguals’ language specific-processing routines. The present study focuses on four nominal morphosyntactic structures of Serbian which are only a selection of stages proposed to investigate whether the informants can handle form variation and information exchange between constituents across three PT stages (lexical > phrasal > inter-phrasal). Description of the structures and their status in the proposed hierarchy can be seen in Table 1. This PT derived hierarchy is hypothesised on the basis of information exchange between constituents. It does not predict that whatever can be processed will indeed be acquired but nevertheless one can test whether these features are “developmentally entrenched” (Pienemann 2005). It is assumed that these structural features are maintained throughout the developmental process and underlie other more complex structures that do not have to be decided on every time a refreshment of parts of the structures is made (Pienemann 2005). In other words the hierarchy is defined as a specific range of morphosyntactic structural options of Serbian available to the learner.

2

Comrie & Corbett (1993) refer to this structure as ‘234 form’ as a remnant of the Proto-Slavonic dual number.

184 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

The selected structures presented in Table 1 distribute over the three stages of the Processability hierarchy: lexical> phrasal> interphrasal. The predictions of the hierarchy imply that: 1) Number and case marking on nouns will be acquired before the number and case agreement in NPs; and 2) The number and case agreement in NPs will be acquired before the Subject – predicative adjective agreement. PT stages

Serbian Morphosyntax

4 Inter-phrasal

Subject – predicative adjective agreement

3 Phrasal

NP agreement (Number: Sing /Pl & Case: Nom /Acc)

2 Lexical

Number marking on nouns (Singular /Plural) Case marking on nouns (Nominative/Accusative)

1 Lemma / Word level

-------------------

Table 1: Hypothesised hierarchy for Serbian Each stage or procedure in this hierarchy is discussed separately below. Stage 2 (lexical): LFG is a lexicalist theory where syntactic structure is driven by the lexicon. Word structure is different from phrase and sentence structure as the order of elements in morphology is always fixed (Falk 2001). Lexical and form variation of nouns in terms of number and case is the characteristic of this procedure. Such features characterize one of the major lexical categories, i.e., the nouns. Nominative and accusative cases are assumed to be the default markers for core grammatical relations. Nominative endings of nouns mark the subject function. The recognition of nouns with nominative case inflection is faster than other case inflections (Lukatela, Mandiü, Gligorijeviü, Kostiü, Saviü and Turvey 1978; Lukatela, Gligorijeviü, Kostiü and Turvey 1980) in contrast to verbs and adjectives that show frequency-based processing (Feldman and Fowel 1987; Kostiü and Katz 1987; Kostiü 1991). The direct object is expressed by accusative case. The lexical form variation of a noun such as ‘medved’ (bear) is illustrated in (3). This procedure does not require any exchange of information with other constituents and is therefore computed at the lexical stage. The task

Lucija Medojeviü

185

for the learner is to acquire the singular /plural alternation expressing values of the NUM(ber) feature and/or the nominative /accusative variation for the case feature. (3) a. Medved N (Ĺ PRED) =‘bear” (Ĺ NUM) = SING (ĹGEN) =MASC (ĹCASE)=NOM b. Medved-i

N

(Ĺ PRED) =‘bears” (Ĺ NUM) = PL (ĹGEN) =MASC (ĹCASE) = NOM

c. Medved-a

N

(Ĺ PRED) = ‘bear” (Ĺ NUM) = SING (Ĺ GEN) =MASC (Ĺ CASE) = ACC

d. Medved-e

N

(Ĺ PRED) =‘bears” (Ĺ NUM) = PL (Ĺ GEN) =MASC (Ĺ CASE) = ACC

Stage 3(Phrasal): Two types of NP structure discussed previously typically require the phrasal procedure that is, feature unification between constituents in the phrase. The variation of feature value of NPs required for the phrasal procedure to be acquired is illustrated in (4). (4) Noun Phrasal procedures a. Modifier (Nom, Sing) + N (Nom, Sing) Modifier (Acc, Sing) + N (Acc, Sing) b. Quantifier + (Modifier (Gen, Pl)) + N (Gen, Pl) Structures listed in (4a) represent the agreement of modifiers with the noun in number and case as presented in Figure 1 in section 3. An example is given in (5). (5) Ov-u-ACC.S.F plav-u-ACC.S.F haljin-u-ACC.S.F ‘This blue dress’

186 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

The quantifier NP in (4b) represents agreement of elements in plural and can occupy an overt or oblique case position. A formal description is presented in Figure 2. The numeral imposes the genitive of quantification onto the adjective and noun that agree in number, case and gender as in (6) below. (6) Tri-PL velik-a-GEN.PL.M medved-a-GEN.PL.M ‘Three big bears’ In every case the values of the feature(s) expressed in the agreeing elements must be compatible, e.g., if the value of the nominal NUM(ber) feature is S(ingular) then the feature of the agreeing modifier must also be S(ingular). To sum up, in order to achieve the process of unification in the structures presented, the learner must identify the head of the phrase and exchange the feature values information with the modifiers. Stage 4 (Inter-phrasal): This inter-phrasal level of information exchange corresponds to the level of the sentence which requires exchange of information between phrases with different heads (Di Biase and Kawaguchi 2002). The lexical entries in (7) show the information distribution between the various phrasal elements. (7) Devojþica N (Ĺ PRED) = ‘girl’ (Ĺ NUM) = SING (Ĺ GEN) = FEM (Ĺ CASE) = NOM je

AUX

bila

V

(Ĺ PRED) = ‘to be’ < SUBJ, COMP> (Ĺ TENSE) = PAST (Ĺ PER) =3 (Ĺ NUM) = SING (Ĺ GEN) = FEM

umorna.

ADJ

(Ĺ ADJ PRED) = ‘tired’ (Ĺ NUM) = SING (Ĺ GEN) = FEM (Ĺ CASE) = NOM

‘Girl was tired’.

(Ĺ PRED) = ‘to be’ (Ĺ TENSE) = PRESENT (Ĺ PER) =3 (Ĺ NUM) = SING

187

Lucija Medojeviü

The noun ‘Devojþic-a’ (‘girl’) and the predicative adjective ‘umorn-a’ (‘tired’) must agree in number and case (Singular and Nominative) as well as gender. The copula ‘je bila’ (‘to be’) also agrees in number and gender but is not assessed in this study as it focuses on information exchange between nominal elements in the sentence. Figure 3 also illustrates the corresponding c-structure to f-structure mapping.

SUBJJ

PRED NUM GEND CASE

‘girl SING FEM NOM

S

PRED ‘to be’ NP

VP

TENSE PAST V’

COMP

ADJ PRED ‘tired’ d’ NUM SING GEND FEM CASE NOM

N

Cl

Devojþic-a je ‘Girl was tired.’ ‘G

AP V bila

A umorn-a.

Figure 3: Parallel to c- to f-structure mapping ‘Devojcica je bila umorna’ To sum up, acquirers must match feature information for gender, number (singular) and case (nominative) across different phrases (noun and predicative adjective). Form variation in terms of gender (feminine, masculine and neuter) is used as a pointer whether the learners have acquired this stage.

188 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

The hierarchy just illustrated above is proposed as a metric to test the language attainment, in Serbian, by Serbian-Australian teenage bilinguals because it shows the level of language processing skills that the bilinguals possess and, indirectly, reveals the rate of this language development in prevailing Australian conditions. Given the restricted conditions of Serbian language acquisition in Australia, the following question is posed: Question 1: Are the language maintenance efforts of Serbian-Australian teenage bilinguals helping them attain the inter-phrasal level of morphosyntactic achievement in the proposed Serbian PT-based hierarchy? It is expected that the teenage bilinguals will be able to communicate effectively, in Serbian, with members of their family. Serbian is one of the numerous community languages in Australia spoken by more than 100000 migrants and their descendants (Dimitrijeviü 2004a, b). The process of language shift from Serbian to English in the Serbian community is an ongoing process, and was possibly slowed down by such factors as the continuous migration of Serbs during the period following the Second World War and the 1990s, Australian multicultural policy, and the strong presence of the Serbian Orthodox Church (Dimitrijeviü 2004c). Assuming that these teenage bilinguals, attending the secondary Australian school (year 7 to 12), have a positive attitude towards the language maintenance and use Serbian in the home domain, the following hypothesis is proposed: Hypothesis 1: Serbian-Australian teenage bilinguals will reach the interphrasal level in the proposed PT hierarchy. If the predictions of hypothesis 1 are supported, the proposed Serbian PTbased hierarchy will only show the use of limited Serbian language structures by the three bilingual teenagers that are assumed to be developmentally entrenched. Given the richness of the pragmatic-syntax choices in Serbian language and the flexibility of their use, the proposed hypothesis is not enough to assess their use of Serbian language. Thus, PT’s extension is used to predict the bilinguals’ use of some pragmatic features of the language and is discussed in the next section.

5. Pragmatic-syntax choices in Serbian This part of the chapter describes a selection of pragmatic-syntax choices and non-default mapping of argument roles utilised in Serbian and

Lucija Medojeviü

189

presents the hypothesis for the bilinguals’ language behaviour using the current extension of PT (Pienemann et al. 2005). The description of the informants’ non-default mapping skills is necessary to complete the picture of their language use given the typological characteristics of Serbian. As mentioned previously Serbian word order is flexible and its alternation ensures effective communication. However, it may cause learning problems for the bilinguals as skills in handling such structures require higher levels of morphosyntactic development than those proposed in the Serbian PT-based hierarchy for (obligatory) morphosyntactic structure. The deployment of discourse-pragmatic choices minimally requires learning to map grammaticised discourse functions (Bresnan 2001) as well as special lexical features. Thus, if the bilinguals do attain the highest level of the proposed PT hierarchy, that does not imply that they will necessarily have skills in non-default mapping of argument roles or deviate from the canonical sentence schema. The inter-phrasal structure proposed in the Serbian PT-based hierarchy requires default SVO word order and is the order which can be implemented with the least processing cost (Pinker 1984; Pienemann et al. 2005).This implies the use of the basic architecture of syntax with its three parallel levels of structure in a lineal order, that is the Unmarked Alignment Hypothesis where the parallel structures are mapped onto each other in a one-to-one manner. This corresponds to the inter-phrasal stage in the Serbian PT-based hierarchy as well as the default SVO word order. Serbian children and adult speakers, as we saw earlier, rely on morphological markings and word order strategy in sentence interpretations (Slobin 1973; Uroševiü et al. 1986, 1988). To test the pragmatic-syntax structures of the language, it is necessary to propose a selection of structures that would capture the deviation from the default Unmarked Alignment Hypothesis. Serbian topic-comment structure allows for any constituent in the sentence to be emphasised by assigning a more prominent position, in linear order, to the topicalised element. Differentiation of the Topic and Subject in Serbian can be achieved by alternative word order that disrupts the linearity of the Unmarked Alignment Hypothesis. A speaker is able to assign the TOP function to a core argument other than SUBJ by mapping it onto the most prominent (first or early) position in the sentence. For example, (8a) shows default word order in Serbian while (8b) shows an alternative word order with the object being assigned prominence by placing it in initial position (whereas it is canonically post-verbal).

190 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

(8a)

Canonical SVO in Serbian Macka je pojela rib-u. S V O Cat-NOM eat-PAST fish-ACC. ‘The cat ate the fish.’

(8b)

Object topicalization in Serbian Rib-u je pojela macka. TOP V S Fish-ACC eat-PAST cat-NOM. ‘The cat ate the fish.’

In this example, the object of a transitive verb has been topicalized and assigned the first position in the sentence demonstrating the OVS word order. This does not require any change of the morphological properties of the directly marked NPs or any change of lexical mapping as TOPic is a grammaticalized function. The object is still marked by accusative case and the sentence is in active voice. To be able to alternate the word order, the bilingual must rely on morphological markers for the sentence interpretation and not on the default word order. Thus, word orders such as OVS or OSV can be expected as examples of topicalization when the speaker is able to assign the TOP function to a core argument other than SUBJ (Pienemann et al. 2005). This implies that functional assignment (Sprocedure) must be in place. Further, Serbian also allows a different mapping of roles on to grammatical functions altogether, i.e., with the passive voice. Basic passive voice in Serbian requires similar processing procedures and resources that were described for the Lexical Mapping Hypothesis (Pienemann et al. 2005) in section 2. It involves the mapping of the patient onto the Subject function and suppression of the agent role or its demotion to a non-core adjunct function (Bresnan 2001). The use of passive in Serbian is not widely used due to the flexibility and effectiveness of the free word-order (Novakoviü 2004). This further deviates from the default SVO and relies on additional mapping principles and exceptional lexical entries. An example is presented in (9).

191

Lucija Medojeviü (9a)

Active voice

Agent

Patient

SUBJECT

VERB

ODJECT

Mack-a je poje-la rib-u. Cat-NOM be-AUX eat-3.S.PAST fish-ACC. ‘Cat ate the fish.’ (9b)

(thematic role) (grammatical function) (constituent structure)

Passive voice

Patient SUBJECT

Agent VERB

ADJUNCT

(thematic role) (grammatical function)

Rib-a je pojede-na od strane Mack-e. (constituent structure) Fish-NOM be-AUX eat -3.S.PASSIVE by cat-GEN. ‘The fish was eaten by the cat’. As the patient is mapped onto the subject function it receives nominative case ending. A passive particle (-na) is added to the perfective verb stem and the agent may be expressed as an Adjunct (therefore it is optional) in the ‘by’ clause that starts with a dual preposition (‘od strane’). Acquirers of Serbian must thus demonstrate the use of these processing routines and morphological marking on the elements of the sentence. Topicalisation and Passivization represent the options available to the learner in the Serbian pragmatic-syntax interface. Based on the theoretical groundings and predictions presented above, the following question is asked: Question 2: Are Serbian-Australian teenage bilinguals able to express the optional pragmatic structural choices through alternative word order and/or passive constructions? Under the assumption that these teenage bilinguals do attain high levels of language processing as well as have positive attitude towards the language maintenance and use Serbian in the home domain, it can also be

192 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

expected that they will be able to use at least some of the pragmatic-syntax choices of Serbian discussed above. Thus, the proposed hypothesis is: Hypothesis 2: Serbian-Australian teenage bilinguals will be able to use some of the pragmatic-syntax choices of Serbian, such as topicalization (alternative word order) and non-default mapping of argument roles (passive constructions). In an experimental study of English voice choices, Tomlin (1995) hypothesises that the speaker marks the referent in focal attention as the syntactic subject of the utterance under current formation. Tomlin (1995) tested the hypothesis with a computer animated clip Fish Film designed to elicit the interaction between the syntactic subject and focal attention. This Fish Film consists of 32 trails where one fish eats another fish identical in shape and size but different in colour. At equal speed the two fish approach each other and one swallows the other. The fish colour is selected randomly and the direction of the agent is counterbalanced. Half of the trails are agent-trails, half patient-trails. Each trail is visually primed with an arrow to attract focal attention on the referent that is either an agent or a patient. The timing between the priming and the jaw closing is less than 150ms, thus blocking the attention movement elsewhere as this is possible in such visual representations. Kawaguchi (2005b) and Wang (2006) demonstrated that learners of English as a second language utilised passive voice constructions only if they had attained the required processing skills. Wang (2006, this book) used the Fish Film task for data elicitation in his study and pointed out the very compelling visual context of it in prompting alternative voice constructions by learners. The present study also utilises this visual task to investigate whether the Serbian-Australian bilinguals utilise word order alternatives and/or passive constructions when visually cued to do so. The following section of the chapter describes the design of the study and the bilingual informants.

6. Method Informants The design of the study is cross-sectional with a quantitative approach (Doughty and Long, 2003). The three teenage Serbian-Australian bilinguals who participated in this study were born in Australia or migrated to Australia in their infancy with both parents originating from Serbia. Two of the informants were attending a secondary school in

Lucija Medojeviü

193

Sydney and acquired Serbian in their home environment; the other informant, also a secondary school student who acquired Serbian in a home environment and in community but, in addition, he is receiving formal instruction in Serbian through the Saturday School of Community Languages. This latter informant was born in Serbia and arrived in Australia with his family at one and a half years of age. The other two informants were born in Australia. All three informants are acquiring two languages as their first language as in common conditions of bilingual first language acquisition (BFLA) in a migration situation in what Vihman and McLaughlin (1982) called “an environment-bound language, with one language (Serbian in this case) at home and another (i.e., English) in the community” (cf. Qi, Di Biase and Campbell 2006). In addition, a native adult Serbian speaker participated in the study as the comparison informant to insure that the range of structures to be elicited is actually produced by natives in similar situations of elicitation. The bilingual teenagers’ parents were all born in Serbia and all speak the same Ekavic dialect as they originate from Vojvodina, a northern part of Serbia. The native speaker also originates from this same region. Each informant is assigned a fictional name to protect their privacy. Ethics approval was obtained from the University of Western Sydney Ethics Panel (Human Subjects) for this study. Table 2 below offers a description of each informant. Name Trish

Year at school Year 7

Ned

Year 9

Don

Year 12

Nicole

-------

Other information Thirteen years of age. Born in Australia. Fifteen years of age. Born in Australia. Seventeen years of age. Arrived in Australia as an infant aged 1.5 years. Two years of attendance of Serbian language classes at the Saturday school of Community languages. Thirty five years of age. Arrived in Australia at the age of 20 and has lived in Australia for 15 years

Table 2: An overview of the informants

194 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

Procedure Data elicitation was dyadic in nature (the informant and the researcher who is a native speaker of Serbian) and was conducted in the informants’ homes. The session with each informant consisted of four parts reflecting different tasks used for data collection. Besides some spontaneous conversation (part 1) informants were asked to perform a variety of communicative tasks including a spot-the-difference task (part2), narration (part 3) and an online description of a video (part 4). Interviews lasted 3540 minutes, in an attempt to elicit a high number of contexts for producing the test structures. The audio data was transcribed using the transcription conventions in Di Biase (2000). Part 1 was a short spontaneous conversation where informants were asked some personal experience questions for the purpose of gathering information on language maintenance strategies. Part 2 was a “Spot–the-difference” task where the informants were presented with two pictures both with a park scene with minor differences. They were asked to pick one of the pictures and not to reveal to the interviewer which one they picked. The task was to spot at least seven differences between them by asking questions. Part 3 of the session was designed to elicit a narration. This is a story-telling task where informants were presented with a sequence of pictures with illustrations of a well known story for children, “Goldilocks and the three bears”. They were asked to re-tell the story in Serbian following the pictures. These tasks were used to elicit, through near-naturalistic conversation, a range of structures presenting various degrees of complexity or processing problems. The tasks were obtained from the COASIT Teachers Inservice CD (Di Biase 2003). Part 4 was a short 4 min 40 second video clip Fish Film by Tomlin (1995) as described earlier. This particular task elicits on-line speech production by informants in conditions that would usually make native English speakers use passive constructions (Tomlin 1995).The instructions for the task were translated literally into Serbian from Tomlin’s (1995) design. Data analysis A full distributional analysis was used to analyse the data according to language specific structures (cf. Di Biase & Kawaguchi 2002). The theoretical framework of this study is Processability Theory (Pienemann 1998a; Pienemann et al. 2005) which specifies ‘emergence’ criteria for syntactic and morphological structures in determining the stages of the informants’ language acquisition rather than using accuracy measures.

Lucija Medojeviü

195

Emergence is defined as a point in time corresponding to the first systematic and productive use of structures (Pienemann 1984). In this cross-sectional design ‘emergence’ means ‘presence’ (Palotti 2007) that will show that the structures in the proposed hierarchy have emerged at some time and are now being used systematically and productively. The acquisition criterion in this study is based on types rather than on tokens. More precise representation of actual morpheme production by these bilinguals would not provide us with a clear understanding of their basic language use in terms of processing. Reducing the complex Serbian noun morphology to a particular inflectional type would require well controlled tasks that would not embrace the purpose of this study. A precise criterion for establishing what emerges must therefore be provided. The following exclusion criteria were formulated. All nouns and phrases were included in distributional tables expect a) echoes of words or phrases uttered by the interviewer, b)self-repetitions of words or phrases by the informant c) uninflectible nouns and d) words with inaudible morphological endings. Turning to factorization, case assignment (NOMinative / ACCusative) and number assignments (Singular / PLural) on nouns as well as noun and modifier agreement in case and number were considered for stage 2 and 3, disregarding gender. For the inter-phrasal stage in the Serbian PT-based hierarchy gender (feminine, masculine and neuter) was used to verify lexical and form variation due to subject and predicative adjective agreeing in same case (NOM) and number (S). For each informant a distributional table was formulated for each stage separately. Each category of noun (NOM.S, NOM.PL, ACC.S and ACC.PL) in the lexical stage, category of phrasal (NP agreement) and inter-phrasal (Sub-pred,adjective agreement) was classified as being supplied, oversupplied or not supplied in obligatory context. For example, at the lexical stage ‘supplied’ meant the use of category ACC.S in an object-like function. ‘Oversupplied’ meant the use of the ACC.S inflectional morpheme in a function other than direct object (replacing another case requirement). ‘Not supplied in obligatory context’ meant that the ACC.S was not used in (for ex. NOM.S used instead of ACC.S) in the function of direct object. Productive use is demonstrated if at least two morphological and /or morphosyntactic minimal pairs are produced at the given hypothesised stage in the hierarchy. The requirement that the structure should occur more than once ensures that the structure produced is not formulaic in nature. All distributional tables for the proposed hierarchy were summarised in one table presented in the results section based on the first

196 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

three parts of the interview with each informant. The total number of occurrences of respective structures and the number of valid positive instances were counted and recorded in tables presented in the results.

7. Results The results for the proposed Processability hierarchy for Serbian are presented in Table 3 below. The broken line in Table 3 marks the level in the PT hierarchy that the informants attained in this study. The results show that all informants produced a sufficient number of structures for each hypothesized level and attained the inter-phrasal stage in the proposed Processability hierarchy for Serbian in accordance with the emergence criteria specified in the previous section. The total number of contexts where a particular feature or structure has been ‘supplied’ is marked by (+) if used in appropriate context, by (-) if ‘not supplied in obligatory context’ and by (>) if the feature or structure appears to be ‘oversupplied’. An example of the interpretation of the table is given for stage 2. Some examples are provided for lexical and form variation in respect of number and case features. Trish, for example, supplied 12 occurrences of accusative singular nouns (ACC.S) in appropriate contexts (+12). An example is presented in (10). (10)

Moja samo ima one blue Mine-POSS.PRO only have-3.S that blue

i onak i … and then and …

yellow suknj-u. yellow skirt-ACC.S.F ‘Mine only has that blue and then and, yellow skirt” Trish however failed to provide the accusative plural (ACC.PL) in one obligatory context (-1) in an example presented in (11) below. She used genitive plural marker on the noun ‘krevet’ (‘bed’) instead. (11)

Onak su videli krevet-a što su bili tu. Then see-3.PL.PAST bed-GEN.PL.M that be-3.PL.PAST here. ‘Then they saw the beds that were there’

197

Lucija Medojeviü

Trish

Ned

Don

Nicole

Stage 4: S-Pred. Adjective Agreement Neuter Masculine Feminine Stage3: NP agreement

+1 (1), -1 +1 +2, -1

+1, -1 +4 +1

- - +2, -1 +4

+1 +2 +2

NOM. S +4 ACC. S +8, >11 NUM + (mod) + N +5 (1) Stage 2: Category & features

+17, >2 +14, >11 +6 , -1

+9 +11, >2 +6 (2)

+13 +18 +10 (1)

NOM. S ACC. S NOM. PL ACC. PL

+2, >1 +10, >10 L - - +2

+5 +5, >7 L, -3 +1 +4, >1

+7 +12, >1L +3 +6

+1 , >1 +12, >12 L +2 +7, -1

Table 3: Results for the proposed PT hierarchy for Serbian Trish also oversupplied singular accusative endings in 12 contexts (>12). An example is presented in (12). The accusative case feature was used instead of more marked oblique cases (in this example locative case), while the number feature is correctly supplied. (12)

Ali radim samo malo u škol-u. But work-1.S.PRES only little in school-ACC.S.F. ‘But I only do a little bit of work at school’.

The three examples used above illustrate how the spoken performance of the informants was scored in respect to the required number and case

198 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

features. According to the emergence criteria discussed previously, Trish supplied sufficient positive evidence of lexical and structural variation in obligatory contexts required for the (noun) category procedure (NOM.S = 1, ACC.S = 12, NOM.PL = 2, ACC.PL = 7). Other informants in the study also supplied lexical and structural variation in all the required procedures as Table 3 shows. Thus, the inter-phrasal stage in the developmental hierarchy is attained by all three informants in this study. The results in terms of the informants’ performance in response to the Fish Film task, used to elicit topicalization and/or passives in Serbian, show their range of responses in handling these pragmatic-syntax choices. The patient-cued episodes (sixteen) in this task are represented in Figure 4. Responses were divided into four possible categories, each representing structures elicited by the task. The ‘Agent-active’ category represents responses in active voice that start with the agent in initial position (as subject) despite the patient being cued. In active voice the lexical verb is formed from an INPER(fective) verb stem and tense markers as given in (13). (13)

Crven-a Red-NOM

rib-a je fish-NOM be.AUX

poje-la eat.INPER-3.S.PAST

plav-u rib-u. blue-ACC fish-ACC. ‘The red fish ate the blue fish.’ Here informants utilise the Unmarked Alignment Hypothesis (Pienemann et al. 2005), i.e., they map argument roles canonically onto the surface form. An example is in the canonical SVO in Serbian that corresponds to this category. The second category is labelled “passive”. An example is given in (14). (14) Plav-a rib-a je pojede-na Blue-NOM fish-NOM be.AUX eat.PERF-PASSIVE od strane crven-e rib-e. by red-GEN fish-GEN. ‘The blue fish was eaten by the red fish’. The Lexical Mapping Hypothesis (Pienemann et al. 2005) predicts that this kind of structure may be produced by learners who have

Lucija Medojeviü

199

acquired the S- procedure, but we need to remember that using a passive rather than active construction is an option for the speaker. In passive voice the lexical verb is formed from a PERF(ective) verb stem and passive particle ‘na’. The category of ‘Patient-actives’, on the other hand, refers to the informants’ missed attempts to form passives. The patient in these utterances is mapped onto the subject initial position, however the passive morphosyntax is inappropriate e.g., inappropriate passive particle ending on the verb or the use of a verb in active voice as in (15). (15)

Plav-a rib-a je poje-la Blue-NOM fish-NOM be-AUX eat.INPER-3.S.PAST od strane crven-a rib-a. by red-NOM fish-NOM. ‘The blue fish ate by the red fish’.

The category “Alternative WO” represents the informants’ use of alternative word order (OVS, OSV etc.), which is perfectly grammatical in Serbian, and follows the predictions of the Topic Hypothesis (Pienemann et al. 2005) as presented in (16). (16)

Plav-u rib-u TOP Blue-ACC fish-ACC crven-a

je be-AUX

rib-a. S red-NOM fish-NOM. ‘The red fish ate the blue fish.’

poje-la V eat.INPER-3.S.PAS

200 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

Fish Film result for 16 patient-cued episodes 16 14 12 10

Passive Agent-active

8

Patient-active 6

Alternative W O

4 2 0 Trish

Ned

Don

Nicole

Figure 4: The informants’ individual speech production results in the Fish Film task To detail the interpretation of the results in the graph, the findings are summarised in Table 4 below. Then the findings are discussed for each informant individually. Informants

Agent-active (UAH)

Passive Patient-active Alternative WO (LMH) (Missed) (TOP hypothesis)

Trish

12

--

4

--

Ned

7

4

5

--

Don

1

5

--

VOS =5 OVS =5

Nicole

13

--

1

OVS =2

Table 4: Summary of findings for the Fish Film in Serbian

Lucija Medojeviü

201

Trish acquired Serbian in a limited home environment. She attained the inter-phrasal level of language processing as demonstrated previously in this chapter. She made four attempts to formulate the passive construction in patient-cued trails of the Fish Film task. She failed to make a clear distinction between the past tense ending and passive particle suffixed to the verb. Instead, they were used together. An example is given in (17). A lexical verb in the Serbian passive construction does not contain the information about the tense (perfective verb stem + passive particle). This is perceived from the auxiliary accompanying the passive form of the lexical verb. Nevertheless, the patient in Trish’s example was mapped onto the subject first position following the predictions of LMH and was put into the category of patient-actives. (17)

Ima jopet crna i bila je pojeda-na-la za drugi. Have again black and was eat- PASSIVE-PAST for other. ‘There is a black one again and it eaten-ate by the other one’.

Out of 16 patient-cued events Trish produced four patient-actives discussed above where the patient was mapped onto the subject position. She produced agent-actives in all other events even though the arrows cued the patient fish as (18) shows. (18)

Sada ima jopet dve .. i bel-a je poje-da-la Now have again two ..and white-NOM eat.PERF-PAST ov-u drug-u. other-ACC one-ACC. ‘There are two again and the white one ate the other one.’

Ned, the second informant, also acquired the inter-phrasal level of processing skills in a limited home environment. He produced four passive constructions as exemplified in (19). (19)

Roza riba je bila ubije-na Pink-NOM fish-NOM was kill.PERF-PASSIVE od crven-u rib-u. from red-ACC fish-ACC. ‘The pink fish was killed by the red fish.’

202 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

In five out of sixteen patient-cued episodes of the Fish Film, Ned mapped the cued patient onto the subject position but used a verb in active voice that consequently allowed for the patient to be interpreted as an agent in the sentence. These examples are treated as patient-actives, as in (20) where he used the verb ‘je umrla’ (died). This verb does not require a ‘by’ clause and the patient fish became the experiencer. (20)

Crvena riba je umr-la Red-NOM fish-NOM be-AUX die-PAST od siv-u rib-u. from grey-ACC fish-ACC. ‘The red fish died by the grey fish’.

Ned demonstrated the use of passive construction and utilized non-default mapping of argument roles. However, this was not consistent as he also produced seven agent-active examples despite the cue. Don also acquired Serbian at home and, in addition, he has been attending Saturday community school classes for two years. He produced five passives. Further, Don produced alternative word order to topicalize the patient in the sentence following the predictions of Topic Hypothesis. He used the OVS word order where he mapped the patient onto the first position in the sentence while maintaining object mapping as in (21). (21)

Bel-u

rib-u O White-ACC fish-ACC

je

poje-la V be-AUX eat.INPER-PAST

crn-a

rib-a. S black-NOM fish-NOM. ‘The black fish ate the white fish’. Don also used VOS word order as in (22).

Lucija Medojeviü (22)

203

Pojela je siv-u rib-u . V O Eat.INPER-PAST be-AUX grey-ACC fish-ACC plava

riba. S blue-NOM fish-NOM. ‘Blue fish ate the grey fish.’ Don demonstrated skills in non-default mapping of argument roles and the appropriate use of morphosyntax of passives. Unlike Trish and Ned, Don is also able to use alternative word orders to topicalize the patient-cued fish. Nicole, the comparison informant, is a 35 year old native speaker of Serbian. She produced two examples of OVS word order, such as in (23). (23)

Roza riba pliva .. nju je proguta-la Pink fish swim-PRES ..her-ACC be-AUX swallow-PAST O V plav-a ribic-a. blue-NOM fish-NOM. S ‘The pink fish is swimming .The blue fish swallowed her.’

She also had one example of what I would call patient-active presented in (24). The cued patient was mapped onto the initial position. However, Nicole reversed the roles of the agent and patient as she used the verb ‘progutala’ (swallowed) after a brief pause possibly out of confusion. She turned the mapped patient into an agent of the utterance. (24)

Ljubiþast-a ribic-a .. proguta-la je crven-u ridic-u. Pink-NOM fish-NOM .. swallow-PAST red-ACC fish-ACC. ‘The pink fish (pause) swallowed the red fish.’

In the rest of the patient-cued trails, Nicole overwhelmingly used agent-actives. As a native speaker Nicole utilizes topicalization in two examples from the task. She did not produce passives to demonstrate nondefault mapping of argument roles.

204 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

8. Discussion and conclusion The results for all informants suggest that the proposed Processability hierarchy for Serbian was successfully utilized to test the bilinguals’ morpho-syntactic processing skills. Using the proposed Serbian PT hierarchy, the hypothesis tested was supported. Results show that all informants attained the inter-phrasal level of processing skills in the proposed PT hierarchy. However, they also suggest that not all informants are equally accurate in the language. It simply shows that they have the capacity to process these structures. The use of the case marking system, for example, shows significant differences between the informants. The generalized use of accusative instead of other oblique cases (Locative, Dative, and Instrumental) in the language of the two bilinguals who did not attend additional Serbian language classes (hence the high occurrence of oversupplied ACC.S presented in Table 3) points towards a reduced case system in these speakers. An example of oversupplied accusative case instead of genitive was given in (20) and presented in (25) below. The preposition ‘od’ (‘from’) typically requires genitive case in this context. (25)

Od siv-u rib-u. From grey-ACC fish-ACC

This characteristic has also been observed in the use of Serbian language by immigrants in diaspora (Ĉuroviü 1983; Dimitrijeviü 2004a, b, c). The limited use of the case marking system in the bilinguals’ language is important when discussing the options that they may choose when assessed for pragmatic-syntax interface of the language with the Fish Film task. The hypothesis regarding the pragmatic-syntax choices produced by bilinguals was also supported. The bilingual informants in this study were able to produce some discourse-pragmatic choices. However, some differences observed between the bilingual informants are apparent especially in the processability of non-default mapping of argument roles and the use of alternative word order. The acquirer who attended additional instruction in Serbian attained a higher level of alternative word order production and skills in non-default mapping of argument roles as well as an effective use of passive than the two bilingual informants who only acquired Serbian in a home based environment. While the first uses all the options available i.e. topicalization and passives, the latter are only pragmatically aware (hence the use of missed passives called patientactives) of the mapping principles available or possible but we do not know

Lucija Medojeviü

205

whether they are developmentally ready to produce these structures (inaccurate use of the morphological markings on missed passives) given their limited case system. It is possible to assume that instructed acquirers of Serbian may have an advantage over the acquirers that learn Serbian only at home in terms of having an extra domain of language input and use. Research on second language acquisition demonstrates that language instruction helps the learner notice the form of marked structures that are less frequent in language production (Pavesi 1986; Van Patten 1996). Instruction is effective as attention to form promotes development of interlanguage grammar and acquisition (Daughty 1991; Di Biase 2002). At the same time, Pienemann’s (1989, 1998a) Teachability Hypothesis holds that while stages of acquisition cannot be skipped as a result of formal instruction, formal instruction is, nevertheless, most beneficial if it focuses on the ‘next stage’ of acquirer’s development. In other words, it may accelerate the rate of acquisition as Di Biase (2002) shows. This may impact on the importance of community language schools and its role in the maintenance of minority languages in Australia (cf. Medojeviü 2007). Furthermore, the native speaker overwhelmingly used agent actives and only two object-first examples in her speech production regarding the required pragmatic-syntax choices. The results from her speech production are in line with Myachykov, Tomlin and Posner (2005) findings for Russian adult native speakers who also utilise agent-actives (SVO) overwhelmingly irrespective of the cuing in the Fish Film task. The observed differences between the speech production by the native speaker and the bilinguals are possibly due to the more automatic linguistic processing skills that the native speaker possesses. The load of language processing is minimised and the native speaker can devote time to other processing skills such as thinking (Myachykov, Posner and Tomlin 2007). The speech production for this task is restricted by the constraints of the speaker’s linguistic knowledge. The bilingual’s two languages do not necessarily reach the same automatic processing level (Pienemann 1998a) particularly in this study due to the restricted home domain of Serbian language use compared to dominant English. The differences observed between the bilingual informants may therefore be due to the manner of language processing in terms of non-default mapping of argument roles and the use of alternative word order. It would be interesting to investigate whether there is a correlation between the limited use of pragmatic-syntax choices and the limited use of the case-marking system that potentially hinders the use of these structures.

206 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

In conclusion, the empirical data from this study has some implications on the applicability of Processability Theory and its current extension. As a universal theory of language acquisition, PT has been applied to Serbian, a language not tested for previously. Thus, this study supports and extends its typological plausibility. The predictions of the current extension of PT (Pienemann et al. 2005) proposed for second language acquisition (in particular the Topic hypothesis and the Lexical Mapping Hypothesis) were also profitably utilised for hypothesising the pragmatic-syntax choices used by the Serbian language acquirers and in fact proved useful in discriminating between advanced and less advanced speakers, thus helping to characterize bilinguals who have acquired one of their languages in more limited contexts. The results of this study show, then, that this theoretical framework is suitable to test the acquisition of bilinguals’ mother tongue in relatively constrained conditions of development. The study also throws some light on Serbian language degrees of attainment by Serbian-Australian teenage bilinguals.

References Bresnan, J. (2001): Lexica Functional Syntax. Malden, MA: Blackwell Publishers. Comrie, B. & G. Corbett (1993): The Slavonic Languages. London: Routledge. De Hauwer, A. (1990): The acquisition of two languages from birth: A case study. Cambridge: CUP. Di Biase, B. (1999): The acquisition of inflectional morphology in learners of ItalianL2 with some reference to the role of phonological word structure. Paper presented at the EUROSLA 9, June 10-12, Lund, Sweden. Di Biase, B. (2000): Second language acquisition notes and exercises. Sydney: Language Australia. Di Biase, B. (2002): “Focusing strategies in second language development: a classroom-based study of Italian L2 in primary school”. In Developing a second language: acquisition, processing and pedagogy issues in Arabic, Chinese, English, Italian, Japanese and Swedish edited by B. Di Biase. Melbourne: Language Australia: 95-120. Di Biase, B. & S. Kawaguchi (2002): “Exploring the typological plausibility of Processability Theory: Language development in Italian second language and Japanese second language”. Second language Research, Vol. 18 (2002): 274-302.

Lucija Medojeviü

207

Dimitrijeviü, J. (2004a): “Language shift and Contact induced change: Serbian in contact with English”. Melbourne papers in Linguistics and Applied Linguistics, Vol 2 (2004): 1-19. Dimitrijeviü, J. (2004b): “Code-switching: structure and meaning”. Linguistics and Literature, Vol. 3/1 (2004): 37-46. Dimitrijeviü, J. (2004c): “Code-switching and borrowing in a Serbian language community”. Melbourne Papers in Linguistics and Applied Linguistics, Vol 4/1(2004): 1-12. Doughty, C.J. (1991): “Second language instruction does make a difference: evidence from an empirical study of Second Language relativisation”. Studies in Second Language Acquisition, Vol. 13 (1991): 431-469. Doughty, C. J. & M. H. Long (2003): The handbook of second language acquisition. Malden, MA. : Blackwall. Ðuroviü, L. (1983): “The case systems in the language of diaspora children”. Slavica Lundensia, Vol. 9 (1983): 21-95. Falk, Y. N. (2001): Lexical-functional grammar: an introduction to parallel constraint-based syntax. Stanford: CSLI Publications. [Electronic version] Retrieved 22/05/2007 from: http://cslipublications.stanford.edu Feldman, L.B. & C.A. Fewler (1987): “The inflected noun system in Serbo-Croatian: lexical representation of morphological structures”. Memory and Cognition, Vol. 15/1 (1987): 1-12. Franks, S. (1995). Parameters of Slavic Morphosyntax. New York: Oxford University Press. Glahn, E., G. Hakansson, B. Hammarberg, A. Holmen, A. Hvenekilde & K. Lund (2001): “Processability in Scandinavian second language acquisition”. Studies in Second Language Acquisition, Vol. 23 (2001): 389-416. Hakansson, G. & M. Pienemann (1999): “A Unified Approach towards the Development of Swedish as L2”. Studies in Second Language Acquisition, Vol. 28 (1999): 383-420. Hakansson, G., M. Pienemann & S. Sayehli (2002): “Transfer and typological proximity in the context of L2 processing”. Second Language Research, Vol. 18/3 (2002): 250-273. Hammond, L. (2005): Serbian: an essential grammar. London: Routledge. Itani-Adams, Y. (2007). Lexical and grammatical development in Japanese –English bilingual first language acquisition. In F. Mansouri (Ed.), Second Language Acquisition Research: Theory-construction and Testing (pp. 173-198). Newcastle, UK: Cambridge Scholars Press.

208 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

Kaplan, R, & J. Bresnan (1982): “Lexical-functional grammar: a formal system for grammatical representation”. In The mental representation of grammatical relations edited by J. Bresnan. Cambridge, Mass.: MIT Press: 173-281. Kawaguchi, S. (2000): “Acquisition of Japanese verbal morphology: applying Processability Theory to Japanese”. Studia Linguistica, Vol. 52/20 (2000): 238-248. Kawaguchi, S. (2002): Grammatical development in learners of Japanese as a second language. In Developing a second language: Acquisition, processing and pedagogy of Arabic, Chinese, English, Italian. Japanese and Swedish. edited by B. Di Biase. Melbourne: Language Australia : 17-29. Kawaguchi, S. (2005a): “Argument structure and syntactic development in Japanese as a second language”. In Cross-linguistic aspect of Processability Theory edited by M. Pienemann. Amsterdam: John Benjamins: 253-298 . Kawaguchi, S. (2005b): Acquisition of Japanese as a second language: A Procesability Perspective. Ph. D. Dissertation, University of Western Sydney. Kostiü, A. (1991): “Informational approach to the processing of inflected morphology: Standard data reconsidered”. Psychological research, Vol. 53 (1991): 62-70. Kostiü, A. & L. Katz (1987): “Processing differences between nouns, adjectives, and verbs”. Psychological research, Vol. 49 (1987): 229-236. Levelt, W. J. M. (1989): Speaking: From intention to articulation. Cambridge, MA: MIT Press. Lukatela, G., Z. Mandiü, B. Grigorijeviü, A. Kostiü, M. Saviü & M.T. Turvey (1978). “Lexical decision for inflected nouns”. Language and Speech, Vol. 21(1978): 166-173. Lukatela, G., Grigorijeviü, B., Kostiü, A. & Turvey, M.T. (1980). Representation of inflected nouns in the internal lexicon”. Memory and Cognition, Vol. 8 (1980): 415-423. Mansouri, F. (2002): “Exploring the interface between syntax and morphology in second language development”. In Developing a second language: Acquisition, processing and pedagogy of Arabic, Chinese, English, Italian. Japanese and Swedish. Melbourne: Language Australia: 59-72. Mansouri, F. (2005): “Agreement morphology in Arabic as a second language: Typological features and their processing implications”. In Cross-linguistic aspect of Processability Theory edited by M. Pienemann. Amsterdam and Philadelphia: John Benjamins: 117-154.

Lucija Medojeviü

209

Medojeviü, L. (2007): The acquisition of Serbian as a family and community language by Serbian-English bilingual teenagers in Australia: A processability account. Unpublished BA (Honours) thesis, University of Western Sydney. Myachykov, A. & R.S. Tomlin & M.I. Posner (2005): “Attention and empirical studies of grammar”. The Linguistic Review, Vol. 22 (2005): 347-364. Myachykov, A., M.I. Posner & R.S. Tomlin, (2007): “A parallel interface for language and cognition: Theory, method, and experimental evidence”. The Linguistic Review. Novakoviü, I. (2004): “Language transfer in the acquisition of SerboCroatian Object Clitic by French and English Learners of SerboCroatian”. Research Centre for English and Applied Linguistics, Working Papers, University of Cambridge, Vol. 10 (2004): 197-244. Pavesi, M. (1986): “Markedness, discoursal modes, and relative clause formation in a formal and informal context”. Studies in Second Language Acquisition, Vol. 8 (1986): 38-55. Pienemann, M. (1989): “Is Language teachable? Psycholinguistics experiments and hypothesis”. Applied Linguistics, Vol. 10 /1(1989): 5270. Pienemann, M. (1998a): Language processing and language development: Processability theory. Amsterdam: John Benjamins. Pienemann, M. (1998b): “Developmental dynamics in L1 and L2 acquisition: Processability theory and generative entrenchment”. Bilingualism: Language and Cognition, Vol. 1(1998b): 1-20. Pienemann, M., B. Di Biase & S. Kawaguchi (2005): “Extending Processability Theory. In Cross-linguistic aspects of Processability Theory edited by M. Pieneman. Amsterdam: John Benjamins: 199-252. Pienemann, M. (2005): “An introduction to Processability Theory”. In Cross-linguistic aspects of Processability Theory edited by M. Pieneman. Amsterdam: John Benjamins: 1- 60. Pinker, S. (1984): Language learnability and language development. Cambridge, Mass.: Harvard University Press. Palotti, G. (2007): “An operational definition of the emergence criterion”. Applied Linguistics, 28/3 (2007): 361-382. Qi, R., B. Di Biase & Campbell (2006): “The transition from nominal to pronominal person reference in the early language of a MandarinEnglish bilingual child”. International journal of Bilingualism, Vol. 10/3 (2006): 301-329.

210 Applying Processability Theory and its Extension to Serbian as a family and community language in Australia

Slobin, D.I. (1973): “Cognitive prerequisites for the development of grammar”. In Studies of Child language development edited by C.A. Ferguson and D.I. Slobin. New York: Holt, Rinchart and Winston: 175208. Tomlin, R. (1995). Focal Attention, voice, and word order: An experimental, cross-linguistic study. In Word order in discourse edited by P. Downing & M. Noonan. Amsterdam: John Benjamins: 517-552. Uroševiü, Z., C. Carello, M. Saviü, G. Lukatela & M.T. Turvey (1986): “Some word order effects in Serbo-Croat”. Language and Speech, Vol. 29/2 : 177-195. Uroševiü, Z., C. Carello, M. Saviü, G. Lukatela & M.T. Turvey (1988): “Word Order and Inflectional strategies in Syntactic Processing”. Language and Cognition Processes, Vol. 3/1: 49-71. VanPatten, B. (1996): Input processing and grammar instruction: Theory and research. Norwood, NJ: Ablex. Vihman, M. M. & B. McLaughling (1982):” Bilingual and second language acquisition in pre-school children”. In Progress in Cognitive Development Research. Verbal processes in children edited by C. J. Brainerd and M. Pressley. Berlin: Springer Verlag : 35-58. Wang, K. (2006): The acquisition of the English passive construction by Chinese learners of ESL. Unpublished BA (Honours) thesis, University of Western Sydney. Wechsler, S. & L. Zlatiü (2000): “A theory of agreement and its application to Serbo-Croatian”. Language, Vol. 76 /4 (2000):799-832. Wechsler, S. & L. Zlatiü (2001): “Case realisation and identity”. Lingua, Vol. 111 (2001): 539-560. Zhang, Y. (2002): “A Processability approach to the L2 acquisition of Chinese grammatical morpheme”. In Developing a second language Acquisition, processing and pedagogy of Arabic, Chinese, English, Italian. Japanese and Swedish edited by B. Di Biase. Melbourne: Language Australia: 29-45. Zhang, Y. (2004): “Processing constraints, categorical analysis, and the second language acquisition of the Chinese adjective suffix-de (ADJ)”. Language learning, Vol. 54/3: 437-468.

ACQUISITION OF NON-CANONICAL ORDER IN JAPANESE AS A SECOND LANGUAGE: THE CASE OF CAUSATIVE STRUCTURE

1

SATOMI KAWAGUCHI Abstract The objective of this study is to examine how learners of Japanese as a second language (L2) acquire causative structures, that is, how do they learn to go from canonical structures i.e., unmarked mapping of AgentPatient-Verb constructions on linear order towards making more refined, non-canonical choices involving the special mapping required by causative constructions in Japanese as a second language. The framework utilised is Processability Theory (PT henceforth) (Pienemann 1998, Pienemann, Di Biase and Kawaguchi 2005) and Lexical Mapping Hypothesis (LMH) (based on Bresnan 2001). When describing a particular eventuality, a speaker has alternative structural choices. However, the structural choice of a native speaker is not arbitrary but influenced by pragmatic factors such as the discourse prominence on a particular participant (e.g., agent or patient). The choice of a specific grammatical construction over another ensures effective communication. For the learner, however, this may create structural complexities and impose learning difficulties. PT’s Lexical Mapping Hypothesis predicts that it may not be possible for L2 learners to make such a choice if they have not developed the necessary processing skills. Yet, an L2 learner will eventually need to develop the skills to produce optimal structures according to the discourse context in order to meet, efficiently, their L2 communicative needs. This study reports on results from a cross-sectional study involving 28 learners of Japanese L2, which examines the learners’ structural choices when the contexts for causative events are presented in a story-telling task. It will also elucidate the non-canonical nature of the 1 This work was supported by a Seed Grant from the University of Western Sydney, Research Office. I would like to express my gratitude to Bruno Di Biase and Kenny Wang for their comments on earlier drafts and Kathleen Olive for editing assistance. I also would like to thank my students of Japanese at University of Western Sydney who participated in this study. I am responsible for any errors.

212

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

Japanese causative construction (i.e., mono-clausal or bi-clausal) on the basis of Lexical Functional Grammar (Bresnan 2001). Furthermore, it will trace the learners’ structural choices for causative contexts and compare them to their developmental stage thus plotting their path towards Japanese causative structures. The results have implications for language teaching and learning for L2 Japanese.

1. Introduction The objective of this study is to examine the development of noncanonical mapping with special reference to the causative construction in Japanese as a second language within the framework of Processability Theory (PT) (Pienemann 1998; Pienemann, et. al 2005). In this chapter I report results from a cross-sectional study of 28 Japanese L2 learners focusing on morphological causatives (i.e., where the causative morpheme –(s)ase is attached to the stem of the verb). Lexical causatives such as kowasu ‘break’ and korosu ‘kill’ are not considered in this study as these lexical causative verbs yield canonical structures. There is more than one way for a speaker to express the same eventuality or concept. For example, the eventuality of a dog chasing a cat can be realized as “a dog is chasing a cat”, “a cat is being chased by a dog”, “it is a dog which is chasing a cat” and so on. The choice will depend on his/her discourse needs (focus, topicality, etc.). However, the range of choices may not be available to a language learner because some word orders are ‘marked’. In fact, according to PT (Pienemann 1998), the learner can produce only those L2 structures which are processable at his/her current stage of L2 acquisition. PT’s Lexical Mapping Hypothesis (Pienemann, et al. 2005) predicts that the production of more ‘marked’, i.e., non-canonical structures (e.g., passives, benefactives in Japanese) require greater processing resources. Current L1 and L2 acquisition literature (e.g. Morikawa 1991, 1997; Sasaki 1998, Murasugi, et al. 2007) report, for instance, that Japanese (morphological) causative structures are acquired later. However it is not clear what type of procedural skills may be required for this, i.e. WHEN should this happen in L2 development. PT, on the other hand, predicts just when the learner becomes capable of handling non-canonical structures: non-canonical word orders are acquired when the Sentence-procedure is in place. In this paper, I will address the following research questions:

Satomi Kawaguchi

213

Q1. Is the production of causative structures possible only for more advanced L2 Japanese learners? If so, which developmental stage(s) in PT does the learner have to be at? Q2. When learners are unable to produce causative structures, what alternative structures do they produce? Are these alternative structures different according to the learner’s stage of L2 Japanese acquisition, i.e., can the route towards causative constructions be specified? In what follows, Section 2 introduces my framework, PT, and its two hypotheses relevant to this study: The Unmarked Alignment Hypothesis and The Lexical Mapping Hypothesis. This section also shows the hypothesized hierarchy of developmental stages in Japanese as a second language. Section 3 shows a formal representation of Japanese causative structures using Lexical-Functional Grammar (e.g., Bresnan 2001). The experimental design adopted for this study as well as my hypothesized answers to the research questions are presented in Section 4 followed by a discussion of the results in Section 5 and, finally, the Conclusion in Section 6.

2. Processability Theory (PT) PT (Pienemann 1998, Pienemann et al. 2005) is a universal theory of second language acquisition based on speech processing, lexical access and memory capacity. The theory predicts universal second language (L2) developmental sequences and can be applied to typologically different languages. This is possible by utilising the notions of “feature unification” and of the “lexical mapping mechanism” within the framework of LexicalFunctional Grammar (LFG, Bresnan 2001). A number of researchers have applied PT to the acquisition of various L2s with LFG formalism, such as English and German (Pienemann 1998), Italian (Di Biase and Kawaguchi 2002), Swedish (Pienemann and Håkansson 1999), Japanese (Di Biase and Kawaguchi 2002; Kawaguchi 2005, 2007), Arabic (Mansouri 1995, 1997, 2002, 2005) and Chinese (Zhang 2001, 2002, 2004, 2005). LFG utilizes a set of key psychological factors involved in language processing and, consequently, it is compatible with psycholinguistic theories such as Pinker’s (1984) Learnability Theory as well as Levelt’s speech model (Levelt 1989) and Lexical access theory (Levelt, Roelof and Meyer 1999), all of whom use LFG as their grammar-theoretical reference point.

214

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

PT and developmental stages in L2 Because of its psychological and typological plausibility, PT is suitable for describing, predicting, and accounting for the development of L2 syntax and morphology in speech of typologically different second languages. The scope of this theory also includes the hypothesis for the developmental sequence of L2 syntax/morphology in learners’ interlanguage based on the development of procedural components. PT assumes that syntax is lexically driven, an assumption shared by Levelt’s speech production model (1989), Kempen and Hoenkamp’s Incremental Procedural Grammar (IPG) (1987), as well as Lexical-Functional Grammar (LFG) (Kaplan and Bresnan 1982; Bresnan 2001). Based on this assumption, as well as on the incremental nature of speech processing, Kempen and Hoenkamp (1987) propose that, in the course of language production, grammatical encoding is activated in the following order: (1) Speech processing procedures x lemma x the category procedure (lexical category of the lemma) x the phrasal procedure (instigated by the category of the head) x the sentence-procedure and the target language word order rules x the subordinate clause procedure – if applicable When applying the grammatical encoding process described above to L2 contexts, it can be seen that the learner needs to acquire a number of processing resources. In extending Levelt’s (1989) speech model to bilingual speakers, de Bot (1992) explains that the two language systems involved may either be used separately, or influence each other. Pienemann, referring to de Bot’s bilingual speech model, claims that “in all cases where the L2 is not closely related to the L1, different (languagespecific) procedures have to be assumed” (Pienemann 1998, 74). This means that the L2 learner needs to acquire a language-specific lexicon and build components of the formulator that are also language-specific. These include word order rules, syntactic procedures and their specific stores, diacritic parameters in the lexicon, the lexical categories of lemmas and functorisation rules (Pienemann 1998, 74). The beginning L2 learner cannot infer such language-specific components in advance and, therefore, he/she is not able to use the syntactic procedures listed above because they are not yet part of their system. Pienemann (1998) hypothesised that there is a hierarchical relationship for the acquisition of processing resources by

Satomi Kawaguchi

215

L2 learners. This hierarchical order follows the same sequence as the order of activation in the production process listed in (1) above (Pienemann 1998, 78ff.) because the processing resources to be acquired form an implicational hierarchy in the encoding process. Based on Levelt (1989), Pienemann claims that the processing resources at the lower level are a prerequisite for the higher level. He writes: A word needs to be added to the L2 lexicon before its grammatical category can be assigned. The grammatical category of a lemma is needed before a category procedure can be called. Only if the grammatical category of the head of the phrase is assigned can the phrasal procedure be called. Only if the phrasal procedure has been completed and its value is returned can Appointment Rules determine the function of the phrase. And only if the function of the phrase has been determined can it be attached to the S-node and sentential information be stored in the S-procedure. And only if the latter has occurred can the target language word order be arranged. In other words, it is hypothesised that processing devices will be acquired in their sequence of activation in the production process. (Pienemann 1998, 8)

Table 1 summarises a hypothetical hierarchy of processing procedures based on the above description. It lists the five stages of the development of the L2 processor i.e. word/lemma at Stage 1, the category procedure at Stage 2, the phrasal procedure at Stage 3, the S-procedure at Stage 4 and the S’-procedure at Stage 5, and the kinds of resources the learner’s interlanguage builds for each of these stages. The right-most column shows the structural outcome in Japanese L2 based on Kawaguchi (2005). I will now briefly summarise the (general, non-language specific) structural outcome of each stage for both syntax and morphology at different stages of language acquisition, which are described in the above Table 1 (middle column). I will also exemplify the Japanese L2 structural outcomes. The reader is referred to Kawaguchi (2005) for a fuller formalisation and empirical support for each stage in Japanese L2. Stage 1 (Word/lemma) The L2 learner has not developed any language-specific procedures at this first stage of L2 acquisition, and therefore cannot access syntactic information. Consequently, the production of phrases or information exchange between constituents is not possible. Possible outcomes at this stage are single words and formulaic expressions in the L2, which does not

216

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

require any syntactic procedure or information exchange. The beginning learners of Japanese at this stage can produce invariant single words such as ischii “delicious” and fixed expressions such as greetings e.g., Arigatoo “thank you” and Konnichiwa “good afternoon”. Stage Stage 5

Processing procedure Subordinate clause procedure

Structural outcomes (General) Main and subclauses (exchange of information across clauses) Inter-phrasal morphemes (exchange of information across phrases within a clause)

Stage 4

Sentenceprocedure and the target language word order rules

Stage 3

Phrasal procedure (e.g., Noun Phrase, Verb Phrase) The lexical procedure (category of the word, e.g. verb, noun)

Phrasal morphemes (exchange of information within a phrase) Lexical morphemes (no exchange of information), canonical word order

Lemma access

Words (invariant form); no sequence of constituents

Stage 2

Stage 1

Structural outcomes (Japanese) Particles ga/wa distinction in subclause and main clause Case marking according to the feature of the verb Non-basic verb form (e.g. Passive, Causative, Benefactive) “Exceptional” verb V-te V XP + canonical order (e.g, Adjunct-wa + SOV) Nominal marking on noun Verb inflection (TENSE, NEG, etc.) Canonical order (i.e., SOV, Verb last) Invariant form / formulaic expressions e.g., greetings

Table 1: Hypothetical hierarchy of processing procedures (based on Pienemann 1998) Stage 2 (Category procedure) Once the category procedure develops in the learner language, basic lexical category such as noun and verb is assigned to the L2 lexicon

Satomi Kawaguchi

217

and lexical morphology starts to emerge in IL. An example of lexical morphology in Japanese is verb inflection for tense as in tabe-mashita “eat-PAST”. This type of morphological marking is local in nature where there is no need for information exchange between the morphology and other sentence constituents. As for syntax, phrasal procedure and S-procedure have not yet developed and therefore exchange of information is not possible at this stage. The only possible syntactic structure now is the canonical sentence, because canonical sentences (i.e., Subject-Verb-Object in English, Subject-Object-Verb in Japanese) can be created by the direct mapping of the conceptual structure onto the linguistic form. In this case, the learner can utilise the semantic roles of the words (agent, action, patient, etc.) and order them canonically, as Agent-Patient-Action in the case of Japanese, a SOV language (e.g., watashi-wa hon-o yomimasu (I-TOP(ic) bookACC(usative) read “I read a book”)) where the subject watashi “I” may be omitted (Hinds 1982). Also, predicate-last sentence with copula (e.g., watashi-wa hatachi desu (I-NOM 20 years old COPULA “I am 20 years old”) is typical at this stage in Japanese L2. Stage 3 (Phrasal procedure) When the phrasal procedure develops, exchange of information between the head and its modifier becomes possible and phrasal morphology (e.g., -s as in two dogs in English) starts to emerge. As for Japanese, LGF-based analysis would suggest that the V-te V (V-comp V) construction as in ki-te mimasu (‘wear-comp see’ (I/you/ he/she, etc. try wearing ~)) requires phrasal procedure because information exchange is required between two verbs in terms of the ‘combinatoric TYPE’ feature (Sells, 1995, 1999). For the syntax, the learner still needs to resort to the strategy of canonical mapping of the conceptual material to the utterance. However, Pienemann (1998) and Pienemann et al. (2005) predicts the expansion of canonical word order may be possible, such as the addition of an XP (a phrase of any type) as Topic, or a question word to the canonical sentence. Such structures can still be generated without the S-procedure. In Japanese L2, the emergence of a constituent additional to canonical order (i.e., S o TOP + SOV) becomes possible at Stage 3. An example of this structure is: (2)

ADJTOP S O V Kinoo-wa Tanaka-san-ga kono tegami-o kaita yesterday-TOP Tanaka-Mr-NOM letter-ACC write-PAST

218

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

‘Yesterday, Mr Tanaka wrote this letter.’ Stage 4 (S-procedure) At this stage functional appointment Rules and S-procedure are developed in the learner’s IL system. Now, the learner can construct the phrases, assign them their function and attach them to the S-node according to the sentential information. The production of inter-phrasal morphemes is possible with this S-procedural skill. Also, the canonical order is now freed up and the learner is able to construct the sentence according to the constraints of the target grammar. Thus non-canonical mapping between the thematic role and the grammatical function becomes possible at this stage. However, this operation is restricted to the main clauses because the S’-procedure (i.e., subordinate clause procedure) has not yet developed and information exchange across clausal boundaries is not possible at this stage. An example of Stage 4 is the Japanese Passive construction: (3)

Michiko-wa inu-ni kam-are-masita Michiko-NOM dog-ACC bite-PASS-PAST “Michiko was bitten by a dog”

Stage 5 (S’-procedure) At this stage the information exchange between clauses becomes possible. The learner is able to distinguish the main clause from the subordinate clause and construct the sentence accordingly. This stage has not been investigated in the context of Japanese L2 yet. In order to account for the Sentence-procedure, PT has developed the following two hypotheses (Pienemann, et al. 2005): a) The Unmarked Alignment Hypothesis b) The Lexical Mapping Hypothesis While (a) accounts for canonical order in the acquisition of L2, (b) accounts for special word orders and non-canonical mapping between thematic (argument) roles and grammatical functions. Many scholars have pointed out a regularity in the association between grammatical functions and their characteristic thematic roles (e.g. cf. Fillmore’s 1968 pioneer work). The universality of argument-function mapping relations has been studied in typological contexts (e.g. Keenan 1976; Keenan & Comrie 1977; Hopper & Thompson 1980). Both the

Satomi Kawaguchi

219

thematic roles (Agent, Patient, etc.) and the grammatical functions (Subject, Object, etc.) form hierarchical relationships. So, the argument roles higher in the thematic hierarchy in (4.a) are more prominent in human cognition than the lower ones; for example, in an event such as “kicking”, the Agent (that is the kicker) is more salient than the Patient (the kicked) to the human mind when everything else is neutral. These roles are encoded in grammatical functions (or grammatical roles) which also display a hierarchical relationship as in (4.b) whereby if the verb encodes only one argument, it would be SUBJ. (4) a. Thematic hierarchy (Bresnan 2001, 307) Agent > Beneficiary > Experiencer/Goal > Instrument > Patient/Theme > Locative

b. Relational hierarchy (Keenan and Comrie 1977, referred to in Bresnan 2001, p. 96) core SUBJ > OBJ > OBJT 2

noncore >

OBLT > COMPL > ADJUNCT

Going back to the two hypotheses in turn now, the Unmarked Alignment Hypothesis is formulated as follows: In second language acquisition learners will initially organise syntax by mapping the most prominent semantic role available onto the subject (i.e. the most prominent grammatical role). The structural expression of the subject, in turn, will occupy the most prominent linear position in cstructure, namely the initial position (Pienemann, et. al., p. 229).

Like many other approaches to explaining language acquisition (e.g., Meisel 1991; Pinker 1984, 1989; Slobin 1985 and later), PT predicts ‘canonical order’ as the unmarked, default initial syntactic alignment. This may be viewed as an optimal alignment of the most prominent role in the argument structure hierarchy (say the ‘agent’ role) mapping onto the universally most prominent grammatical function (the subject) occupying the most prominent position (the initial position) in surface structure, as in Figure 1. 2

OBJT refers to the secondary object as in Mary gave John the book (book is the secondary object in LFG)

220

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

kick Agent

Patient

argument-structure (thematic roles)

Subject

Object

functional-structure (grammatical functions)

Peter

the dog

constituent-structure (word order)

Figure 1: Canonical mapping: “Peter kicked the dog” As for the Lexical Mapping Hypothesis, it is based on non-canonical mapping of semantic argument roles onto grammatical structure (technically, the association between argument structure and functional structure): “L2 learners initially map the most prominent thematic role onto SUBJ and gradually learn how to attribute prominence to a particular thematic role, e.g., promoting the patient (rather than the agent) role to SUBJ, first in single clauses such as in passive constructions and later in complex predicates such as causative constructions (e.g., in Japanese or Romance).” (Kawaguchi & Di Biase, 2005)

The Lexical Mapping Hypothesis predicts that non-canonical mapping is acquired later, as the non-canonical association of thematic roles and grammatical functions requires the capacity for processing higher cognitive load. In the sentence ‘a dog was kicked by Peter’, the Patient, which is a less prominent role than Agent, is associated with the most prominent grammatical function, the Subject. The identification of the phrases' grammatical function and their functional assignment, e.g., in passive and benefactive constructions, requires that the learner unifies information from different constituents: the V and the N phrases. This calls for a sentence (i.e., inter-phrasal) procedure (for further details of the processing mechanism involved in non-canonical mapping see Kawaguchi 2005). kicked Patient

Agent

argument-structure (thematic roles)

Satomi Kawaguchi

221

Subject

Adjunct

functional-structure (grammatical functions)

The dog

Peter

constituent-structure (word order)

Figure 2: Non-canonical mapping: “A dog was kicked by Peter” In PT, canonical order is hypothesised at Stage 2 (i.e., the category procedure) because the learner can utilise direct mapping of the conceptual structure onto the linguistic form and no information exchange is required between constituents. This is computationally the least costly means of organising syntax (Pinker 1984, 1989). On the other hand, non-canonical mapping is hypothesised at Stage 4 (i.e., the sentence procedure) in PT.

3. Japanese causative structures and developmental stages In order to appreciate the acquisition of Japanese causative as a second language, it is necessary to provide brief formalization of the structure. Syntactically, Japanese is a SOV, Nominative-Accusative language and is consistently head-last. In terms of morphology, Japanese is an agglutinating language, e.g., a verb consists of a stem plus strings of affixed morphemes (c.f., Shibatani (1990, p. 307). Causative verbs are productively formed by attaching the causative morpheme –(s)ase to the stems 3 . One intransitive and one transitive causative construction are given in (5a-b) respectively. In the case of the intransitive causative, the argument roles of the causer is marked as NOM(inative) and the causee is marked either as ACC(usative) or DAT(ive) with the ACC marking yielding a permissive rather than coercive, reading. (5)a

3

Tamiko-ga musuko-o/ni hasir-ase-ta Tamiko-NOM son-ACC/DAT run-CAUSE-PAST “Tamiko made her son run.”

-ase and –sase are allomorphs.

222

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

The transitive causative structure, on the other hand, always marks the causee as DAT(ive) because Japanese has a so-called “double ‘o’ constraint”, so, only the OBJ(ect) in the transitive causative structure is marked with -o (ACC) (Shibatani 1990). (5)b

Tamiko-ga musume-ni keeki-o tukur-ase-ta Tamiko-NOM daughter-DAT cake-ACC make-CAUSE-PAST “Tamiko made her daughter make a cake.”

Based on Lexical Mapping Theory (Bresnan and Kanerva 1989), PT also predicts that the causative structure in Japanese requires the sentenceprocedure. The example (6a-b) illustrates the mapping mechanism of (5ab) above: i.e., intransitive and transitive (coercive) causatives in Japanese. The causative construction involves non-default mapping between argument structure and functional structure (cf. Alsina 1992, 1996). According to Matsumoto (1996), Japanese (coercive) causatives involve causer-controlled sub-events where the logical subject of the embedded clause is fused to the patient of the matrix clause. Consequently, the Japanese causative construction involves an embedded sub-event at argument-structure level but this is realised as a single clause in functional-structure (cf. Iida, Manning, O’Neill and Sag 1994). In (5a) above, the logical subject of the base verb is realised as OBJ in the syntactic structure. Alsina (1996, p. 186) assumes that “the causative verb and the base verb undergo predicate composition yielding one, single, complex, a-structure”. For example, musuko (“son”) in (5a) is the patient of the causative verb and also the agent of the base verb. Likewise, musume (daughter) in (5b) is the recipient patient of the causative verb as well as the agent of the base verb. Thus, two thematic roles are fused into one in functional structure. As a result, the causative verb and the base verb behave like one single predicate in light of case assignment (Alsina, 1996). (6)a. hasiru: ‘cause < [Agent] [Patient] run >’ argument-structure SUBJ

OBJ

functional-structure

Tamiko musuko “son”

constituent-structure

Satomi Kawaguchi

223

b. tukuru: ‘cause < [Agent] [Recipient Patient] make >’a-structure

SUBJ

OBJ

Tamiko musume “daughter”

OBJpatient

f-structure

keeki “cake” c-structure

This construction involves non-default mapping due to the fusion of two thematic roles in the Event and the Sub-event. Therefore, their production, in Japanese, involves non-canonical functional assignment as well as specific morphological operations. In order to produce these causative structures, the learner needs to analyse the syntactic frame of the causative structure and assign the participants’ grammatical functions accordingly. This operation requires integration of information coming from the NPs and the VP. Consequently PT predicts that Japanese causative structures are acquired at Stage 4, when the sentence-procedure is in place.

4. Experimental Design The previous Section provided the formal representation of Japanese causatives and now their non-canonical nature should be evident. Based on the Processability Theory’s claim that non-canonical mapping is acquired later than canonical mapping, my hypotheses for the two research questions are as follows. Q1. Is the production of causative structures possible only for more advanced L2 Japanese learners? If so, which developmental stage(s) in PT does the learner have to be at? Hypothesis: Based on the non-canonical nature of Japanese causative structure, only learners who are at Stage 4 (i.e., Sentence-procedure) or possibly at Stage 3 moving towards Stage 4 are able to produce Japanese causative structures. Q2. When learners are unable to produce causative structures, what alternative structures do they produce? Are these alternative structures different according to the learner’s stage of L2 Japanese acquisition, i.e., can the route towards causative constructions be specified?

224

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

Hypothesis: Learners who have not developed Sentence-procedure cannot produce causative structures. Based on PT, these learners, in describing causative events, will choose canonical sentences instead. They may attempt to construct causative syntactic frames, but with errors in functional assignments of the participants (indicated by Japanese case markers). More advanced (Stage 4) learners have a wider range of structures to describe causative events which may display either canonical or non-canonical mapping. So, the route can be specified. In order to test the two hypotheses above, I conducted a cross-sectional study with 28 intermediate-advanced university learners of Japanese L2 at the University of Western Sydney (UWS), Australia, using a focused picture task. Oral production data was elicited through a self-paced storytelling task looking at a series of 16 pictures, which included at least nine causative events (see appendix). Table 2 lists all the learners who are enrolled at UWS and L1 language background according to their levels of Japanese. All interviews (about 15-20 minutes in length) were recorded and transcribed 4 Learner’s level at UWS

Number of informants (L1 background)

Intermediate level

19 informants: (English (9), Chinese (6), Korean (1), Arabic (1), Greek (1), Tagalog (1))

Advanced level

9 informants: (English (7), Chinese (2))

Native control

1 (Native speaker of Japanese)

Table 2: Informants Summary in the cross-sectional study 4

The Kunrei-style romanisation system was used for transliteration, except for long vowel sounds which I have chosen to transcribe as double vowels (i.e., aa, ii, uu, ee, oo) instead of using the usual diacritic above the vowel (i.e., â, î, û, ê, ô). This romanisation system is widely used in LFG studies of Japanese and it facilitates automatic treatment of the data.

Satomi Kawaguchi

225

As a first step in the analysis, the developmental stages of the 28 learners were identified based on the stage of L2 Japanese shown in Table 1. I then checked whether or not causative constructions were produced by the informants. Further, I analysed other structural choices, besides causative, when the learners encountered causative contexts.

5. Results Table 3 presents a syntactic analysis of the cross-sectional study corpus. The first and second rows show the PT stages and the syntactic structures respectively according to PT’s acquisitional hierarchy in Japanese L2. The left-most column lists the informant’s fictitious names. Then the frequency score of occurrence is entered according to the learner. Results from the native speaker control (Kuki) are listed in the last row. Frequency counts of the causative structure are listed in the left-most column. For the two structures “word/lemma” (Stage 1) and Nominal marking (Stage 2), “+” is entered instead of the actual frequency count due to the large number of these structures by all learners in this study. For the frequency count for some syntactic structures, such as benefactives and causatives, two numbers are given before and after a slash (i.e., “/”) indicating the number of valid, correct, instances and (after the slash) the total number of instances. For example, Informant 25 (Kate) attempted causative structures three times but none exhibited a valid case where both case markings on the noun and the causative morphology on the verb were not correctly marked (this is indicated as 0/3). The bold line around the cells shows the acquisition of particular syntactic structures. The bold line shows that these are arranged in an implicational relation. Based on this frequency analysis, the stage of the informant is listed in the second column. Those who produced Stage 3 structures and causative but not other Stage 4 structures are listed as either Stage 3 or 4. Japanese causative structures involve Stage 4 operations, i.e., Sentence-procedure based on PT analysis. But we cannot say these learners are at Stage 4 since they have produced some causative structures which are hypothesised as Stage 4: the argument would be circular.

For table 3 please refer to the appendix at the end of this chapter

226

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

According to the results shown in Table 3, eleven learners out of 28 attempted causative constructions but eight among these eleven learners did so successfully. They are: Lee, Tim, Dan, Patty, Aly, Kathy, Mary and Henry. These learners achieved at least Stage 3 or clearly Stage 4. Now let us look at some successful examples of causative structures produced by the learners at various developmental stages. Also we will look at the alternative structures describing the causative events according to the stage of the learner. Informant 18 (Kathy) is at clearly Stage 4 because she is able to handle both benefactive and passive structures, which represent non-canonical mappings. Her sentences produced in the causative contexts are listed in (7). Dan, who’s stage may be either 3 or 4, also produced causative structures as in (8). (7) Informant 18(Kathy) itsumo kare-wa watasi-ni koohii-o motte ko-sase-masu always he-TOP I-DAT coffee-ACC bring-COMP come-CAUS-POL.PRES

“He always get me bring coffee (for him)”. (8) Informant 20 (Dan) Sosite bosu-wa watasi-ni koohii-o tukur-ase-masu and boss-TOP I-DAT coffee-ACC make-CAUS-POL.PRES “And my boss gets me to make a coffee.” Being at Stage 4 does not mean that the learner would always produce structures belonging to Stage 4. Instead, learners who are at Stage 4 are able to produce any of the structures belonging to Stages1-4, i.e., they have more structural choices. Three learners at Stage 4, Amy, Shao and Becky, did not produce the causative structures. However, Informant 26 (Becky), for example, employed the benefactive structure to describe events of the mother making the girl eat vegetables for which other Stage 4 speakers may utilise the causative structure, as in two-line example (9). One Stage 4 learner, Henry, also could produce Causative-Passive construction as in (10) where even more complex non-canonical mapping is involved. (9) Informant 26 (Becky) er. watasi-no haha-wa mai maiasa yasai-o tukutte age-masita ga I-GEN my-TOP every morning vegetable-ACC cook-COMP BENE-PAST.POL but

Satomi Kawaguchi

227

watasi-wa yasai-ga sukija arimasen I-TOP vegetable-NOM like NEG.POL “my mother cooked vegetables (for me) every morning but I don’t like vegetables”. (10) Informant 15 (Henry) demo haha-ni tabe-sase-rare-masu but mother-by eat-CAUS-PASS-POL.PRES “but (I) am made to eat vegetables by my mother.” On the other hand, none of the learners at Stage 2 produce causative structure for describing causative events. They, instead, used canonical sentence (i.e., SOV where Agent is canonically mapped onto grammatical subject) as in (11) and (12), which show their restricted structural choices. This may limit their expressiveness. (11) Informant 12 (Liz) er. Watasi-wa er kopii-o sitari. Koohii-o tukaitari.. Bosu-wa daikiraidesu I-TOP copy-ACC do-and coffee-ACC make-and boss-TOP hateful-POL “er I do something like photocopying and making coffee.. I hate my boss.” (12) Informant 8 (Yang) demo (XXX) o sitari (long pause) konpyuutaa-o sitari.. etto watasi-no but (XXX)-ACC do-and computer-ACC do-and well I-GEN bosu bosu-ga bosu-ni koohii-o tukuttari eeto isogasisoodesu boss boss-NOM boss-DAT coffee-ACC make-and well busy-looks like-POL

“But I do (XXX), use computer, well my boss, for my boss I make coffee, well I am busy” Some Stage 3 learners, who have not yet acquired non-canonical argument-function mapping, may end up producing sentences involving incorrect mapping conforming to canonical SOV order as in (13a-b). PT predicts that this type of error may be committed by the learner lacking Sprocedure. So the learner may start the sentence with the most prominent participant at conceptual level but is unable to complete the rest of the sentence. In (13a) for instance, Susan started the sentence with okaasan “mother” whose argument role is causee of the event (i.e., ‘Mother is making me eat’). Producing the correct causative structure requires the speaker to encode the sub-event of the caused event, which, in turn, requires the sentence-procedure. However, as the learner has not acquired

228

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

this procedure, Susan falls back on canonical mapping whereby the sentence ended up meaning “my mother is always eating vegetables”, which is not her intended meaning. Example (13b) also shows such case. (13) Informant 04 (Susan) a. * okaasan-wa itumo yasai tabete imasita my mother-TOP always vegetable eat-COMP PROGG-POL.PAST (lit.)“my mother was always eating vegetables” (intended) “my mother always made me eat vegetables” b. *demo bosu-wa kanozyo-ni zangyoo.. zangyoo... si. simasu.... simasita but boss-TOP she-DAT overtime overtime do-POL do-POL.PAST (lit)“but her boss does.. did over time for her” (intended) “but her boss made her work overtime” Summing up, in this cross-sectional study eight learners out of 28 successfully produced causative constructions. These learners achieved at least Stage 3 or clearly Stage 4. Learners at Stage 2, instead, used only canonical structures in order to describe causative events. Some Stage 4 learners utilised other non-canonical mapping such as Benefactive and Causative-Passive to describe the causative events. Some Stage 2 and Stage 3 learners attempted the causative structure but with the wrong argument mapping. To be able to produce causatives, a learner must have achieved at least Stage 3, which means that Stage 3 is a prerequisite for producing causatives. Eleven Stage 3 learners could not produce any causative (e.g., Gay and Jane). This would indicate that causative constructions should be placed at a different stage. My three-year longitudinal study (Kawaguchi 2005) also shows that production of causative structure emerged in the learner’s interlanguage late – that was in her third year of learning Japanese L2. At this point, this learner could also produce other Stage 4 structures such as Passive and Benefactives. On the bases of the above analysis, we are now in a position to answer the two research questions and their hypotheses. As for Q1, my hypothesis is supported: only eight learners at higher PT stages (Stage 3 and 4) could successfully produce Japanese causative structures. In order to answer Q2, I looked at the learners’ structural choices for describing causative events when they do not use causative constructions. The selection of the causative construction is an option for the speaker.

Satomi Kawaguchi

229

But in the L2 context, this structure is available only to more advanced learners as predicted by PT. Then, the answer for Q 2 is as follows: In order to describe causative events, learners lacking the Sentenceprocedure used either: (a) canonical sentences consistently; or (b) attempted causatives but with the incorrect mapping. The learners at Stage 4, on the other hand, exhibited various structural choices when they did not use causative structures. They used either or both (c) canonical sentences; and/or (d) other Stage 4 structures involving non-canonical mapping such as Benefactive and Causative-Passive structures.

6. Conclusion In this study, I examined learners’ structural choice in Japanese L2 when they encounter causative events. We could see that the learners’ structural choices increase according to their stage of L2 acquisition. The acquisition of non-canonical structures (Stage 4 in Processability Theory) opens up access to more structural choices that learners may utilise according to their discourse needs. Otherwise the learner would simply stick to the canonical structure, which may not fully reveal the intended information structure in a particular discourse (e.g. prominence on a particular argument role). This not only restricts the learner’s expressivity but it also imposes a heavier processing load for the listener. For example, sentence topic or focus may not be assigned in the learner’s sentence, or participants’ roles are not assigned correctly, or the learner’s concept of coercive event may not be explicitly encoded. Thus the sentence uttered by the learner may create confusion and ambiguity. Once Stage 4 procedure is acquired the learner is in a position to access Stage 5 (i.e., the Sub-clause procedure), which in turn provides further structural choices. Thus acquiring non-canonical structures is crucial for the learner to be able to promote more effective, smooth communication. In conclusion, this study shows that learners at higher stages posses more structural choices and consequently they can choose the structure which is more faithful to his/her discourse needs. This has practical implications for both language teaching and learning (when to teach what structures) as well as language assessment.

230

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

References Alsina, A. (1992). On the argument structure of causatives. Linguistic Inquiry. 23, 517-555. Alsina, A. (1996). The role of argument structure in grammar. Stanford, CA: CSLI Publications. Bresnan, J. & Kanerva, J. (1989). Locative inversion in Chichewa: A study of factorization in Grammar. Linguistic Inquiry, 20, 1-50. Bresnan, J. (2001). Lexical-functional syntax. Malden, MA: Blackwell Publishers. Bresnan, J. and S. Mchombo (1987): Topic, pronoun, and agreement in Chichewa. Language, Vol. 63/4: 741-82. Cho, Y.Y. & P. Sells (1995): A lexical account of inflectional suffixes in Korean. Journal of East Asian Linguistics, Vol. 4: 119-74. de Bot, K. (1992). A bilingual production model: Levelt’s ‘Speaking’ model adapted. Applied Linguistics, 13 (1), 1-24. Di Biase, B. & Kawaguchi, S. (2002). Exploring the typological plausibility of Processability Theory: language development in Italian L2 and Japanese L2. Second Language Research, 18 (3), 274-302. Di Biase, B. and Kawaguchi, S. (2005). Second language development at the syntax-pragmatics interface. A paper presented at JSLS 2005 The Seventh Annual International Conference of the Japanese Society of Language Sciences, 25-26 July 2005, Sophia University; Japan. Fillmore, C. J. (1968). The case for case. In E. Bach and R. T. Harms (Eds.) Universals in Linguistic Theory. 1-210. New York; Holt, Renehart, and Winston. Hopper, J. P. and Thompson, S. A. (1980). Transitivity in grammar and discourse. Language 56, 251-299. Hinds, J. (1982). Ellipsis in Japanese. Linguistic Research INC. Iida, M, Manning, C, O’Neil, P. & Sag, I. (1994). The lexical integrity of Japanese causative. Paper presented at the Linguistic Society of America. 1994. Annual Meeting. Ishikawa, A. (1985): Complex predicates and lexical operations in Japanese. Unpublished doctoral dissertation. Stanford: Stanford University. Kaplan, R. M. & Bresnan, J. (1982). Lexical-Functional Grammar: A formal system for grammatical representation. In J. Bresnan (ed.), The mental representation of grammatical relations, 173-281. Cambridge, MA: MIT Press.

Satomi Kawaguchi

231

Kawaguchi, S. (1998). Referential choice in foreigner talk and in learners’ speech in Japanese. In N. Bramley& N. Hanamura (eds), Issues in the teaching and learning of Japanese, Australian Review of Applied Linguistics Series 5 (15), 56-76. Canberra: Journal of the Applied Linguistics Association of Australia. Kawaguchi, S. (1999). The Acquisition of Syntax and Nominal Ellipsis in JSL Discourse. In P. Robinson (ed.), Representation and Process: Proceedings of the 3rd Pacific Second Language Research Forum, 1, pp. 85-94. Tokyo: PacSLRF. Kawaguchi, S. (2005). Argument structure and syntactic development in Japanese. In M. Pienemann (Ed.). Cross-linguistic aspects of Processability Theory. 255-298. Amsterdam/Philadelphia: John Benjamins. Kawaguchi, S. (2007). Lexical mapping theory and Processability Theory: A case study in Japanese. In F. Mansouri (Ed.), Second Language Acquisition Research: Theory-construction and testing. (pp. 39-80). Newcastle Upon Tyne: Cambridge Scholars Press. Kawaguchi, S. (2008). Language typology and Processability Theory. In Keßler, J.-U. (ed.): Processability Approaches to second language development and second language learning. (pp. Newcastle Upon Tyne: Cambridge Scholars Publishing. Kawaguchi, S. and Di Biase, B. (2005). Second language development at the syntax-pragmatics interface. Paper presented in the 7th Annual International Conference of the Japanese Society of language Sciences. Sophia University, Tokyo. (June 25-26, 2005). Keenan, E. O. & Comrie, B. (1977). Noun phrase accessibility and Universal Grammar. Linguistics Inquiry, 8, 63-99. Keenan, E. O. (1976). Toward a universal definition of ‘subject’. In C. Li (Ed), Subject and Topic. New York; Academic Press. Kempen, G. & Hoenkamp, E. (1987). An incremental Procedural Grammar for sentence formulation. Cognitive Science, 11, 201-259. Lee, H. (1999). Aspectual and thematic licensing of grammatical case. CLS, 35, 203-222. Lee, H. (2001a). Optimization in argument expression and interpretation: A unified approach. PhD thesis, Stanford University: http://www.unc.edu/~hanjung/thesis.html. Lee, H. (2001b). Markedness and word order freezing. In P. Sells (ed.), Formal and empirical issues in optimality theoretic syntax, 63-127. Stanford, CA: CSLI Publications.

232

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

Lee, H. (2002). A Lexicalist OT Approach to Case. http://www.unc.edu/~hanjung/otlfg-case.pdf. Levelt, W. J. M. (1989). Speaking: from intention to articulation. Cambridge, Mass.:MIT Press. Levelt, W. J. M., Roelofs, A. & Meyer, A. S. (1999). A theory of lexical access in speech production. Behavioral and Brain Sciences, 22, 1-75. Mansouri, F. (1995): The acquisition of Arabic subject-verb agreement in Arabic as a second language. The Australian Review of Applied Linguistics, Vol. 18/2: 65-84. Mansouri, F. (1997): From emergence to acquisition: Developmental issues in Arabic interlanguage morphology. The Australian Review of Applied Linguistics, Vol. 20/1: 83-104. Mansouri, F. (2002): Exploring the interface between syntax and morphology in second language development. In Developing a second language: Acquisition, processing and pedagogy of Arabic, Chinese, English, Italian, Japanese, Swedish edited by B. Di Biase. Melbourne: Language Australia: 59-72. Mansouri, F. (2005): Agreement morphology in Arabic as a second language: typological features and their processing implications. In M. Pienemann (Ed.). Cross-linguistic aspects of Processability Theory, 117-53 Amsterdam and Philadelphia: John Benjamins: Matsumoto, Y. (1996). Complex predicates in Japanese - A syntactic and semantic study of the notion ‘word’. California and Tokyo: CSLI Publications and Kurosio Publishers. Meisel, J. M. (1991). Principles of universal grammar and strategies of language use: On some similarities and differences between first and Second Language Acquisition. In L. Eubank (Ed.), Point – counterpoint. Universal Grammar in the Second Language, 231-2276. Amsterdam: Benjamins. Morikawa, H. (1991). Acquisition of causative in Japanese. Paper and reports on child language development. Vol. 30. Department of Linguistics, Stanford University, California. Morikawa, H. (1997). Acquisition of case marking and argument structures in Japanese. Tokyo: Kuroshio Publishers. Morimoto, Y. (1998): Dative objects in Japanese -sa Nominalization. In 3rd International Lexical Functional Grammar Conference LFG-98 Proceedings of the Conference edited by M. Butt & T. Holloway King. Stanford, CA: CSLI Publications 252-263. Murasugi, K, Hashimoto, T. & Kato, S. (2007). On the acquisition of causative in Japanese. BUCLS 28 Proceedings Supplement

Satomi Kawaguchi

233

Abstracts. http://72.14.253.104/u/uwslib?q=cache:U7kVwiucMPYJ:128 .197.86.186/posters/Murasugi Retrieved on 6/24/2007 Nordlinger, R. (1998): A grammar of Wambaya. Canberra: Pacific Linguistics. Pienemann, M. & G. Håkansson (1999). A unified approach towards the development of Swedish as L2: A processability account. Studies in Second Language Acquisition, Vol. 21: 383-420. Pienemann, M. (1998). Language processing and second language development - Processability Theory. Amsterdam: John Benjamins. Pienemann, M. (2007). Processability Theory. In B. VanPatten & J. Williams (Eds.). Theories in second language acquisition: An introduction. Pp.137-154. Mahwah, New Jersey: Lawrence Erlbaum. Pienemann, M. Di Biase, B. & S. Kawaguchi (2005). Extending Processability Theory. In M. Pienemann (ed.). Cross-linguistic aspects of Processability Theory. 179-251. Amsterdam and Philadelphia: John Benjamins. Pinker, S. (1984). Language Learnability and language development. Cambridge, MA: Harvard University Press. Pinker, S. (1989). Learnability and cognition: The acquisition of argument structure. Cambridge, MA: MIT Press. Sasaki, Y. (1998). Processing and learning of Japanese double-object active and causative sentences: an error-feedback paradigm. Journal of Psycholinguistic Research, 27 (4), 453-479. Sells, P. (1994): Sub-phrasal syntax in Korean. Language Research, Vol. 30: 351-86. Sells, P. (1995): Korean and Japanese morphology from lexical perspective. Linguistic Inquiry, Vol. 26: 277-325. Shibatani, M. (1990). The languages of Japan. Cambridge: Cambridge University Press. Simpson, J. (1991): Warlpiri morpho-syntax. Dordrecht: Kluwer. Slobin, D. I. (1985). The cross linguistic study of language acquisition. Vols. 1 & 2. Hillsdale, N.J.: Lawrence Erlbaum. Van Valin Jr, R. D. (2001). An introduction to syntax. Cambridge: Cambridge University Press. Zhang, Y. (2001): Second language acquisition of Chinese grammatical morphemes: a processability perspective. Unpublished doctoral dissertation. Canberra: Australian National University. Zhang, Y. (2002): A processability approach to L2 acquisition of Chinese grammatical morphemes. In B. Di Biase (Ed.). Developing a second language: Acquisition, processing and pedagogy of Arabic, Chinese,

234

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

English, Italian, Japanese, Swedish, 29-44. Melbourne: Language Australia. Zhang, Y. (2004): Categorial analysis, processing demands, and the L2 acquisition for the Chinese adjective suffix -de (ADJ). Language Learning, Vol. 54/3: 437-468. Zhang, Y. (2005): Processing and formal instruction in the L2 acquisition of five Chinese grammatical morphemes. In M. Pienemann (Ed.). Crosslinguistic aspects of Processability Theory edited by. Amsterdam: John Benjamins: 155-179.

Satomi Kawaguchi

Appendix An example of story telling task

235

Informant’s stage the informants 2 2 2 2 2 2

3 3 3 3 3

3 3 3 3 3 3 3-4 3-4

Informant’s number 02 (Fred) 05 (Sue) 24 (Anne) 25 (Kate) 16 (Ingrid) 07 (Linda)

23 (Peter) 13 (Ely) 08 (Yang) 11 (Dbie) 12 (Liz)

01 (Gary) 06 (Vicky) 14 (Gay) 17 (Jane) 28 (Bony) 04 (Susan) 19 (Lee) 22 (Tim)

236

+ + + + + + + +

+ + + + + + + +

+ + + + +

+ + + + + +

+ + + + + + + + + + +

(2)

Stage 1 (1)

8/8 15/15 21/21 9/9 10/10 19/19 16/16 22/22

13/13 10/10 9/9 12/12 9/9

12/12 12/12 11/11 15/15 20/20 7/7

(3)

Stage 2

/ 1 7 4 1 12 11 14

1 7 / 1 1

2 2 4 11 9 6

(4)

0 0 4 0 0 1 1 1

1 1 1 2 2

0 0 0 0 0 0

(5)

1 1 5 2 1 / / /

/ / / / /

/ / / / / /

(6)

Stage 3

/ / / / / / / /

/ / / / /

/ / / / / /

(7)

/ / / / / 0/1 / /

/ / / / /

/ / / / / /

(8)

Stage 4

Acquisition of non-canonical order in Japanese as a second language: the case of causative structure

/ / / / / 0/3 4/4 4/4

/ / / / /

/ / / 0/3 0/3 /

causative

Stage 4

4 4 4

4 4 4 (Native control)

10 (Amy) 03 (Shao) 26 (Becky)

18 (Kathy) 21 (Mary) 15 (Henry) Kuki (NS)

+ + + +

+ + +

+ + +

+ + + +

+ + +

+ + +

14/14 24/24 16/16 33/33

8/8 6/6 11/11

14/14 22/22 17/17

9 0 1 2

2 5 8

8 19 5

3 3 11 6

2 1 2

1 7 3

/ 1 1 1

1 / /

1 1 2

1/1 2/3 1/1 3/3

1/1 1/1 1/1

/ / /

1/1 1/1 1/1 2/2

/ / /

0/1 / /

4/4 2/2 1/3 3/35

/ / /

4/4 4/4 3/3

237

Kuki used causative structures describing all causative events where multiple causative events were encoded into one sentence using V1-tari V2tari V3-tari sasemasu “makes (me) to do V1 and V2 and V3”.

5

Table 3: Implicational table for acquisition of Japanese L2 syntax in the cross-sectional study

(1) word/ lemma, (2) Nominal markings, (3) canonical order (i.e., Verb-final), (4) SUBJ-wa OV, (5) V-te V, (6) ADJUNCT-wa+SOV, (7) benefactive, (8) passive. The number before slash = total number of valid instances, the number after slash = total number of instance.

3-4 3-4 3-4

20 (Dan) 27 (Patty) 09 (Aly)

Satomi Kawaguchi

CASE, WORD ORDER AND THEMATIC ROLE ELENA OSSELLA DE FILIPPO

The aim of this research is to prove that, as far as the interlanguage of German L2 is concerned, verb arguments are distinguished by case markers, word order and thematic role. The acquisition of case system in German presents common features among speakers acquiring German as L1 and as L2 (Tracy 1986; Stenzel 1984; Diehl 1991; Jordens 1983), and among adult speakers of different L1 (Diehl 1991; Jordens 1983; Putzer 1994; Westheide 1981). According to previous works, the most remarkable strategies of the acquiring pattern for the case system in GSL are the assignment of a case morpheme according to the word order within the sentence (i.e.: before or after the verb) and according to the thematic role of the arguments of the predicate (Tracy 1986; Klann- Delius 1999; Diehl 1991; Putzer 1994). I like to suggest that these two strategies are used in an inversely proportional way, and that the assignment of case is based on word order during the first acquisitional phases, whereas it is based on thematic role during the last phases of the acquisitional process. Besides, the introduction of the strategy of the thematic role case marking leads to the acquisition of constraints for the use of the accusative case, which since then marks every argument of the predicate except the subject. In order to demonstrate this hypothesis, case assignment for arguments of three- place predicates (e.g.: fragen “to ask for”; sagen “to say”; geben “to give”) will be analysed according to the Role and Reference Grammar approach (Van Valin 2005; Van Valin & Lapolla 1997).

1. Three Place Verbs and Case Assignment Since ditransitives show in their logical structure three arguments, only three place verbs are analysed. According to Woolford (2001: 524- 526), in fact, dative case occurs especially with three place verbs, as the third argument should receive a different case marker from other core arguments (which generally take nominative and accusative case).

240

Case, Word Order and Thematic Role

Consequently, the presence of three arguments in the predicate structure simplify the analysis of case alternation in interlanguages. Although this research is concerned with case system acquisition, target like case markers are not considered as the only evidence for case assignment, as learners can use different strategies in order to mark the relationship between the predicate and its arguments. In other words, in GSL morphological case markers should not be present, especially during first phases of acquisition. Subsequently, case markers are acquired by learners, but they can be used in a deviant way with respect to the target language case system. Nevertheless, the data show regularities in the realization of verb arguments, since the relationships between the predicate and its elements are marked, anyway, although the relationship markers in interlanguage case system are not present in Standard German 1 .

2. Role and Reference Grammar The Role and Reference Grammar (RRG) is a grammatical model that analyses the interaction between semantic and syntactic structures (Van Valin & Lapolla 1997; Van Valin 2005). The semantic structure is universal and it consists of the lexical representation of the predicative element (the lexicon). The syntactic structure is language- specific and it consists of the predicate projection and the operator projection (the parser and the syntactic inventory). The function that relates the two structures is the Macrorole. The Macroroles are two general groups of thematic relations: the actor, whose prototypical element is the “agent”, and the undergoer, whose prototypical element is the “patience”. The linkage between the structures of language is subjected to a set of connection rules, the linking- algorithm.

1

For example, the benefactive argument, that is generally marked with dative case in the standard variety, is realized with a “preposition + determiner (ACC/ NOM/ DAT) + N” string in GSL. Although this strategy is deviant from standard language, it is highly productive for German L2 interlanguage.

Elena Ossella De Filippo

241

Figure 1: The Organization of RRG The syntactic structure consists of two projections: the constituent projection and the operator projection. The constituent projection involves the primary constituents of the sentence, the core and the periphery. The core is made of the predicative element (generally the verb), and the nucleus (that contains predicate arguments). The periphery contains all non predicative elements. The non verbal arguments can be non phrasal arguments (adverbials) and phrasal adverbials (prepositional phrases) (Van Valin 2002, 2005). The operator projection consists of the set of all operators that are part of grammatical categories. One or more operators can change the strata of which the operator projection is made of. There are three classes of operators: nuclear operators, core operators and clausal operators. Nuclear operators modify the action, the event or the state without involving the participants. Negation, aspect, and direction (referred to action or event) are nuclear operators. Core operators modify the connections between arguments and action. Direction (referred to participants), event classification, modality and internal negation are core operators. Clausal operators modify the entire sentence. These operators are state, tempus, expressive strength (expressional modality) and evidentials (epistemic evidences of facts) (Nakamura 1997; Van Valin 2002, 2005; Van Valin & Lapolla 1997). The semantic structure consists of the classification of the verb through the decomposition system in features based on the temporal characteristics of the verb. This system is based on the concept of Aktionsart (“action form”) proposed by Vendler (1967) and on the modified version of Dowty (1979). According to the decomposition system, verbs are subdivided into six classes (Van Valin 2005: 31- 35): State: these verbs indicate states (be cold; posses)

242

Case, Word Order and Thematic Role

Activity: these verbs are dynamic and atelic (run; eat) Achievement: these verbs are punctual and telic. They suppose a change of state (pop; explode) Semelfactive: these verbs are generally considered as punctual, but they can eventually turn into durative if the action is iterated (flash; cough) Accomplishment: these verbs are dynamic and telic. They suppose a change of state (learn; freeze) Active Accomplishment: these verbs indicate a telic use of activity verbs (run to school; eat the meal) Verb classes have a causative variant. For example, the achievement verb “to pop” in “The balloon pops” has a causative- achievement version like “the child pops the balloon”. For each verbal class there exists a Logical Structure (LS). The LS is the formal representation of temporal structure, i.e. of inherent lexical aspect of a verb; according to it, state and activity are the basic classes all the others are derived from. In Table 1 are exposed the logical structures of the verbal classes (Van Valin 2004: 10; Van Valin 2005).

Table 1: Lexical representations of Aktionsart categories. A macrorole is a general set (“macro”) of different thematic relations (or roles), that are considered as similar in grammatical constructions. Actor and undergoer are the two macroroles and they correspond to the two primary arguments of a transitive predicate, or to the single argument of an intransitive predicate. Actor is what generally is defined as the “logical subject” of a sentence and the prototypical actor argument is the agent.

Elena Ossella De Filippo

243

Undergoer is the “logical object” and the prototypical undergoer argument is the patient. The relationship between argument position of a predicate in the LS and the macrorole selection follow a hierarchy: the Actor – Undergoer Hierarchy (Van Valin 2004: 12; Van Valin 2005), as shown in figure 2.

Figure 2: The Actor- Undergoer Hierarchy. According to this hierarchy, in nominative- accusative languages the highest argument in the hierarchy is the argument of DO (agent), and the lowest is the argument of pred´ (x) (patient). Other thematic roles can be ascribed to actor or undergoer macroroles. Among the central arguments of a verb, one should cover the function of privileged syntactic function, that is defined by a set of constraints. This argument is called Privileged Syntactic Argument (PSA). There are two kinds of PSA: PSA controller and PSA pivot. PSA controller rules verbal agreement, it precedes a reflexive and it allows the interpretation of missing elements (shown by [___i) in adjacent sentences, like “Mary decided [___i to go home”. “Mary” is the PSA controller, and “[___i” is the PSA pivot (Van Valin 2005: 95). PSA selection is linked to the Syntactic Privileged Argument Selection Hierarchy (Van Valin 2002: 7; Van Valin 2005: 100): (1) arg of DO > 1st arg of do´ > 1st arg of pred´ (x, y) > 2nd arg of pred´ (x, y) > arg of pred´ (x) In transitive verbs both macroroles are realized. According to these constructions, the agent is an actor macrorole and is the argument of DO, too. It implies, that the agent is the PSA. In intransitive constructions, only one argument is present in the nucleus. The argument is necessarily a macrorole and, therefore, the PSA. Therefore, an undergoer should be a PSA in intransitive constructions. According to the Principles of PSA Selection, the PSA function is accessible depending on language typology.

244

Case, Word Order and Thematic Role

For accusative languages, the PSA is by default the highest argument in the PSA Selection Hierarchy, which is also the highest argument in terms of the Actor- Undergoer Hierarchy, hence the actor.

Figure 3: Summary of components of RRG linking system.

2.1 Case in RRG In RRG, since grammatical relations are not universal, case assignment is a language- specific syntactic function. Furthermore, it is strictly connected to the number of macroroles that a predicate takes as direct arguments. Consequently, case assignment undergoes case marking rules which are applicable to core arguments only, according to language typology. Case assignment rules for accusative languages are (Van Valin 2005: 108): (2) a. Assign nominative case to the highest ranking macrorole argument. b. Assign accusative case to the other macrorole argument. c. Assign dative case to non-macrorole arguments (default).

Elena Ossella De Filippo

245

For instance, in German the argument of intransitive verbs and the first argument of transitive verbs take the same case marker: a) ichNOM[U] schalf (“I sleep”), and b) ichNOM[A] lese das Buch[U] (“I raed the book”). In both examples, the PSA is marked by nominative case, even though in a) the macrorole is an undergoer and in b) is an actor. If one of the arguments can be realized through a pronoun, it is connected to the proper position in the LS, and its morphosyntactic form is assigned according to the case assignment rules. If one of the arguments can be realized through a reflexive, it is connected to the proper position in the LS and it is marked by the appropriate morphosyntactic form, in conformity with the case assignment rules. In accusative languages, in fact, reflexives can be controlled only by a macrorole.

2.2 Three Place Verbs in RRG In RRG, all three place verbs have a common complex LS, that is (Van Valin 2002: 1): (3) [do´ (x, Ø)] CAUSE [BECOME predicate´ (y, z)] In accordance with the decomposition system, three place predicates like give and show have a LS as it is illustrated in (4) and (5) (Van Valin 2002: 11). (4) give [do´ (x, Ø)] CAUSE [BECOME have´ (y, z)] (5) show [do´ (x, Ø)] CAUSE [BECOME see´ (y, z)] The LS of the verb give shows three arguments: x, y and z. Two of them are nuclear core arguments, so they are both a macrorole candidate. Necessarily, the third argument is a non nuclear core argument, and it is a non- macrorole candidate. The PSA takes nominative case, the other macrorole is marked by accusative case and the non- macrorole takes dative case. In the sentence “Pat gave the book to Kim”, Pat is the PSA, the book is the undergoer macrorole and to Kim is the third non- macrorole argument. The sentence has a semantics – to – syntax linking template like the one in figure 4 (Van Valin 2002: 12):

246

Case, Word Order and Thematic Role

Figure 4: semantics- to – syntax linking. In languages like English, case assignment rules ((2)a, (2)b, and (2)c) are valid only for pronouns. Otherwise, the macrorole assignment rules change according to the marked/unmarked choice assignment of the undergoer (Van Valin 2005: 114), as it is shown in the example in (6) (Van Valin 2002: 12). (6) a [do´ (Pat, Ø)] CAUSE [BECOME have´ (Kim, book)] b. Pat [Actor] gave the book [Undergoer] to Kim. Unmarked choice c. Pat [Actor] gave Kim [Undergoer] the book. Marked choice The verb give is, in languages like English, a dative shift verb. Dative shift phenomena imply that the lowest ranking argument in terms of the ActorUndergoer Hierarchy is not selected to be a macrorole, but is supposed to become the non- macrorole argument (that is the second highest ranking argument). Consequently, it is allowed “a marked undergoer assignment in terms of the actor- undergoer hierarchy”. The marked undergoer construction involves a non nuclear core argument to be an undergoer and the most frequent one is the beneficiary (Van Valin 2005: 121). As far as the marked choice is concerned, “... if there is an animate argument as a potential undergoer, then it must be undergoer” (Van Valin 2005: 123) 2 . Therefore, the argument predicted to be an undergoer in three place predicates is the benefactive/ recipient. The marked undergoer choice 2

See also Van Valin & Lapolla 1997: § 7.4.1.

Elena Ossella De Filippo

247

allows dative shift or every sort of “object status change”, although this is not predicted by the Actor-Undergoer hierarchy (Van Valin 2005: 124). Even though RRG assumes a distinction between marked and unmarked undergoer choice for ditransitives, in languages like German a distinction based on case markers alternation is not possible, since case affixes are marked on arguments. (7) a. [do´ (Polizist, Ø)] CAUSE [BECOME have´(Mann, Brief)] b. Der Polizist [A] gibt den Brief [U] dem Mann c. Der Polizist [A] gibt dem Mann den Brief [U] In (7)b and (7)c, the PSA is the actor and takes the nominative case. The other nuclear core argument is an undergoer and it is realized by an accusative mark. The non macrorole core argument is marked by dative. Case markers imply that, in languages like German, the marked choice of the undergoer is not directly signaled on NPs, but by a morpheme on the verb. In (8), the difference between the German verbs streichen and bestreichen (spread) is shown (Van Valin 2002: 16). (8) a. [do´ (Max, [spread´ (Max, Farbe)])] CAUSE [BECOME be-on´ (Wand, Farbe)] b. Max [A] hat die Farbe [U] an die Wand gestrichen. c. Max [A] hat die Wand [U] mit der Farbe bestrichen.

3. Elicitation of Data 3.1 The Learners The data are elicited by the transcriptions of the tasks, to which the learners of Second Language Acquisition by Adult Immigrants (SLA) project of the European Science Foundation (ESF) were exposed to. There are six ESF speakers who acquire German as L2, three Italians and three Turks (Perdue 1993: 191- 206). The Turkish learners are all aged between 15 and 16. There are two men and a woman and their length of residence (LOR) is long. The speakers learn German partly in a formal context, since they attend a 500 hours language course, the MBSE (Maßnahmen zur Beruflichen und sozialen Eingliederung – “measures for work- and social integration”), and partly in a non formal context, as they live and work in Germany. Ayshe is a woman aged 16, and she attended secondary school. She speaks English as L2.

248

Case, Word Order and Thematic Role

Çevdet is a man aged 15 and he did not attend secondary school. He speaks English as L2. Ilhami is a man aged 16 and did not attended secondary school. His L2s are English and French. The Italian learners who took part in the SLA project are aged between 20 and 23. There are two men and a woman, and they all come from Southern Italy. Their LOR is long and they did not attend any German language course. Angelina is a woman aged 20 and she attended secondary school. She does not speak any foreign language. Marcello is a 23 year old man and he attended secondary school. He speaks French as L2. Tino is a man aged 20. He did not attend secondary school and he does not speak any foreign language.

3.2 The Tasks Although they show remarkable examples for case acquisition, the tasks were not conceived in order to elicit data on case realization in GSL. From the different tasks to which the learners of SLA project were exposed, I have analysed the data elicited by three tasks: a free recall task, Modern Times, and two play scene tasks Pullover and Fahrrad (Perdue 1993: 92103). These tasks are planned as predetermined narrative patterns, so that the interpretation of relationships between verbs and their arguments in the interlanguage should be not ambiguous. Free Recall. The learner is supposed to describe a short film or some scenes of a film soon after he/ she has seen it. The six learners saw “Modern Times” by Charlie Chaplin. Play scene. This is a role play with a background. The learner is given a background, and through the received information he/ she plays a role during the recording. Fahrrad (bicycle): the learner is given a bicycle by the interviewer to go to a shop/ bank. While he/she is in the bank (or shop) the bicycle is stolen. The learner must explain what happened and should find a solution to this problem. Pullover: the learner has bought a jumper in a shop. After the first wash, the jumper has shrunk. The learner must explain what happened to the shop assistant (the interviewer), and should try to get back the money or, at least, a new jumper.

Elena Ossella De Filippo

249

The data are subdivided in three phases, according to the SLA project division. The learners were exposed to the task sets every 3/ 4 weeks (Perdue 1993: 90).

4. Analysis Learners realize one, two or three arguments of ditransitive verbs, according to four types of syntactic template: “one argument realization” (two strategies), “two arguments realization”, and “three arguments realization”. There are two strategies for “one argument realization” (OAR) template. According to the first OAR strategy, the learner realizes one argument, generally the subject. A peripheral argument, for example an adjunct, can be present. In conformity with the second OAR strategy, one core argument is realized and it is introduced as an object clause or a direct speech quotation. In “two arguments realization” (TAR) template, the PSA is generally present within the clause. The second argument is either the macrorole argument or the non macrorole one. As far as the “three arguments realization” (ThAR) template is concerned, all positions of three place verbs are fulfilled by the learner. Phase 1 The verbs used by learners during the first set of tasks are: bringen (to bring/take); fragen (to ask); geben (to give) and sagen (to say). There are 99 occurrences of three place verbs in phase 1: 25 for bringen; 6 for fragen; 5 for geben and 63 for sagen. bringen

fragen

geben

sagen

Angelina

3

0

1

7

Tino

0

0

0

5

Marcello

11

0

0

5

Ayshe

2

6

4

26

Ilhami

5

0

4

12

Çevdet

4

0

0

8

Table 2: verbs in phase 1

250

Case, Word Order and Thematic Role

There are 60 occurrences of OAR template. In the 85% of the 60 occurrences, the argument that is realized is a PSA, although German interlanguage is a pro- drop variety. In 15 examples, “one argument (+ ADJ)” strategy is used. (9) ich [A] habe gefragt I have asked Ayshe, „Pullover“, comment 278 In 45 examples, “one argument + object clause/ direct speech quotation” strategy forms occur. (10) se [A] ham gesagt nein das wirt ni so klein des des. they have said no this doesn’t become so small this this Ayshe, “Pullover”, comment 234 Twenty examples of TAR template are noticed in phase 1. (11) er [A] hat gefrag weltsche überraschung [U] he has asked which surprise Ayshe, “Modern Times”, comment 613 Finally, there are 23 occurrences of ThAR template examples. (12) sie [A] gebe a charlot [NMr] ein geschenke [U] she gives charlot a present Tino, “Modern Times”, comment 283 The majority of occurrences appear in sentences in which one central argument and an object clause or a direct speech quotation are realized. Avoiding the realization of all arguments, except for PSA, should be a strategy in order not to make mistakes. Also, there are 63 occurrences of sagen. Since learners are not yet able to transfer dialogue parts into indirect speech, they use this strategy during the retelling task.

Elena Ossella De Filippo

22%

44%

251

OAR Obj.Cl./ Dir. Speech OAR TAR

19%

ThAR 15%

Graphic 1: percentages of occurrences of templates. Phase 2 In phase 2 there are six three place verbs in the interlanguage of learners: bringen (to bring/take); erklären (to explain), erzählen (to tell), fragen (to ask); geben (to give) and sagen (to say). Ditransitive verbs occur 124 times in data: 15 bringen, 1 erklären, 1 erzählen, 3 fragen, 16 geben and 88 sagen.

bringen

erklären

erzählen fragen

geben

sagen

Angelina 2

0

0

0

1

10

Tino

1

0

0

0

2

22

Marcello 5

0

0

0

1

4

Ayshe

2

1

1

1

8

27

Ilhami

2

0

0

2

3

15

Çevdet

3

0

0

0

1

10

Table 3: verbs in phase 2 Occurrences of OAR template are nearly stable with regard to phase 1, as there are 64 examples. For all examples, the argument realized is the PSA. There are 21 examples of “one argument (+ ADJ)” strategy. (13) un sie [A] bringen auf die kleine camion [T.it.] and they take to the small lorry Tino, “Modern Times”, comment 80

252

Case, Word Order and Thematic Role

In 43 sentences, “one argument + object clause/ direct speech quotation” strategy is used: (14) dann erzählen wir [A] was uns passiert than we tell what it is happened to us Ayshe, “Fahrrad”, comment 529 There are 31 occurrences of TAR template examples. (15) ich [A] ihr [NMr] gesagt habe daß... I have told her that... Çevdet, “Pullover”, comment 678 In phase 2, ThAR template examples increase to 29. (16) ich [A] gebt mir [NMr] oder die geld oder die neue pullover [U] I give me either the money or the new jumper Angelina, “Pullover”, comment 616 Again in phase 2, sagen is the most frequent verb in the data of learners, with the result that, although “one argument + object clause/ direct speech quotation” strategy is less frequent than in phase 1, this is the preferred strategy for learners. On the other hand, the percentages of use of two and three arguments realization change. As a matter of fact, the number of occurrences of TAR template increases a little bit with respect to phase 1. 23% 35%

OAR Obj. Cl./ Dir. Speech OAR TAR ThAR

25% 17%

Graphic 2: percentages of occurrences of templates.

Elena Ossella De Filippo

253

Phase 3 Learners use six three place verbs in phase 3: bringen (to bring/take), erzählen (to tell), fragen (to ask), geben (to give) leihen (to lend) and sagen (to say). There are 183 occurrences of ditransitive verbs, which are thus distributed: 13 bringen, 2 erzählen, 7 fragen, 16 geben, 3 leihen and 146 sagen. bringen

erzählen fragen

geben

leihen

sagen

Angelina 4

0

0

1

0

18

Tino

2

0

2

4

0

24

Marcello 0

0

3

1

0

9

Ayshe

3

0

0

3

0

38

Ilhami

1

2

0

2

1

29

Çevdet

3

0

2

5

2

28

Table 4: verbs in phase 3. OAR template occurs 90 times in data of this phase. The privileged syntactic argument is realized in 94% of examples. The number of “one argument (+ADJ)” decreases to 13 from phase 2. (17) ich [A] erkläre I explain Ayshe, “Modern Times”, comment 461 The number of sentences with one argument and an object clause or direct speech quotation increases to 77 occurrences. (18) eine mann [A] sage gibt mir eine stück von hols a man says give me a piece of wooden Angelina, “Modern Times”, comment 504 There are 37 examples of TAR template in interlanguage data: (19) er [A] kann vielleicht eh sein fahrrad [U] geben maybe he can give his bicycle

Case, Word Order and Thematic Role

254

Marcello, “Fahrrad”, comment 210 In phase 3, there is a doubling of ThAR occurrences, which increase to 60 examples. (20) ich [A] geb ihne [NMr] mein telefonnummer [U] I give you my telephone number Ayshe, “Pullover”, comment 548 Even though “one argument + object clause/ direct speech quotation” is the template used more frequently by learners in phase 3, “one argument (+ ADJ)” template use definitely diminishes in interlanguage data, and TAR template percentage use is stable in phase 2 and 3. ThAR template occurrences increase up to 32%, becoming the second more used strategy.

32% 41%

OAR Obj. Cl./ Dir. Speech OAR TAR ThAR

20%

7%

Table 5: percentages of occurrences of templates.

4.1 Argument Realization and Case Alternation According to whether the verb is realized with an OAR, a TAR or a ThAR template, arguments undergo case assignment. The assignment of case markers is based on two possible strategies, which are frequent and highly productive according to the time of the interview: word order strategy and thematic role case assignment. Word order strategy consists of the assignment of the nominative morpheme on the first element before the verb in the sentence and of the accusative morpheme on every element after the verb or after the argument marked with nominative morpheme (Jordens 1983; Putzer 1994; Tracy 1986) 3 . Case assignment based on 3

According to Tracy, the canonical word order in German as L1 is NPnomVNPACC (Tracy 1986: 59). Learners of GSL use the same word order (Putzer 1994: 255; Jordens 1983: 189; Pishwa 1989: 89).

Elena Ossella De Filippo

255

thematic roles consists of the assignment of nominative to agents, accusative to patiences and dative to beneficiaries (Jordens 1983; Putzer 1994). Word order is a sort of mechanicalism for the learner, while case assignment based on the thematic role is connected also with semantics. In accordance with the strategy used, case assignment is subjected to alternation. OAR templates generally imply the use of NOM markers on the only present argument, which is an actor most of the time. Case assignment for TAR templates is a little bit more complex, since the second argument can be both an undergoer and a non nuclear core argument (a non macrorole argument). Thus involves that arguments in TAR templates can be realized with a NOM-ACC or with a NOM- DAT alternation. As far as NOM- DAT alternation is concerned, the second argument is always considered as a non macrorole argument. Personal pronouns are more subjected to be marked with dative case 4 . In phase 1 there are 19,4% of occurrences of TAR with NOM- DAT alternation: (21) dann er [A] hat gesag für de Mädschen [NMr] wenn sie jetzt gehn würde... than he has told the girl if she would like to go now... Ayshe, “Modern Times”, comment 598 In phase 2 the percentage of occurrences is 19,4%: (22) ich [A] erzähl dir [NMr] aber vielleich is besser wenn du I tell you but maybe it is better if you... Marcello, “Pullover”, comment 000 In phase 3 percentage of NOM- DAT marking is 23,28%: (23) wenn ich [A] von jemanden [NMr] leihe if I lend to someone Çevdet, “Fahrrad”, comment 81 The accuracy of the use of dative case in TAR templates on NPs increases over phases. As a matter of fact, in the last phase the occurrences of dative

4

For further reading, see also Dittmar (1997) and Tracy (1986).

Case, Word Order and Thematic Role

256

marking on NPs double with respect to dative marking on personal pronouns. 14 12 10 8

Personal Pronoun

6

NP

4 2 0 Phase 1

Phase 2

Phase 3

Graphic 3: Occurrences of NOM –DAT with personal pronoun or NP NOM- ACC alternance is the most productive for TAR template. There are two types of NOM- ACC alternances: NOM-ACC alternance and NOM*ACC alternance. In the latest, the accusative case marker realizes an argument that should be a non macrorole argument (that is generally marked with DAT) in Standard German, but that is realized as an undergoer in the interlanguage of learners. Assignment of macroroles is most of the time target like for NOM – ACC alternance. In phase 1 the percentage of use of NOM- ACC marking is 52.7%: (24) er [A] gibt wieder diese karte [U] he gives these documents again Ayshe, “Modern Times”, comment 551 In phase 2 occurrences of NOM- ACC are 66.6% of data: (25) ich [A] muß die differenz [U] geben I must give the difference Tino, “Pullover”, comment 621 In phase 3 the percentage of use is 57.5%: (26) ich [A] hab das alles [U] erzählt I have told all of that Ilhami, “Fahrrad”, comment 474

Elena Ossella De Filippo

257

The third kind of case alternance for TAR template is NOM - *ACC marking. This implies that a non macrorole is betrayed as a macrorole, an undergoer, and it is marked with accusative case. In phase 1 occurrences are 27.7% of data: (27) er [A] sage die mädchen [U] wir wollen... he says (to) the girl we want ... Tino, “Modern Times”, comment 260 In phase 2 the percentage of NOM- *ACC marking use is 13.8%: (28) der [A] hat gesag den geschäftsmann [U] des war diese mädschen nischt der mann he has told to the shopkeeper that was this girl not the man Ayshe, “Modern Times”, comment 449 In phase 3 the percentage of occurrences is 17.8%: (29) ich [A] sage die andere leute [U] I tell to other people Angelina, “Fahrrad”, comment 33

50

43

40 31

30 20 10 0

NOM-ACC

19 10

7

5 7

Phase 1

Phase 2

13

17

NOM-*ACC NOM-DAT

Phase 3

Graphic 4: Nom-Acc/Nom-*Acc/Nom-Dat alternations With ThAR template, all arguments are realized, so that, as far as Standard German language is concerned, all case markers are supposed to be used by the speaker. In GSL, there are two possible case alternation: NOM-ACC-DAT or NOM-ACC-ACC.

258

Case, Word Order and Thematic Role

For NOMINATIVE – ACCUSATIVE – DATIVE alternation there are two possible word order settings, according to whether the non macrorole core argument is a noun phrase or a personal pronoun. Word order is SVOdirOind if dative case marks a noun phrase. In phase 1 the 18,2% of NOM- ACC –DAT frames occur with a noun phrase as non macrorole argument: (30) de polizei [A] aber gebt eine brief [U] in der mann [NMr] the police give a letter to the man Angelina, “Modern Times”, comment 354 In phase 2, NPs marked with dative case are 12.5%: (31) die polizei [A] haben gegeben eine brief [U] zu charlie chaplin [NMr] the police have given a letter to charlie chaplin Çevdet, “Modern Times”, comment 353 In phase 3, the occurrences of NPs marked as a non macrorole argument increase to 50%: (32) charlie chaplin [A] hat das brot [U] dem mann [NMr] gegeben charlie chaplin has given the bread to the man Ayshe, “Modern Times”, comment 461 Word order is SVOindOdir if dative marks a personal pronoun: In phase 1, learners use personal pronouns 82.8% of times: (33) isch [A] gebe dir [NMr] ein brief [U] I give you a letter Ilhami, “Modern Times”, comment 43 In phase 2 the percentage of dative marking on pronouns is 87.5%: (34) chaplin [A] brings her [T.en.] [NMr] die brot [U] chaplin brings her the bread Marcello, “Modern Times”, comment 89 In phase 3 the number decreases to 50%: (35) eine mann sache gibt mir [NMr] eine stück von hols [U] a man says give me a piece of wooden

Elena Ossella De Filippo

259

Angelina, “Modern Times”, comment 504

10 8 6

Personal Pronoun

4

NP

2 0 Phase 1

Phase 2

Phase 3

Graphic 5: Occurrences of NOM – ACC – DAT with personal pronoun or NP. Regarding NOMINATIVE – ACCUSATIVE – ACCUSATIVE marking, the only possible word order shown by data is SVOindOdir. As far as this kind of case frame is concerned, ditransitive verbs in the data of GSL interlanguage show the same behaviour of three place verbs structure of English marked undergoer choice. Namely, the undergoer is not the lowest ranking macrorole in conformity with the Actor- Undergoer hierarchy, but it is the second highest ranking one. Consequently, the undergoer is the animate argument, i.e. the benefactive. This kind of case marking decreases with time. In phase 1, the percentage of the NOM- ACC- ACC marking frame is the 44,4%: (36) der charlie chaplin [A] die kinders [U] geschenk gegeben charlie chaplin gives a present to the children Ilhami, „Modern Times“, comment 92 In phase 2, occurrences increase to the 66,6%: (37) der [A] hat die kinder [U] auch schokolade gegeben he has given the children also chocolate Ayshe, “Modern Times”, comment 501 In phase 3, NOM- ACC- ACC marking frame occurrences decrease to 25%: (38)

Case, Word Order and Thematic Role

260

ich [A] kann nur sie [U] eine zigarette geben I can give you only a cigarette Tino, “Fahrrad”, comment 260

20

18

15 10 5 0

8 5

4

4

NOM-ACC-DAT 6

NOM-ACC-ACC

Phase 1 Phase 2 Phase 3

Graphic 6: Non-Acc-Dat/Nom-Acc-Acc alternation Although these case assignment strategies are the marked choice in RRG, since the undergoer macrorole is assigned to the second highest ranking argument, NOM-ACC-ACC and NOM-*ACC marking frames are the preferred choices for learners, especially the beginning. According to the time of the interview, data show that occurrences of the NOM – ACC – DAT case frame with NPs and personal pronouns in dative case (Nom-Acc-Dat) increase with the passing of time, with respect to occurrences of the NOM – ACC – ACC marking strategy with a SVOindOdir word order (WO) .

30 25 20 15 10 5 0

WO Nom-Acc-Dat

Phase 1

Phase 2

Phase 3

Graphic 7: accusative/dative alternation through phases.

Elena Ossella De Filippo

261

Since phase 1, dative case is assigned to the benefactive (Putzer 1994; Jordens 1983). So that, the increasing of occurrences of NOM- ACCDAT case frame cannot be ascribed to the improved accuracy of dative marking on a specific thematic role.

40

32

30 20

16

Patience

15

Benefactive

10 0

0

3

0

1° Trim.

2° Trim.

3° Trim.

Graphic 8: Animate and inanimate arguments realized by accusative case On the contrary, accusative case is initially assigned both to patient and to benefactive roles (i.e. the highest and the lowest ranking argument). As a matter of fact, in early phases of acquisition, accusative case is used to realize all arguments, except for the PSA. Then, learners delimit the use of accusative morpheme to the patient. 60 50 40 30 20 10 0

54 42 28

22 19

20

1° Trim.

2° Trim.

Patient Benefactive

3° Trim.

Graphic 9: benefactors and patients realized by accusative case Therefore, the choice of the second highest ranking argument as undergoer is not due to a “marked choice” of the learner, but to the word order strategy, which leads to marking every argument after the verb with the accusative.

262

Case, Word Order and Thematic Role

5. Conclusion Learners of German L2 usually realize the arguments of three place verbs with four types of templates: “one argument (+ ADJ)” realization, “one argument + object clause/ direct speech quotation” realization, “two arguments” realization and “three arguments” realization. “One argument + object clause/ direct speech quotation” realization is the most frequent strategy used by learners. Occurrences of “two- and “ three arguments” realization increase especially in phase 2 and phase 3. With reference to TAR and ThAR templates, there are two case marking strategies: word order and thematic role case assignment strategies. In phase 1 and phase 2, the learner generally uses WO strategy, since he/she assigns accusative case marker on every argument after the verb, so that the order of the arguments of the predicate is SVOindOdir. In phase 3, learners assign case markers according to the thematic role. That is, dative preferably marks benefactives, nominative marks agents and accusative marks patients. Consequently, the order of the arguments of the predicate is SVOdirOind. The use of the dative case on benefactives increases as the production with the accusative case is limited by the learners only to patients.

6. Bibliography Diehl, E. (1991): Lernstrategien im Fremdspracherwerb: Untersuchungen zum Erwerb des deutschen Deklinationssystem. Tübingen: Niemeyer. Dittmar, N. (1997): „Die Rolle der Kontrastivität beim Fremdspracherwerb“. In Deutsch als Zweit und Fremdsprache edited by N. Dittmar & M. Rost-Roth. Frankfurt am Main: Lang: 107- 139. Dowty, D. R. (1979): Word Meaning and Mountage Grammar. Dodrecht: Reidel. Jordens, P. (1983): Das deutsche Kasussystem im Fremdspracherwerb. Tübingen: Narr. Klann – Delius, G. (1999): Spracherwerb. Stuttgart – Weimar: Metzler. Nakamura, W. (1997): A Constraint – Based Typology of Case Systems. M.A. thesis. Buffalo: State University of New York. Perdue, C. [ed] (1993): Adult Language Acquisition: cross linguistic perspectives, vol. I: fields and methods. Cambridge: CUP. Pishwa, H. (1989): Erwerb der deutschen Kongruenzregel. Berlin: TUB. Putzer, O. (1994): Fehleranalyse und Sprachvergleich. Ismaning: Hueber.

Elena Ossella De Filippo

263

Stenzel, A. (1994): “Case Assignment and Functional Categories in Bilingual Children – Routes of Development and Implications for Linguistic Theory”. In Bilingual First Language Acquisition, French and German grammatical development edited by J. M. Meisel. Amsterdam: Benjamins: 161 – 208. Tracy, R. (1986): “The acquisition of case morphology in German”. Linguistics, Vol. 24 (2005): 49- 78. Van Valin, R. J. (2002): The Role and Reference Grammar Analysis of Three- Place Predicate. New York: Buffalo University. ___ (2004): Lexical Representation, Co – Composition, and Linking Syntax and Semantics. New York: Buffalo University. ___ (2005): Exploring the Sintax- Semantic Interface. Cambridge: Cambridge University Press. Van Valin, R. & R. Lapolla. (1997): Syntax: Structure, Meaning and Function. Cambridge: Cambridge University Press.. Vendler, Z. (1967): Linguistics in Philosophy. Ithaca: Cornell University Press. Westheide, H. (1981): Grenzen des Zweitsprachenerwerbs? Eine linguistiche Analyse. Frankufurt am Main: Lang. Woolford, E. (2001): “Case Patterns”. In Optimality-Theoretic Syntax edited by G. Legendre, J. Grimshaw, S.Vikner. Cambridge: MIT: 509543.

PROFILING CHILD ESL ACQUISITION: PRACTICAL AND METHODOLOGICAL ISSUES KRISTIN KERSTEN Immersion programs have been claimed to be the most effective educational programs for the acquisition of a second language. This study focusses on ESL data from an immersion elementary school in Germany, which are analyzed within the framework of Processability Theory (PT, Pienemann 1998, 2005) and subsequently compared to PT data from naturalistic L2 acquisition. The paper puts a special focus on methodological issues of data analysis, especially with regard to coding decisions resulting from the form-function interface of linguistic structures. A fine-grained analytical grid is suggested, which is mainly based on the work of Pienemann (1998) and Pallotti (2003, 2007). The results indicate that, after four years of immersion schooling in a monolingual German environment, the participants in the program reached the final two stages of the processing hierarchy suggested by PT (stages 5 and 6) in L2 English, and are thus comparable to children learning English as a second language in a naturalistic context. With regard to data elicitation, it could be demonstrated that the profiling procedures suggested by PT can also be applied, with some limitations, to data sources not directly related to the PT framework.

1. Introduction This paper focusses on data from an immersion (IM) elementary school in Kiel, Germany, in which monolingual German children aged 6-11 are instructed in English in almost all parts of the curriculum. The immersion method has been called the most successful educational program for second 1 language acquisition in schools (Genesee 1987, Wesche 2002, Wode 2004, cf. also results from a comparative study by Pienemann et al. 2006 2 ). In order to shed more light on the effectiveness of such programs,

1 2

This article will not differentiate between the terms second and foreign language. Pienemann, Keßler & Liebner (2006) present data they collected at an immersive primary school in Kiel-Altenholz for a comparative analysis of different school programs.

266

Profiling child ESL acquisition

semi-spontaneous oral narratives from a four-year longitudinal study with four children of the Kiel immersion school, originally elicited in the framework of temporality and narrative structure (e.g. Berman & Slobin 1994, Dietrich et al. 1995, for an overview on temporal semantics cf. Bardovi-Harlig 2000), are profiled with the help of acquisitional stages for L2 English as suggested by the profiling hierarchy within the framework of Processability Theory (PT, Pienemann 1998, cf. also online information by Pienemann et al. on Rapid Profile 3 ). Developed by Manfred Pienemann in his influential 1998 book, PT is based on the premises of Levelt's (1989) model on language production and relies mainly on Kempen & Hoenkamp's (1987) Incremental Procedural Grammar and Kaplan & Bresnan's (1982) Lexical-Functional Grammar (for a concise introduction to the theory see Pienemann 2005, chapter 1). A special emphasis of the discussion in this paper lies on the criteria used for data coding. In the first section of this paper, the program, the data elicitation procedure, and the research questions will be introduced. The next part will describe the structures the analysis focusses on, and discuss other theoretical and methodological issues relevant to the coding of the data, especially with regard to the form-function interface in the acquisition of grammatical structures. Afterwards, the operationalization of criteria used for the analysis will be discussed in detail. The following two sections will present the results of the analysis and relate them to attainment levels of naturalistic L2 learners of English (Pienemann & Mackey 1993).

2. The study 2.1 The immersion project The data presented here were collected in a bilingual elementary school, the Claus-Rixen Grundschule in Kiel, Germany. The school is one of several educational institutions, ranging from preschool to secondary levels (ages 3-18), which are monitored under the supervision of Henning Wode at Kiel University (see Kersten 2005, Wode 2001 for more information on the Kiel Bilingual Project). The elementary school incorporates a partial immersion program. Bilingual classes (grades 1-4) mainly composed of monolingual German children from a middle class background are taught in English in all

3

http://groups.uni-paderborn.de/rapidprofile

Kristin Kersten

267

content matter except for German language arts. This amounts to approximately 70% of L2 input throughout the curriculum.

2.2 Data elicitation procedure The present analysis is part of a larger study which was designed to elicit guided, semi-spontaneous oral narratives in the framework of the acquisition of temporality based on a comparative research study by Berman & Slobin (1994). The data are currently being analyzed with respect to the acquisition of narrative structures (e.g. Möller 2006) and verbal morphology in the framework of temporal semantics (Kersten 2007, 2008). Data was elicited longitudinally over four years in the children’s L1 and L2 at the end of each grade (grades 1-4), and crosssectionally with an L1 English-speaking comparison group 4 from an elementary school in White Bear Lake, Minneapolis, in the USA, grades 1-4. Subjects: The analysis of this study was carried out on the samples of four subjects of the L2 data set, i.e. subjects 03, 06, 07, and 08. For each of the children, one test is available at each grade level. All subjects are female and started learning the L2 at age 6, with the exception of child 06, who had prior experience with the L2 in a bilingual preschool. Method: A picture story was used for data elicitation. 5 During each elicitation, the narration was delivered twice in the L2, first with a German-speaking interviewer whom the children were able to ask for vocabulary (Test A), and subsequently with an interviewer who was known to the children as a native speaker of English (Test B). During the second test, no questions were permitted. Both versions provided data for this study. The grade 4 data also include a short interview on personal stories preceding the narration. The tests were audio- and videotaped and subsequently transcribed. The samples average about 7:50 min of recording.

4

The term control group is avoided since it implies that all variables except for the dependent one are kept constant. As this was not possible at the time of data collection in the USA, I prefer the term comparison group to indicate that some, but not all of the variables are comparable. Comparable variables include the grades and the age of the children. 5 Frog, where are you? (Mayer 1969), cf. Berman & Slobin (1994).

268

Profiling child ESL acquisition

Coding conventions: Direct repetitions within the narrations (e.g. where is my frog, where is my frog), repetitions of interviewers' utterances (up to 3 turns), hesitations and self-corrections, and uninterpretable elements such as tokens with unclear endings were excluded from the analysis.

2.3 Research questions The developmental stages suggested by PT are used to evaluate the effectiveness of language acquisition in an IM educational program. As the data were collected in a different theoretical framework, the first step of the analysis will focus on the applicability of PT to this specific set of child data. Only after having established the coding criteria for the analysis can the results be compared to naturalistic L2 acquisition. Thus, the study will focus on the following research questions: 1. Do the data confirm the stages predicted by PT? 2. What are the operational criteria for an application of PT to this data set? 3. Are the results of L2A in an immersion school as indicated by the PT stages comparable to results achieved by learners from a naturalistic learning context?

3. Theoretical and methodological considerations The operational methodology applied here is based on Pienemann (1998) and includes criteria suggested by Pallotti (2003, 2007), 6 who, in his recent work, sought to operationalize the coding criteria presented in Pienemann (1998) for his application of PT to Italian. The following section introduces the selection of structures focussed on in the analysis, and discusses some problematic issues which have emerged in the process of data interpretation.

3.1 The structures The classification of linguistic structures into six developmental stages of ESL identified by PT is based on Pienemann (1998), Rapid Profile, and on Pienemann & Johnston (1987) and will not be repeated here. In order to

6

I am very much indebted to Gabriele Pallotti for helpful comments and discussions on the first draft of this paper. For a full account of his work on the operationalization of the emergence criterion see Pallotti (2007).

Kristin Kersten

269

use PT as a profiling measure for data not elicited in the PT framework, it is important to identify those structures which occur with a relatively high frequency and thus increase data density. stage syntactic structures 6 cancel inv 5 4 (wh-)copula inv yes-no inv part-verb 3 do-front topical wh-front adv-front 2 SVO neg+V

morphological structures 3.sg –s

aux+ing aux+en poss pro/det obj.pro past reg past irreg IL-ing plural –s

1 single wds Table 1: PT-Structures Verbal inflections are likely candidates since they appear in almost every clause of the narrations. Table 1 gives an overview of the syntactic and morphological structures coded for analysis. Irregular past was classified together with regular past and IL (interlanguage) -ing as stage 2 since no exchange of phrasal information is involved. The two auxiliary constructions not present in Rapid Profile were classified as stage 3 because of the unification of values between auxiliary and the lexical verb, which is an instance of phrasal agreement (Pienemann 1998:175). Note that this does not require the target-like use of the auxiliary. The targetlike agreement of auxiliary and subject is not relevant here; otherwise the structure would have to be placed at stage 5. To illustrate this with examples from the data: Child 06.1: and the dog falling down off the window and the boy are looking to a tree

(IL-ing, stage 2) (aux+ing, stage 3)

Since there were too few occurrences of do-front and aux 2nd in the sample they were excluded from the analysis.

270

Profiling child ESL acquisition

3.2 Form-function interface and the emergence criterion: some problem cases 7 Research in SLA has repeatedly shown that a learner's IL-system is a system in its own right with its individual rules, which do not necessarily need to correspond to those of the target language (Pienemann 1998: 138f, 160f). Especially in the beginning stages of acquisition, investigations into the distribution of linguistic features run the risk of confusing the emergence of interlanguage rules with random variation and chunks (Housen 1995, Perdue 1993, Pienemann 1998 chapter 4, 117ff). Such confusion can be avoided when researchers explicitly define "what forms and what functions are considered evidence for the emergence of a certain structure" (Pallotti 2003:1). The emergence criterion is a concept crucial for data analysis within the framework of PT. For an analysis of learner language, PT proposes a methodological operation which is based on the criterion of the first emergence of morpho-syntactic elements in the language produced by a learner. More specifically, Pienemann characterizes emergence as the "first systematic use of a structure" in question (1984:191). This refers to the moment "at which certain operations can, in principle, be carried out" (1998:138). In recent years, Pallotti has further developed the definition of the term emergence (2003, 2007). In his 2007 article, he seeks to operationalize the different components of emergence as quoted above to validify data analyses based on this criterion. The relation of form vs. function of a structure becomes especially important in two aspects of his definition, i.e. the notions of the productive 8 and the systematic use of grammatical morphemes: The learner may in fact be supplying the morpheme randomly, with no clear function, in free allomorphic variation. In this case, one would not say that a systematic form-function association has emerged, but rather that the learner is still experimenting with a phonological form. A criterion must specify a way of differentiating such cases from systematic uses. ... Only when a form begins to be used with a specific, selective function can one conclude that a rule has 7

I am grateful to Satomi Kawaguchi for comments on this issue in an earlier draft of this paper. 8 Examples for the productive use of a structure are minimal pairs or creative constructions. In a minimal pair of the plural morpheme, the -s will be attached complementarily to nouns in plural context, thus indicating a functional use of the morpheme.

Kristin Kersten

271

emerged. This point is also made by Pienemann (1998: 126). (Pallotti 2007:371f)

It has to be pointed out that the term function should refer solely to the grammatical funtion of a linguistic element as opposed to a conceptual or semantic function in a specific context. To give an illustration of this distinction: if the V-s inflection is used by a learner in many different contexts, most of which do not refer to the grammatical function of 3rd pers. sg., V-s cannot be interpreted as having emerged in terms of a systematic use of the structure. The application of V-s rather seems to be used randomly in the data. However, this example only relates to the suppliance of the grammatical function of 3rd pers. -s. The temporal (i.e. conceptual) function of present tense which the -s inflection carries as well should not be taken into consideration in the analysis, as PT does not make any claims about conceptualization but only about the processing load in terms of syntactic and morphological grammar formulation. This distinction will be discussed in more detail below (section 3.2.3). While the description of morpho-syntactic forms seems to be a quite straightforward operation, the description of their various functions thus seems a somewhat more complicated matter. It is therefore necessary to carefully operationalize the criteria which lead to the conclusion that a structure has emerged in an interlanguage system with respect to the first systematic occurrence (Pallotti 2007). To account for these criteria, Table 2 shows a distributional analysis which takes into account the variables context, target-like types, and target-like tokens of a given element, as well as its under- and over-suppliance (i.e. the null-hypothesis). Table 2 is an example for two structures which have emerged in an interlanguage system, i.e. 3rd sg. -s and plural -s (subject 06, grade 1): To illustrate this distribution with the example of 3rd pers. sg. -s: In 59 contexts of 3rd pers. sg., the child uses V-s with 11 types (i.e. with different lexical verbs) and 21 tokens (i.e. total occurrences independent of the verb type). 38 contexts of 3rd pers. sg. are used without the -s inflection (i.e. as base form V-ø), which is counted as under-suppliance, but V-s is never used in a context other that the 3rd pers. sg., thus there is no instance of over-suppliance in this example. For this reason, the structure was assigned the status emerged in the analysis (cf. chapter 4).

272

Structure

Profiling child ESL acquisition

Contexts/target-like types/ target-like tokens 59/11/21 38 0 10/2/10 11/5/11 0 0

3.sg -s – under-suppliance 3.sg ™ s – over-suppliance -s ™ 3.sg others ™ –s pl -s – under-suppliance pl ™ s – over-suppliance -s ™ pl ™ without Table 2: Example of a distributional analysis (Child 06.1) While in Table 2 the form-function interface seems unproblematic, difficulties may arise in the interpretation of corpora like the one at hand with regard to other kinds of functions. The following sections will discuss the notions of grammatical vs. conceptual function and their relevance to a PT analysis with reference to the interpretation of reported speech vs. relative clauses as an example of the former notion, as well as to maturational and lexico-semantic influences on learner language as an example of the latter. 3.2.1 Conceptual vs. grammatical function The preceding section has hinted at certain difficulties in coming to terms with the form-function relationship which has to be taken into consideration when assigning the emergence criterion to interlanguage data. The term function is, in itself, not without problems, though. As already indicated earlier, it is necessary to tease apart different functional aspects of a linguistic element. Taking again the function of 3rd pers. sg. -s as an example, two different aspects can be distinguished: On the one hand, the inflection denotes subject-verb agreement. This specific exchange of inter-phrasal grammatical information pertains to the S-procedure in the Formulator of Pienemann's model (1998, 2005, based on Levelt's 1989 model of language production). On the other hand, the inflection marks present tense. In other words, one can differentiate between a grammatical and a semantic function (Tarone 1988) of the linguistic element. In the example of 3rd pers. sg. -s the grammatical function would be inter-phrasal agreement, and the conceptual or semantic function relates the respective

Kristin Kersten

273

event within a temporal sequence. Thus, in the tense/aspect system of a language it is the second, the semantic function, which comes into play (cf. also Huddleston 1993:80f). In Levelt's model of speech production (1989:9) which underlies PT, time reference is generated in the Conceptualizer (cf. also Pienemann 1998:76), whereas the diacritic feature for tense marking is part of the lemma information stored in the Lexicon. PT's predictions on the processing hierarchy focusses solely on the functioning of the Formulator (Pienemann 1998:74). It pertains fully to syntactic structures, whose functions are often described as "grammatical relations" (Huddleston 1993:7): there is no conceptual difference in saying Turn off the radio or Turn the radio off; nor in Where are you? vs. *Where you are? Thus, in the processing of syntactic structures, it is the grammatical function which indicates that a specific procedure is at work. Still, when it comes to data coding, not all syntactic structures in learner language are easy to interpret in this framework. An example will be discussed in section 3.2.2. In section 3.2.3, a more detailed discussion of form and function with reference to time and morphological tense-marking follows, to which the predictions of PT refer only in part. 3.2.2 Cancel inversion In his characterization of cancel inversion, Pienemann states that “word order phenomena observed in direct questions do not apply in the context of indirect questions” (1998:170). He gives the example I wonder whether/why/where (he had lunch yesterday). An example for this structure from this study is found in the data of child 08.3: 08.3 ... shouted: "Where is my frog?" ... shouted where the frog is.

The introductory verb shout is not quite target-like, but the opposition of both phrases points to a cancel inversion in the context of an indirect question in the second phrase. But consider the following example: 06.2 ... looks where the bees are. ... to see where is the frog.

274

Profiling child ESL acquisition

According to Greenbaum & Quirk (1990:298f), only cognitive verbs or verbs describing a mental activity introduce indirect questions. However, it seems obvious that, in 06.2, the child overgeneralizes the inversion rule from indirect contexts to relative contexts introduced by look and see. It is apparent in the data that the difference between indirect questions and other forms of subordination is not as clear-cut as theory would have it. The same wh-elements that introduce interrogative clauses can as well introduce relative clauses (Greenbaum & Quirk 1990:309, 367f). And Huddleston (1993:396) points out that the same fronting mechanisms are at work in both kinds of clauses. What adds to the confusion in the latter example of 06.2 is the fact that the antecedent is missing, i.e. the element which the relative pronoun/adverbial refers to. Huddleston (1993:396) calls these structures fused relative constructions. He points out that they even partially overlap with interrogative constructions (as e.g. in he spent what they gave him vs. she told me what they gave him, p. 404). Without going into too much detail, it obviously seems rather unlikely that the learners distinguish between relative and interrogative whelements in their interlanguage. It is much more probable to assume that once a lexical item is annotated for a certain grammatical function, in this case for inversion, it is initially overgeneralized to all contexts, and only later in development becomes differentiated with respect to different grammatical functions. This might be different, however, in contexts of relative clauses with antecedent. There are many instances of such full relative clauses especially in the data of the third and fourth year. The last section raises the question of whether these elements, too, should be classified as contexts for cancel inversion. Indeed, some instances of the corpus suggest that inversion seems to be an issue for the children even in full relative contexts: 07.4 ... a place where we don't find them ... a hole where he shouted in ... a Markt [market] where can you buy Chinese things 08.3 ... the little frog who he catched ... the little frog who has he catched

The question remains which factors influence this development. One factor which may come into play here is the complexity of the structure to be cancelled (i.e. copula-inversion vs. verb-complexes in 08.3). This might

Kristin Kersten

275

represent an example of an intra-stage development (compare Mansouri & Håkansson 2007). This analysis thus suggests that three types of relative contexts have to be differentiated, i.e. indirect, fused and full relative contexts. And for the learner, this differentiation seems not straightforward at all. 3.2.3 Past tense marking As stated above, in order to establish the first systematic use of a structure, the researcher has to operationalize the notions of first and systematic. In Pallotti's operational criteria for emergence, this is where the function of a morphological structure comes into play, illustrated above with the example of the 3rd pers. sg. -s-inflection, which requires the processing of inter-phrasal agreement. With respect to past reference, which will be discussed in this section, Pienemann states that the use of grammatical information ... proceeds without reliance on temporary storage. An example is the morphological marking of reference to time. The information about tense is contained in the verb lemma with the value 'past' for the diacritic feature 'tense'. This means that the diacritic feature in question is available in the same location where the morpheme for the marking of past (i.e. '-ed') has to occur and no information has to be deposited into any syntactic procedure to achieve this process. I call the resulting class of morphemes 'lexical'. Since lexical morphemes can be produced without phrasal procedures they can develop before phrasal procedures. (Pienemann 2005:11)

The -ed-inflection is thus processed by the category procedure and classified at stage 2 in the processing hierarchy. In the present data set however, the question arose as to when the morpheme V-ed could be classified as emerged according to the operational criteria used in the analysis. One problem came up with respect to establishing its systematic use, another in connection with "the marking of past (i.e. -'ed')". Both problems arise in the broad context of the form-function relation. In the following two sections, they will be illustrated with examples from the data and subsequently addressed with reference to the theoretical framework of PT. 3.2.3.1Developmental issues One reason which can render it difficult to establish the systematic use of verbal inflections especially in corpora of child narratives are maturational

276

Profiling child ESL acquisition

constraints. To establish the systematicity of use, a stable reference point in the linguistic context is indispensable. 16.1 B (Grade 1)

06.2 A+B (Grade 2)

and the boy is in the water and hears something and then the boy said to the dog quiet (dir) and then the boy looked behind the tree and there is his frog with another frog and babies and then the boy has a little baby from the mother and the daddy

and the boy fall on the floor ... and then the dog and the boy fell in the water and then the dog finds the beehive and [Hesitation: they thi/] the dog think that’s a ball and jumped because he(?) want to toys with the boy ... and there came many many bees(beens?) out and the boy looked in a hole in a tree and there cames a owl out and say huh huh ...

Table 3: Tense marking (L2 English) This may be, for example, the predominant tense of a narration. In narrative child language data, however, it becomes obvious that stable time reference is a specific competence which has to be developed over time. Tables 3-6 vividly illustrate this phenomenon. Table 3 shows a picture very familiar to everyone working with child language. A look at the temporal function of the verbs reveals a high amount of variation regarding tense marking. There are tense shifts within one sentence, within one verb type, and even two different tense markers on one and the same verb. The same is true for aspect marking (Table 4). 08.1 B (Grade 1)

08.2 B (Grade 2)

and the boy say au (dir) and then the boy looking in a hole and the bees fly away and then the dog are running away ... and then the boy staying on the stone ‘n then the boy are sitting on a deer ... n’ then the boy are say psst (dir)

and then the boy is looking in a hole in a tree and then he look above the trunk

Table 4: Aspect marking (L2 English)

Kristin Kersten

277

Again, we find a range of intra-individual variation: activites like fly and run are sometimes marked with a progressive marker, sometimes they are not; there is variation on the inflections of a single verb type, ect. Interestingly, however, this phenomenon is not restricted to the L2 stories. A similar distribution occurs in the L1 stories (Tables 5 and 6), which are taken from other transcipts in the corpus: 16.1 (Grade 1)

07.2 (Grade 2)

er ging zu der Schildkröte und bellte sie an und dann beißt die Schildkröte den Hund in seine Pfote und dann leckt er sich die Pfote und der kleine Junge nahm seinen Eimer und seine Schaufel und packte sie weg

dann läuft der Hund zur Schildkröte und bellt sie an dann kam der Junge dichter und dichter und dann biß plötzlich die Schildkröte den Hund in die Pfote

[and he went to the turtle and barked at her and then the turtle bites the dog in his paw and then he licks his paw and the little boy took his bucket and his spade and put [past] them away]

[then the dog runs to the turtle and barks at her then the boy came closer and closer and then, suddenly, the turtle bit the dog in the paw]

Table 5: Tense marking (L1 German) Tense shifts like went and bites, barks and came as in Table 5 are frequently found. The same is true for English as L1 (Table 6). Here, verbs expressing a rather punctual action like yell through a hole or call for the frog are expressed in the progressive. A time-shift is present in both examples as well, from is yelling to came, went running to climbs and so forth. The question arises as to the reasons for this distribution. Are the children, even in their L1, not capable of target-like tense distinctions? Are we observing an artefact derived from the specific task, i.e. the acquisition of a narrative competence in children? Do the children shift back and forth between narration style and picture description mode? In order to answer these questions, a much more detailed analysis would need to be carried out. Yet, unfortunately this is out of the scope of this paper. What becomes obvious, though, is the following problem: if this variation cannot be explained satisfactorily for the L1 stories, it is even more difficult to make claims about the function of verbal inflections in

278

Profiling child ESL acquisition

the L2, and thus to establish coding criteria for the systematicity of their usage. 25.1 U

04.2 U

then he looks for him and so does the dog then he’s barking at a beehive and the boy’s yelling through a hole then a squirrel came up and bited his nose ... and the dog went running, looking for the frog and then he climbs up a XXX rock and holds onto a branch looking for the frog and then he got stuck by a moose the dog’s barking

the dog slipped out the window the boy is holding the dog and the dog licked the boy now the boy is calling for the frog

Table 6: Tense and aspect marking (L1 English) Whether or not these observations are related to the elicitation task, the fact that the same variability can be observed in the L1 stories of children of the same age strongly hints at a specific cognitive-maturational development of that age. One strong argument in favor of this point is that the temporal switches from past to present gradually decrease in the transcripts from higher grades both in the L1 and the L2 – even though the elicitation task remained the same. 3.2.3.2 Lexico-semantic marking The second problem in establishing the first systematic use of the past inflection is the assumption that -ed indeed functions as "the marking of past" (Pienemann 2005:11). Even if we assume systematicity in the use of L2 inflections, these inflections do not necessarily need to be used – even though systematically – with the same function a native speaker would use them. It has been suggested that verbal inflections, in the early stages of the acquisition process, tend not to encode the grammatical functions ascribed to them at all. The so-called Aspect Hypothesis (AH) as formulated by Andersen & Shirai (1994) complicates the form-function issue further by suggesting that in the beginning verbal inflections are used to encode the inherent semantic properties of the verbs they are attached to rather than grammatical functions. This has been referred to as the lexical aspect of the verb. The categories of lexical aspect are distinguished with regard to their inherent expression of duration, telecity

Kristin Kersten

279

(including an endpoint or goal), and dynamic properties, i.e. state verbs, activities, accomplishments, and achievements (Vendler 1967, Andersen & Shirai 1994): Lexical aspectual classes Semantic features Punctual Telic Dynamic

States -

Activities +

Accomplishments + +

Achievements + + +

Table 7: Semantic features of the four categories of inherent lexical aspect (taken from Rohde 2002:137) According to the AH, English inflections encoding 3rd pers. sg, progressive, and past tense are complementarily affiliated with different verb-categories in that -ing co-occurs mainly with activity verbs, past and perfective inflections with so-called achievements and accomplishments, and 3rd ps. -s with state verbs. It is assumed that this distribution is, to some extent at least, also inherent in the input (Distributional Bias Hypothesis, Andersen & Shirai 1994), and that it grants the learnability of verb inflections in both L1 and L2 acquisition (Rohde 1997; see also Rohde 2002 for a concise overview of these issues). In a study by Kersten et al. (2002) and Kersten (2007, 2008) the AH was confirmed for the same data corpus in an analysis on lexical aspect (cf. Table 8). 9 As a consequence, in the L2 narrations, lexical aspect marking, which has a semantic connotation, has to be carefully distinguished from "the marking of past". It is much more probable that V-ed, in the beginning stages of language acquisition, is used by the learner to indicate a functional notion different from past tense.

9

The effect of this distribution was dominated, in data from grade 1, by the -ing inflection, which seemed to be used as a general verb marker of invariant form. It has been proposed that this could be due to its being more saliently marked than the other forms, to its frequency and multiple functions in the input, and/or due to its phonological resemblance to the German infinitve ending -en (Rohde 1997). Support for this idea can be found in the data. 07.1 B: "The boy rufen the frog." vs. "The boy rufing the frog." But in order to make specific claims about function, an analysis of this kind should be carried out for each single child.

280

Profiling child ESL acquisition

09.1 (Grade 1) Clause - the dog looking in the glass - and the boy looking on the tree - the dog fell down - and them felled the boy and the dog in the water - shut, the Eul [owl] hats shut the eyes

Aspectual Category activity activity achievement achievement state

Table 8: Semantic aspect marking (L2 English) The examples from the data corpus in Table 8 illustrate the predictions of the AH in that activity verbs are inflected with -ing, achievements with past morphology, and state verbs with -s (note the two creative forms of *fell-ed and L1 *hat-s [has], which indicate that the learners use the distribution of verbal morphology productively on forms which have not occurred in the input). To sum up, the coding difficulties in establishing the systematicity of verbal tense morphology in these L2 narratives are twofold: first, the amount of variation due to the cognitive development yields an unstable reference point in the children's L2 narratives; and second, other acquisitional principles seem to interfere with the functional use of morphological structures. Having illustrated these two phenomena in the language of the learners, it is necessary to recur to the original PT model in order to address the issues of systematicity and form and function of past reference, since the question remains as to when past morphology can be classified as having emerged in the corpus. As stated above, PT predictions pertain to the processing operations in the Formulator of the speech production model. The Lexicon, which contains the lemmata, is seen as a separate component. However, in the beginning stages of language acquisition, the Lexicon is not yet completely annotated (Pienemann 1998:76) and may thus lack specific diacritic features such as tense. The findings with regard to the Aspect Hypothesis even suggest that in interlanguage these annotations may differ from those of the target language and will eventually be changed in the process of further language acquisition. One theoretical aspect has to be borne in mind, however: The fact that the respective inflection does not refer to the same conceptual function in the Lexicon as in the target language does not influence the functioning of the processing procedure in the Formulator. The category procedure is at work in both cases, irrespective of whether the inflection carries a reference to the past or not. In the same sense the target-like past

Kristin Kersten

281

reference is irrelevant for the operationalization of systematic use, as long as the morphological form is used in a creative and productive way. The temporal function of the inflections in question can thus be neglected as long as specific analytical criteria are operationalized for variability and productivity of their use. These criteria are discussed in the following section.

4. Data analysis and interpretation In accordance with the previous discussion, the following section will outline the criteria used for this study based on the suggestions by Pienemann (1998) and by Pallotti (2003, 2007) in his definition of operational criteria for data coding.

4.1. Criteria There are several factors which can potentially undermine the validity of an analysis. The first factor to account for is insufficient evidence. The criteria must therefore include a sufficiently high number of contexts for each structure. For syntax, Pienemann (1998:145) has suggested >4. He concedes, however, that "some degree of ambiguity remains in this analysis when it comes to judging if the number of linguistic contexts is sufficient for a given rule to decide if the rule has been applied or not" (p. 146). Second, the exclusion of chunks has to be guaranteed, which can be achieved by a certain degree of lexical variability in the data. It can be argued that this is a necessary prerequisite for syntactic structures as well, because here the risk runs high that some frequently used structures are learned as formulas (Pienemann 1998:147). Finally, Pienemann mentions the occurrence of random hits. These can be accounted for by the number of over-suppliances or over-use of the structure in question (Pallotti 2003, 2007). Table 9 summarizes these criteria. threat to validity exclusion of insufficient evidence exclusion of cunks exclusion of random hits Table 9: Validity criteria

criterion o number of contexts o variability o overuse

Profiling child ESL acquisition

282

Pallotti 10 emphasises that the rate of over-suppliances (inferred from the matrix for a distributional analysis in Pienemann 1998:158) is the crucial factor for the exclusion of random hits, as illustrated in a hypothetical example of a distributional analysis (Table 10): Plural –s N+ -s N+ -ø

pl 8 15

sg 2 43

o number of -s 10, number of over-suppliances 2 = 20 %

p number of pl 23, number of under-suppliances 15

Table 10: Hypothetical example of a distributional analysis Under-suppliance of a yet unacquired linguistic form as represented in the columns is generally to be expected in interlanguage. To illustrate this: if a feature occurs five times correctly with different types, there is probably no difference in the status of its emergence, independent of whether there are 15 or 50 under-uses. It would, however, make a difference had there been 15 over-suppliances instead of two. According to this logic, the percentage of over-suppliance has to be taken as an indicator for randomness. The following cut-off points have been used for the rate of over-suppliance in the analysis: Over-suppliance target-like + overuse + target-like (+) target-like (–) target-like – target-like

=100 % t 75 % t 50 % t 25 % < 25 %

With reference to Pallotti's criteria, this study uses a fine-grained analysis including bracketed labels to indicate insufficient evidence. The operational criteria used in the analysis are presented in Table 11. There is of course some arbitrariness to any cut-off point in an analysis. Thus, this chart represents a work-in-progress list which has to be refined. st structure 6

10

cancel inv

contexts

under -use

over -use

?

inv

/

(personal communication)

variability examples from data / (types) comments ?

he wants to know where the frog is he looks where the frog is = fused rel

Kristin Kersten

st structure

5

aux/do 2nd

3.sg –s 4

wh-cop inv yes-no inv

part-verb

3

topical do-front

wh-front

adv-front

( )

* aux+en

( )

* aux+ing

contexts

under -use

over -use

283

variability examples from data / (types) comments

he sees in the glass where the frog is = rel context / t4 t4 or 3 incl. mp: where are you looking? / 1mp ™inv where is my frog going? (types of wh+aux) exclusion of be/have t5 3.sg™s s™3.sg t5 or 3 + 1mp/creat mp: he walks / they walk creat: past irreg+s context / t4 t4 or 3 incl. mp: where are you? / where is my frog? ™inv 1mp (types of wh+V) context t3 t3 or 2 incl. overuse: this can you inv™ have context 1mp ™inv mp: are you here? / is he (types of here? S+V) creat: the dog came the context / t3 t3 or 1 + beehive down ™movemt. 1creat (mp ?) (different verbs) objects and subordinate / / t3 t3 clauses (different elements) (mp ?) underuse: he go there? t4 context™ do-front t4 do-front ™context (do with diff. contexts) / Ntl t4 or 3 incl. underuse: he where is? t4 mp: where are you? / position 1mp within S, where is my frog? (types of (not final) wh+V) exclusion: clause-linking Ntl / t3 t3 or 2 + conj and then; there is position 1mp underuse: the frog in the within S, or 1 + 1 night go to his family (not final) creat mp: now the boy wake (types of up/now the frog is away adverbs) creat: downside, there are a lake *aux+en t5 of aux™en t5 or 3 + 1 underuse: is fallen/falled/fell, gone aux+en en™aux (= +Vø) creat creat: have goed = aux+*en IL-ing aux t5 of t5 or 3 + 1 aux past + aux present creat: are rufing = aux+ aux+ing ™ing creat L1-ing ist looking = L1-aux+ing

Profiling child ESL acquisition

284

st structure

2

contexts

under -use

over -use

variability examples from data / (types) comments

poss.

t5

context ™ poss

poss ™ context

t2

obj.pro

t5

context ™ obj.pro

t2

SVO

t4

SOV/VSO random distrib.

obj.pro ™ context /

neg+V (acc. to Rap.Prof)

t3

postverb. neg?

/

past irreg

t5 of V-past

(X-irreg)

t5 or 3 + 1 creat

past reg

t5 of V-past

(X-ed)

t5 or 3 + 1 creat

creat: comed = *past-ed camed = irreg+ed

IL-ing

t5of (aux+) V-ing t5

*past-ed irreg+ed irreg+s / (V-reg +irreg) /

the boy (is) not walking / don't walk underuse: he looks not? creat: don't can come/don't finds/didn't were exclusion of be/have creat: *full=fell = *irreg

(X-ing)

t5 or 3 + 1 creat

creat: the boy rufing = L1-ing

t4 (with varying constit.) t3 or 1 + 1 creat (types of V with neg.)

his dog underuse: put he hands on he nose exclusion of 2nd sg. he see them underuse: he see they

mp: frog / frogs t5 or 3 + 1mp/creat creat: childs, childrens / / / dog, boy, der frog 1 single wds t2 (X): unlikely to occur; ntl: non-target-like; ™: without; /: no evidence in the data or not possible; mp: minimal pair; creat: creative construction plural –s

pl™s

s™pl

Table 11: Operational criteria Following Pienemann, the number of required contexts is lower for syntax than for morphology. On the other hand, in syntax some structures run a higher risk of chunking than others, like for example the collocation of wh-words with auxiliaries (what is, what do etc.). This is why four instead of three contexts were used in such cases. To illustrate this: the non-targetlike position of the wh-form within the sentence (He where is?) can be considered as under-use of wh-front. There was no instance of oversuppliance in the data (and it seems difficult to think of any example for it). The notion of minimal pair was also adopted for syntax. An example from the data would be the productive alternation of the auxiliary with the same question word, as in Where are you? vs. Where is my frog?

Kristin Kersten

285

4.2 Illustration A combination of these criteria yields the results of a distributional analysis expressed as '+' and '–'. Table 12 gives an overview of the gradation system for emergence. In the case of positive contexts, the columns of negative evidence are disregarded. random hits variability contexts underuse status: status status status status emerged + + + (–)/(+) + (+) (+) – (+) (–) – + + – (–) – (+) + (–) / / / Table 12: Operational criteria: gradation of status "emerged" Table 13 illustrates how this analysis can be applied to the data. structure random hits variability (negative evidence) status: tl over-u % stat. tl min. creat. stat. contexts stat. under stat. emerged tokens (/cont) -use (t75%) types pairs constr. morphol. pl. –s 15 0 100 6 2 1 15 0 + + + IL-ing / / / p. reg 5 0 100 4 5 0 + – (+) p. irreg 4 0 100 (+) 2 1 6 2 (+) (+) poss 12 0 100 2 12 0 + + + obj.pro 5 0 100 1 5 0 + – (+) aux+ing 6 0 100 6 6 0 + + + aux+en 1 2 33 1 6 3 (–) – + + (–) 3.sg –s 44 1/11 98 16 5 1 56 12 + + + syntax SVO 59 59 0 + / + neg+V 4 4 + + + adv-front 3 3 + + + wh-front 7 2 1 7 0 + – (+) do-front / / / topical. 4 4 + + + part-V 1 1 (+) – (+) move yes/no / / / inv wh-cop 7 2 1 7 0 + – (+) inv aux/do / / / 2nd cancel / / / inv +: emerged; (+): insufficient evidence for emergence; (–): insufficient evidence against emergence; –: not emerged; tl: target-like

286

Profiling child ESL acquisition

Table 13: Operational criteria applied (Child 06.3) The table shows the analysis of the morphological and syntactic structures used in the narrative of Child 06 in grade 3 with regard to the operational criteria established above. To exemplify: in Table 13, the structure 3rd sg. -s is interpreted as emerged ('+') since there are a) enough contexts, b) random hits are excluded, and c) there proved to be enough variability with regard to the criteria. The structure aux+en has not emerged ('–') because there are not enough contexts, thus a high risk of chance, and not enough variability. The feature past irreg may have emerged ('(+)'), but there are not enough contexts and variability according to the criteria.

5. Results and discussion 5.1 Language development The following tables show an overview of the development of the four subjects from grade 1-4. Table 14 represents an implicational analysis of cross-sectional and longitudinal data of all four subjects. Some of the structures (bold print) which occur late in development have not enough linguistic contexts. But since they had already been confirmed in previous stories of the child, as for example neg+Verb with child 07 (confirmed in year 3), they were assumed to still be part of the child's interlanguage system (cf. continuity assumption, Pienemann 1998). 11 At a first glance, all structures are in line with the PT predictions, except for yes-no inversion and aux+past participle. In most of these cases the structures are already present in the data but there is not sufficient evidence according to the criteria. However, as long as other structures of that stage have emerged, this is a sign that the procedure actually is at work. Thus, there may simply be other factors which delay the acquisition of this specific feature as e.g. higher complexity, lower frequency in the input, etc. (cf. Mansouri 2005, Mansouri & Håkansson 2007 on intra-stage developments). Therefore they should not be considered counter-examples to the implicational pattern as long as one of the structures belonging to a certain stage has emerged. There are, however, many gaps in the data set indicating that it is not as robust as the two ESL corpora presented in Pienemann (1998). 11

"If a structure has been acquired it will be a constant part of the interlanguage system at later levels of development." (Pienemann 1998:147)

Kristin Kersten

stg. structure 3.1 7.1 8.1 6.1 3.2 6 cancel / / / / / inv 5 3.sg –s / / – (+) + 4 wh-cop / / / (+) (+) inv yes/no / (–) / (+) (–) inv part-V / / / / (+) 3 topical. / / / (+) / wh-front / / / (+) (+) adv-front / + + (+) + aux+en / / / / (–) aux+ing (+) (+) + + (+) poss (–) / (+) (+) (+) obj.pro / / / / (+) 2 SVO + + + + + neg+V / / / / (–) p. irreg / / / / (+) p. reg / / (+) / (+) IL-ing + + + + + pl. –s (+) (+) + + (+) 1 single / + / / / wds bold +: structure provides not enough previous years

287

3.3 7.2 6.2 6.3 3.4 8.2 7.3 7.4 8.3 6.4 8.4 / / / / / (+) (+) (+) + + + + + + + + + + + + + + (+) (+) (+) (+) (+) (+) (+) (+) (+) (+) (+) (+)

+

/ (+) (+) + (–) (+) + / + (+) (+) + + + /

/ (+) (+) + (–) / (+) + + + + (+) / + /

/

/

/

(–)

+ + / (+) (+) + + (+) (+) (+) (+) (+) / + + + (–) – (+) / +? + + + + + + (+) / (+) (+) / + + + + + + + (+) (+) (+) + + + + + (+) / / / + + + + + / / / /

+ (+) (+) (+) + – + (+) + + + + + / + /

+

(–)

/

/

(+) + + + (+) + + + (+) (+) (+) (+) + + + + (+) + (+) + + + / + + + + + + (+) + + + + + + + + + + + + + + + + + + / / / / + + + + / / / /

contexts in that data set but has been mastered in

Table 14: Implicational Analysis - Development (A+B) Table 15 shows that, when reduced to the different attainment levels, there is indeed enough evidence for each stage in almost every learner according to the operational criteria. stage 3.1 3.2 3.3 3.4 6 5 + + + 4 (+) (+) (+) 3 (+) + + + 2 + + + + 1 / / / / bold +: structure provides previous years

6.1 6.2 6.3 6.4 7.1 7.2 7.3 7.4 8.1 8.2 8.3 8.4 + (+) (+) (+) + + (+) + + + + – + + + + + (+) + + (–) + (+) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + / / / / + / / / / / / / not enough contexts in that data set but has been mastered in

Table 15: Attainment in development (A+B) Only child 03 does not provide enough evidence for stage 4 throughout all samples. Stage 4 however is confirmed in the data of all the other subjects. Data which are not specifically elicitated within the framework of PT run the risk of not providing enough contexts for all structures exemplified in the PT hierarchy. Therefore, the steps taken in the current approach

Profiling child ESL acquisition

288

were, firstly, the addition of some morphological features, secondly, a refinement of operational criteria as formulated by Pallotti (2003, 2007), and thirdly, the application of the continuity assumption in the reading suggested by Pienemann (1998:147). Insufficient evidence for the stages in Table 15 amounts to less than 14%, and none of it contradicts the theory.

5.2 Comparison with naturalistic child L2-acquisition To examine the effectiveness of the IM program in the Kiel elementary school, the data were then compared to a naturalistic child ESL study (Pienemann & Mackey 1993). The data are ordered with respect to length of residence in the ESL study, and time of input in the IM program respectively (Tables 16 and 17). LOR 6 7 8 8 8 9 11 11 (months) stage 1:4 1:1 1:7 1:2 1:6 2:1 1:3 1:5 6 + 5 + + + + 4 + + + + + 3 + + + + + + + + 2 + + + + + + + + 1 + / + / / / / / LOR Length of Residence (taken from Pienemann 1998:179)

11

22

22

44

60

2:2

2:4

2:6

2:5

2:3

+ + + + /

+ + + + /

+ + + + /

+ + + + /

+ + + + /

Table 16: Attainment in development: ESL study according to LOR The age of the children in both studies is comparable (8-10 in ESL vs. 610 in IM). The longest time of input is 60 months (five years) in the ESL study, as opposed to 46 months (four years) in the IM context. Both studies reveal a high amount of inter-individual variation, indicating that learners learn at a very individual pace, even given the same input situation in the IM program. TOI 10 10 10 10 mon stage 3.1 7.1 8.1 6.1 6 5 (+) 4 (+) 3 (+) + + + 2 + + + + 1 / + / / TOI Time of Input (months)

22

22

22

22

34

34

34

34

46

46

46

46

3.2 7.2 6.2 8.2 3.3 6.3 7.3 8.3 3.4 7.4 8.4 6.4 (+) (+) + (+) + + + + + + + + + + + + + + (+) + + (+) (+) + + (+) + + + + + + + + + + + + + + + + + + + + + + + + + + + + / / / / / / / / / / / /

Table 17: Attainment in development: IM study according to TOI

Kristin Kersten

289

In neither study the time of input seems to be predictive for ultimate attainment, but further factors need to be known to make claims about this. Stage 5 seems to be reached earlier by the naturalistic acquirers. However, after 4 years of input more IM children have reached the final stage than the naturalistic learners. It becomes obvious from this comparison that IM is a very efficient program of language teaching which yields results comparable to those of naturalistic L2 learners with respect to the emergence of linguistic features.

6. Conclusion To conclude, the study has shown that PT is a powerful instrument of analysis which is applicable also to data corpora with elicitation methods outside the PT framework. The operationalized criteria suggested in this paper proved to be helpful tools in analyzing the data. The results showed that all stages predicted by PT could be confirmed in the L2 narrations of German children learning English in an IM elementary school. These children's attainment level within the PT hierarchy is especially significant in comparison with L2 learners who learned the language in a naturalistic context. The study shows that children from an IM elementary program are capable of reaching comparable levels of attainment in L2 acquisition. Early IM can thus be considered a highly efficient program of language teaching. A comparison to results from other teaching programs would be interesting in this context. With regard to refinements of the profiling methodology it would be valuable for future studies to apply the coding criteria as used in this analysis to a larger corpus of L2 data.

References Andersen, R. W. & Y. Shirai (1994): "Discourse motivations for some cognitive acquisition principles." Studies in Second Language Acquisition, Vol. 16 (1994): 133-156. Bardovi-Harlig, K. (2000): Tense and Aspect in Second Language Acquisition: Form, Meaning, and Use. Oxford: Blackwell. Berman, R. & D. I. Slobin (1994): Relating Events in Narrative: A Crosslinguistic Developmental Study. Hillsdale, New Jersey: Lawrence Erlbaum. Dietrich, R. & W. Klein & C. Noyau (1995): The Acquisition of Temporality in a Second Language. Amsterdam, Philadelphia: John Benjamins.

290

Profiling child ESL acquisition

Genesee, F. (1987): Learning Through Two Languages: Studies of Immersion and Bilingual Education. Cambridge, MA: Newbury House. Greenbaum, S. & R. Quirk (1990): A Student's Grammar of the English Language. Essex: Longman. Housen, A. (1995): It's About Time. The Acquisition of Temporality in English-as-a-Second Language in a Multilingual Educational Context. Brussels: Vrije Universiteit Brussel. Huddleston, R. (61993): Introduction to the Grammar of English. Cambridge: CUP. Kaplan, R. & J. Bresnan (1982): "Lexical-Functional Grammar: a formal system for grammatical representation." In The Mental Representation of Grammatical Relations edited by J. Bresnan. Cambridge, MA: MIT Press: 173-281. Kempen, G. & E. Hoenkamp (1987). "An incremental procedural grammar for sentence formulation." Cognitive Science, Vol. 11: 201-258. Kersten, K. & C. Imhoff & B. Sauer (2002): "The acquisition of English verbs in an elementary school immersion program in Germany." In An integrated view of language development: Papers in honor Henning Wode edited by P. Burmeister, T. Piske, A. Rohde. Trier: WVT: 473497. Kersten, K. (2005): "Bilinguale Kindergärten und Grundschulen: Wissenschaft und Praxis im Kieler Immersionsprojekt." In Bilingualität im Kindergarten und in der Primarstufe. Bessere Zukunftschancen für unsere Kinder edited by P. Baron. Opole: Niemieckie Towarzystwo Oswiatowe: 22-33; also available as pdf: www.kristinkersten.de/media/Kersten$202005$20Bilinguale$20Kitas$20und$20 Grundschulen$20$28Publ$29.pdf Kersten, K. (2007): "The interaction of lexical aspect and discourse in learner language." Paper presented at EUROSLA 17, University of Newcastle. Kersten, K. (2008): "Der frühe Erwerb von Verbflexionen: L2 Bildergeschichten aus einer immersiven Grundschule". Paper presented at 5. Arbeitstagung für den wissenschaftlichen Nachwuchs in der Fremdsprachenforschung, University of Hannover. Levelt, W. J. M. (1989): Speaking. From Intention to Articulation. Cambridge, MA: MIT Press. Mansouri, F. (2005): "Conceptualising intra-stage sequencing in the learner language." Paper presented at the 5th International Symposium on Processability, Second Language Acquisition and Bilingualism, Deakin University, Victoria, Australia.

Kristin Kersten

291

Mansouri, F. & G. Håkansson (2007): "Conceptualising intra-stage sequencing in the learner language." In Second Language Acquisition Research: Theory-Construction and Testing edited by Fethi Mansouri. Cambride: Cambridge Scholars Press: 95-117. Mayer, M. (1969): Frog, Where Are You? New York: Pied Piper. Möller, C. (2006). "The development of narrative in German-English immersion." Paper presented at EUROSLA 16, Bo÷aziçi University, Antalya. Pallotti, G. (2003): "Cross-linguistic tests of Processability Theory: An operational definition of the emergence criterion." Paper presented at EUROSLA 13, University of Edinburgh. Pallotti, G. (2007): "An operational definition of the emergence criterion." Applied Linguistics, Vol. 28/3: 361-382. Perdue, C. (ed.) (1993): Adult Language Acquisition: Cross-Linguistic Perspectives. Cambridge: Cambridge University Press. Pienemann, M. (1984): "Psychological constraints on the teachability of languages." Studies in Second Language Acquisition, Vol. 6: 186-214. Pienemann, M. (1998): Language Processing and Second Language Development: Processability Theory. Amsterdam, Philadelphia: John Benjamins. Pienemann, M. (2005): "An introduction to Processability Theory." In Cross-linguistic Aspects of Processability Theory edited by M. Pienemann. Amsterdam, Philadelphia: John Benjamins. http://groups.uni-paderborn.de/rapidprofile/docs/INTRO.NEW.pdf (retrieved July 5, 2008) Pienemann, M. & M. Johnston (1987): "Factors influencing the development of language proficiency." In Applying second language acquisition research edited by D. Nunan. Adelaide: National Curriculum Research Centre, Adult Migrant Education Program: 45141. Pienemann, M. & J.-U. Keßler: Rapid Profile: Linguistic Profiling. Paderborn: University of Paderborn. http://groups.uni-paderborn.de/rapidprofile/ (retrieved Jan. 11, 2008). Pienemann, M. & J.-U. Keßler & M. Liebner (2006): "Englischerwerb in der Grundschule: Untersuchungsergebnisse im Überblick." In Englischerwerb in der Grundschule edited by M. Pienemann, J.-U. Keßler, E. Roos. Paderborn: Schöningh: 67-88. Pienemann, M. & A. Mackey (1993): "An empirical study of children’s ESL development and Rapid Pofile." In ESL development. Language and Literacy in Schools, Vol. 2 edited by P. McKay. Commonwealth of

292

Profiling child ESL acquisition

Australia and National Languages and Literacy Institute of Australia: 115-259. Rohde, A. (2002): "The aspect hypothesis and the L2 reacquisition of verbal inflections." In An Integrated View of Language Development: Papers in Honor Henning Wode edited by P. Burmeister, T. Piske, A. Rohde. Trier: WVT: 135-151. Rohde, A. (1997): Verbflexion und Verbsemantik im Natürlichen L2Erwerb. Tübingen: Narr. Tarone, E. (1988): Variation in Interlanguage. London: Edward Arnold. Vendler, Z. (1967): Linguistics in Philosophy. Ithaca: Cornell University Press. Wesche, M. B. (2002): "Early French immersion: How has the original Canadian model stood the test of time?" In An Integrated View of Language Development: Papers in Honor Henning Wode edited by P. Burmeister, T. Piske, A. Rohde. Trier: WVT: 357-379. Wode, H. (2001): "Multilingual education in Europe: What can preschools contribute?" In Language as a Tool: Immersion Research and Practices edited by S. Björklund. University of Vaasa: Proceedings of the University of Vaasa, Reports: 424-446. Wode, H. (2004): Frühes Fremdsprachenlernen. Englisch ab Kita und Grundschule: Warum? Wie? Was bringt es? Kiel: FMKS.

CONTRIBUTORS Malin Ågren, PhD, Lund University, Sweden. Camilla Bettoni, PhD, is professor of applied linguistics at Verona, Italy. She studied at Padua (Italy), and Edinburgh (Scotland), then worked at the Universities of North Queensland, and Sydney (Australia) for 17 years. Her main interests are language contact, bilingualism, and second language learning and teaching. Her recent publications include Imparare un’altra lingua (2001) and Usare un’altra lingua (2006), both published by Laterza in Rome. Bruno Di Biase, PhD, is Senior Lecturer in Second Language Acquisition, Linguistics and Italian, and is Associate Head of the School of Humanities and Languages, University of Western Sydney. His research interests are in Second Language Acquisition, particularly Processability Theory (with a current focus on the Syntax and Discourse-Pragmatics interface), as well as Bilingual First Language Acquisition and Applications of Language Learning Research to Pedagogy and Public Policy on Languages. Yuki Itani-Adams, University of Western Sydney, Australia. Satomi Kawaguchi, PhD, is Lecturer in Second Language Acquisition, Japanese as a second language and Research Methods at the School of Humanities & Languages, University of Western Sydney. Her main research interests are in second language acquisition particularly the interface between lexicon, syntax and discourse-pragmatics. She is also interested in theory-practice link in teaching and learning and implementation of SLA research in CALL (Computer Assisted Language Learning). Dagmar Keatinge works as a Researcher and Lecturer at Paderborn University, Germany. She worked on number of research projects in the field of second language acquisition, including bilingual language acquisition of children and L1- transfer (for her MA thesis). For her PhD thesis, she works on the topic “The interface between comprehension and production in second language acquisition”. Kristin Kersten, Magdeburg University, Germany.

Contributors

Jörg-U. Keßler, PhD, is professor of applied linguistics at Ludwigsburg Germany. He has worked on the evaluation and calibration of Rapid Profile, a computer-assisted diagnostic tool to trace L2 development. His research interests encompass (instructed) SLA and Immersion programmes from preschool to secondary education. In his current research he works on various applications of Processability Theory to the foreign language classroom. His publications include books on the acquisition of L2 English in formal settings and a range of articles on taskbased language teaching. Lucija Medojevic is a PhD candidate at the School of Humanities and Languages at the University of Western Sydney (Australia). She developed the first application of Processability Theory and its extension to the acquisition of Serbian language as a family and community language in Australia. Her ongoing PhD research is on the effect of the first year of school on Bilingual First Language Acquisition (from a processability perspective) with Serbian-Australian bilingual children. Elena Nuzzo is research fellow at the University of Modena and Reggio Emilia, and part-time assistant professor of Didattica della lingua italiana at Bergamo University. Her main interests are pragmatics and second language learning and teaching. Her recent publications include Imparare a fare cose con le parole (2007), published by Guerra Edizioni, Perugia. Elena Ossella de Filippo, University of Naples – Federico II, Italy. Yumiko Yamaguchi is a PhD candidate at the School of Humanities and Languages at the University of Western Sydney (Australia). Her ongoing PhD research is on the developmental sequences in Child ESL acquisition based on Processability Theory. Her doctoral workshop submission was awarded one of the five prizes in EuroSLA 17 held in Newcastle, U.K. in 2007 Kenny Wang is a PhD candidate at the University of Western Sydney. His current research focuses on the L2 English acquistion of noncanonical syntax by Mandarin speakers and Mandarin-English trainee interpreters. He holds a BA (Interpreting & Translation) and a BA (Honours) in Linguistics from the University of Western Sydney.