Empty Category and the Effect of Teaching in Sentence Processing

Empty Category and the Effect of Teaching in Sentence Processing Takako KAWASAKI Hosei University 2-17-1 Fujimi, Chiyoda-ku Tokyo, 102-8160 JAPAN takako.kAdream.com Kiyoshi ISHIKAWA Hosei University 2-17-1 Fujimi, Chiyoda-ku Tokyo 102-8160 JAPAN lingAiij4u.or.jp Abstract Different syntactic frameworks have different ways to deal with dislocation construc- tions. One major disagreement is whether or not empty categories should be assumed. Researchers have been working on this issue on the ground of "psychological reality" (of empty categories), yet have not come to an agreement. The first aim of this paper is to propose an experimental scheme to settle the issue of empty categories. Our second aim is to propose that the application of the experiment to L2 is beneficial to see the teaching effect in L2 acquisition. If native speakers' result supports no-empty-category analysis, yet, L2 learners exhibit the different result, L2 teaching of "wh-movement" created a category which is not in native speakers' mental grammar. 1 Introduction This paper proposes an experimental scheme which solves two independent problems in sentence processing. The first problem concerns the role of empty categories in sentence processing. The second is regarding the sentence processing in L2 acquisition. 2 Backgrounds One major disagreement among syntacticians is whether to posit an empty category or not in analyzing dislocation constructions. For example, consider (1): (1) Which man do you think Mary loves? Under the phrase-structure-based theories, whether transformational (GB/MP; Chomsky 1981, 1993) or non-transformational (GPSG; Gazdar et al. 1985), the dependency between which man (the filler) and loves (the head that selects it) is usually mediated through an empty category (trace) located in the filler's "canonical" position. (2) Which man i do you think Mary loves ei? The filler is the antecedent for the trace, which is an argument of the selecting head. On the other hand, non-phrase-structure-based theories such as HPSG (Pollard and Sag 1994, Sag 1997, Bouma et al. 2001), LFG (Kaplan and Zaenen 1989), and various versions of categorial grammars (Steedman 1996, 2000; Pickering and Barry 1991, Pickering 1993) link the filler wh- phrase to the selecting head without the mediation through empty categories. For example, the gist of the GPSG representation (in a more or less HPSG style) of (2), which assumes a trace, would be something like (3a), while the gist of the HPSG representation without a trace would be something like (3b): 456 VP[ SLASH 1 V� SE SLASH 1 V you which man do S[ SLASH 1 1 denotes an NP, V� Nr VP[ SLASH 1. do you d V� S[ SLASH 11 ] NP� VP[ SLASH COMPS SLASH � loves think Mary (3) a. 1 NP S Sf SLASH b.� S think� NP�VP[ SLASH 1 Mary V[ COMPS ( 1 )] 1 [ SLASH 1 loves NP which man Given that the tag XI SLASH 1 denotes an X with an NP missing. In (3a), loves is a transitive verb that takes an NP as its complement (indicated by the non-empty COMPS value), which is, in this case, an empty category e. …

Content truncated to 500 words