tailieunhanh - Báo cáo khoa học: "Diagnostic Processing of Japanese for Computer-Assisted Second Language Learning"

flection on the errors and inappropriateness, detected by a diagnostic parser, of the sentences composed by the students filling the blanks set up in the given contexts and situations. In this paper we report on prototyping the diagnostic parser implemented using LTAG formalism as a component of the system. LTAG(Lexicalized Tree Adjoining Grammar) is a lexicalized grammatical formalism (XTAG Research Group, 1995). For ease of diagnosing the erroneous sentences composed by the students, lexicalized type of grammars seemed most suitable. Comparing HPSG(Head-driven Phrase Structure Grammar) ( et al., 1994) and LTAG, the well-known two (almost-)lexicalized grammars, LTAG looked more. | Diagn stic Pr cessing f Ja anese f r C m uter-Assiste Sec n Language Learning Jun ichi Kakegawa Hisayuki Kan a Eitar Fuji ka Mak t Itami K hji It h De artment f A lie Electr nic Science Univer ity f T ky 2641 Yamazaki N a- hi Chiba-ken 278-8510 JAPAN kakegawa kan a eitar 76 itami it hg@ a. A str ct As an a licati n f NLP t c m uter-assiste language learn-ing CALL we r se a iag-n stic r cessing f Ja anese eing ale t etect err rs an ina -r riateness f sentences cm se y the stu ents in the given situ-ati n an the c ntext f the exercise texts. Using LTAG Lexicalize Tree A j ining Grammar f rmal-ism we have im lemente a r t -ty e f such a iagn stic arser as a c m nent f a CALL system eing evel e . 1 Intr ucti n In the recent classr m f sec n language learning c mmunicativ appr ach . Wi -ws n 1 77 is r m te in which it matters f r the stu ents t ec me aware f the language use . the functi nality f language usage an it s e en ence n the sit-uati ns an the c ntexts f c mmunicati n. In rer t achieve the jective acc r ing t c nstructivistic int f view f learning t al. 1 1 the stu ents are en-c urage t r uce sentences y themselves in vari us situati ns an c ntexts an gui e t rec gnize y themselves the err ne us r ina r riate functi ns f their misuse ex-ressi ns. We have alrea yr se a mput r-Assist d Languag L arning ALL system t al. 1 7 which r vi es the stu-ents with sam le texts r m ting their re- flecti n n the err rs an ina r riateness etecte y a iagn stic arser f the sentences c m se y the stu ents filling the lanks set u in the given c ntexts an situ-ati ns. In this a er we re rt n r t tying the iagn stic arser im lemente using LTAG f rmalism as a c m nent f the system. LTAG Lexicalize Tree A j ining Grammar is a lexicalize grammatical f rmalism XTAG Research Gr u 1 . F r ease f iagn sing the err ne us sentences c m-se y the stu ents lexicalize ty e f grammars seeme m st suita le. C m aring HP G Hea - riven Phrase tructure Grammar