Date: Mon, 28 Jul 2003 04:56:23 +0000 From: Jenifer Larson-Hall Subject: Generative Approaches to the Acquisition of English by Native Speakers of Japanese
Wakabayashi, Shigenori, Ed. (2003) Generative Approaches to the Acquisition of English by Native Speakers of Japanese, Mouton de Gruyter, Studies on Language Acquisition
Reviewed by Jenifer Larson-Hall, Kyushu University, Fukuoka, Japan.
Using the syntactic framework of Chomsky's (1995) Minimalist Program and clarifications to it such as Hawkins (2001) and Radford (1997), the 8 articles in this volume explore syntactic issues in second language acquisition through a focused lens of Japanese learners of English. All of the articles include solid experimental data aimed at addressing the question of the nature of the initial state of interlanguage and whether Universal Grammar (UG) is active in adult second language (L2) acquisition. As a phonologist, I found all of the articles to be quite accessible and clearly spell out their syntactic premises.
Wakabayashi's introduction notes that the study of Japanese learners of English is helpful to the UG debate because English and Japanese are typologically distant and have differing values for many proposed parameters, such as the head parameter, wh-parameter, and null subject parameter, thus setting up a situation where it is easier to see what kind of influence the first language (L1) may have on L2.
Although all of the articles address the question of whether UG is active in L2 acquisition, not all of the researchers in this volume come to the same conclusions about the influence of L1 and/or UG. For example, some researchers conclude that L1 influence accounts for the patterns that are found while others find strong evidence that functional categories (morpho-syntactic differences among languages are considered to be largely associated with functional categories) are available to L2 learners.
SUMMARY AND EVALUATION
1. 'Japanese learners' acquisition of English motion verbs with goal PPs by Shunji Inagaki. Inagaki examines sentences such as 'John walked to school', 'John ran into the house', and 'John went to school walking' in this chapter. He tests whether Japanese speakers will accept such sentences, as Japanese does not accept manner of motion verbs (like run, walk) with goal PPs (*John-ga gakko-ni aruita). The differences between Japanese and English are attributed to different incorporation patterns in the syntax. A written grammaticality judgment task with pictures was used, and Inagaki concludes that Japanese learners succeeded in acquiring manner of motion verbs with goal PPs (John walked to school), suggesting that such argument structure is not difficult to acquire when positive evidence is available. However, Japanese accepted other sentences that NS found unacceptable, such as 'John went to school by walking' and 'John walked and went to school'. Inagaki suggests that without negative evidence, the Japanese cannot recognize that these sentences are unnatural in English.
2. 'Unaccusatives versus passives in L2 English' by Makiko Hirakawa. Hirakawa studies differences in case licensing by passive and unaccusative verbs. Cases where Japanese and English behave similarly and divergently are examined. The arguments are clearly explained but somewhat complicated to replicate here, so let it suffice to say that the difference lies in that English unaccusatives (such as 'arise' or 'exist') license partitive case optionally, so that there is a definiteness effect (so 'There arrived a man in Storrs' is OK but 'There arrived Uli in Storrs is bad) in English, but none in Japanese. On the other hand, both Japanese and English must raise the NP for passive sentences. Hirakawa found that learners showed sensitivity to partitive case assignment and the definiteness effect with 'be' but not with unaccusative verbs. This means that the learners *were* able to acquire subtle properties such as a sensitivity to the definiteness effect that were probably never taught explicitly.
3. 'The acquisition of the nominative and accusative cases in English by Japanese learners at an early stage' by Koji Suda and Shigenori Wakabayashi. This study investigates the acquisition of nominative and accusative cases of English pronouns. The authors had to test very young English learners (7th graders) to find Japanese who were willing to accept such sentences as '*He likes she'. But if the L1 grammar transfers to L2 grammar at the initial stage, this acceptance is predicted from the Japanese facts, since Japanese can say 'kare-ga kanozyo-ga sukida' (both NPs use a subject marker) as well as 'kare-ga kanozyo-o sukida' (one NP uses a subject marker, the other an object marker). Indeed, the authors found that '*He likes she' was accepted as the correct translation when preceded by the Japanese Sub-ga Obj-ga Stative Verb construction was used. However, the English Sub(Acc) Obj (Acc) case ('*Him likes her') was accepted as correct regardless of the Japanese sentence type that preceded it. Suda and Wakabayshi claim that neither a Full Transfer/Full Access position nor a Minimal Trees position (Vainikka and Young-Scholten, 1996) can explain these findings, and propose their own Lexical Learning/Lexical Transfer model, which has been previously explained in Wakabayashi (2002). Briefly, it states that: 1) Operations, such as Merge and Attract/Move, are innately given, and operate in L2 grammar; 2) L1 transfer takes place along with lexical learning.
4. 'The acquisition of a second language C-system by Japanese learners of English' by Tomohiko Shirahata. This report is a bit different from the others in the book in that it is based on two longitudinal observations. Shirahata focuses on the acquisition of C and Complementizer Phrase (CP) by Japanese child learners. Shirahata assumes that C in English interrogatives is strong so it overtly attracts constituents to move, while Japanese C is not and so nothing is overtly moved. To test whether the child learners have acquired strong English C, the author examines their production of C elements such as that/if/for, their formation of yes/no questions, and their formation of wh-questions. She finds that all 4 subjects produced complex sentences with that-clauses in month 3 or 4, and production of 'if' and 'for' occurred later. The children did not always produce the correct forms, but were able to make yes/no questions within several months. Finally, all the children produced wh-questions even during the first month. Shirahata concludes that these learners had little difficulty in acquiring the English C-system, which supports Lakshmanan and Selinker's (1994) claim that CP is available from the initial stage of L2 acquisition.
5. 'Pied-piping and stranding in oblique relative clauses in Japanese EFL learners' interlanguage grammars' by Hiromasa Ohba. This study's interesting conclusion is that Japanese learners have no trouble learning preposition stranding in English and that they do not transfer aspects of their L1. Ohba cites McDaniel, McKee and Bernstein (1998) in support of the idea that NS of English actually only develop preposition stranding in their mental grammar, and that pied piping of prepositions is a prescriptive artifact. Japanese has postpositions which are obligatory in questions and declarative sentences, but obligatorily omitted in relative clauses (sore-ga watashi-ga sunde iru ie desu/that's the house I live in). Ohba used a grammaticality judgment task (GJT) as well as a sentence-combining task with 3 levels of learners on relative clauses, and found that the learners judged both preposition stranding and pied-piping to be correct at levels over 60%. As for production, the learners tended to produce more pied-piping structures as their proficiency increased, but quite rarely omitted prepositions, as is obligatory in Japanese. Ohba concludes that no L1 transfer is taking place, and that Japanese adult learners' grammar starts with preposition stranding, as it does for NS. Ohba seems to claim that his learners have easily learned preposition stranding, but 60% accuracy on the GJT is not very high. This reviewer would have liked to see the question of why even advanced learners only accepted relative clauses in English with stranding at about 75% accuracy addressed. I also would have liked to see more statistics about group differences included.
6. 'Is an interlanguage a 'possible grammar'?: How Japanese speakers learn CP structures in English' by Chieko Kuribara. This paper is probably the most theoretically challenging of the collection. It addresses the question of whether L2 acquisition is based on UG or whether it involves higher-level cognitive functioning. Kuribara comes down squarely on the side of cognitive resources and the exploitation of non-UG knowledge. She investigates the acquisition of the wh-feature and hypothesizes that if learners have acquired this feature they should notice the ungrammaticality of a cluster of grammatical violations, such as what she calls *C-SpecCP ('*They couldn't tell us that what time they were finishing work') or *SpecCPx2 ('*How much money to whom do I have to pay?'). Subjects responded to sentences with either OK, not OK, or ? and underlined the part of the sentence that was unacceptable. Kuribara finds that although advanced learners were highly accurate in rejecting some kinds of ungrammatical sentences, the *SpecCPx2 structure was judged at chance rates (50%) and reamined at that level even as proficiency increased. The implication is that the subjects have not reset a parameter at all, but instead use general learning mechanisms to improve in their mental representations of the L2 structures. One thing about that experiment that was not clear to me was how '?' answers to the sentences were scored, or whether they were even counted, as Kuribara asserted that 50% was the level of chance, which would be true given there were only two answers. Kuribara includes a learning model at the end of the chapter that seemed to quite ambitiously address the question of how learners come to the conclusions they do.
7. 'Japanese learners' errors on long distance wh-questions' by Shigenori Wakabayashi and Izumi Okawara. This paper is meant to not only provide answers to the research question of what kinds of wh-questions Japanese learners of English will make, but also to illustrate the use of an elicitation technique. The authors note that work on wh-questions has usually consisted of grammaticality judgement data, and this experiment provides production data. The wh-questions were elicited within a framework of a game with different figures. Japanese learners produced wh-questions that children do not produce, and the authors use explanations involving strong complementizers such as [+Q] in English, null complementizers, and weak functional categories in Japanese to explain their results. The authors then were able to make some generalizations, such as that when feature strength is different between the grammras (such as a weak [+wh] vs. strong [+wh]) that the feature in the interlanguage grammar is weaker.
However, these explanations left this reader puzzling as to why simpler explanations for the learners' behavior were not explored. The complex operations that are posited for the learners' behavior may, in fact, be accurate, but I would have liked to see such explanations as chunking or no movement taking place at all also addressed (these possibilities are mentioned in a footnote, but not explored in depth).
8. 'N400 in the brain potential responses of second language learners: What ERPs suggest' by Yuichi Tomita, Kazuhiko Fukuda and Natsuko Tatsuta. This paper begins with the annoying claim that data using Event-Related Potentials (ERPs) can more directly address issues of competence (rather than performance) than do psycholinguistic experiments. Given the fact that ERPs do nothing more than present graphs showing peaks of electronic measurements of the brain in response to stimuli which must be interpreted as to their cause and meaning, this claim surely overreaches itself. Clearly, ERPs may help in the search for the cause of L2 learner behavior, but will not explain the cause of it any more directly than physical measurements of a depressive person^s brain will explain the cause of his or her depression. In this experiment, 14 Japanese learners of English heard 40 stimulus sentences with syntactic anomalies (such as 'Mike listened to Frank's orange about politics'), and had to respond by deciding whether the sentence was correct or not. The authors had divided subjects up into groups based on their proficiency, and found their accuracy on judgements of sentences was significantly different, but found no significant difference for group in the ERP responses (they were looking specifically at a peak around 400 ms, the N400 latency, which occurs when people seek to extract meaning from senseless utterances). They somewhat reluctantly admit in a footnote that their findings may be compatible with the idea that ERP results are influenced by the starting time of exposure to L2, and not by proficiency, but they urge more studies investigating whether differences in proficiency will result in ERP differences.
Chomsky, N. (1995). The Minimalist Program. Cambridge, M.A.: MIT Press. Hawkins, R. (2001). Second Language Syntax: A Generative Introduction. Oxford: Blackwell. Lakshmanan, U. and Selinker, L. (1994). The status of CP and the tensed complementizer that in the developing L2 grammars of English. Second Language Research 10: 28-45. McDaniel, D., McKee, C. and J. Bernstein. (1998). How children^s relatives solve a problem for minimalism. Language 74: 308-334. Radford, A. (1997). Syntactic Theory and the Structure of English: A Minimalist Approach. Cambridge: Cambridge University Press. Vainikka, A. and M. Young-Scholten. (1996). Gradual development of L2 phrase structure. Second Language Research 12:7-39. Wakabayashi, S. (2002). The acquisition of non-null subjects in English: A minimalist account. Second Language Research 18:28-71.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER Jenifer Larson-Hall has a Ph.D. in Linguistics from the University of Pittsburgh. She is teaching English at Kyushu University in Fukuoka, Japan. Her main areas of research are second language acquisition of phonology, the role of phonological theory in predicting learners' errors, the Critical Period in phonology, bilingualism and early language experience, and language attrition.