Papers from the Second International Workshop on Japanese Syntax Edited by William J. Poser (review)

REVIEWS211 Papers from the Second International Workshop on Japanese Syntax. Edited by William J. Poser. Stanford: Center for the Study of Language and Information , 1988. Pp. vii, 241. Reviewed by John B. Whitman, Cornell University The eight papers in this volume were originally presented at the second of three workshops on Japanese syntax funded by the System Development Foundation , held at Stanford University on 7-9 March 1986. Proceedings of the first workshop appeared as Kuroda 1986, and proceedings of the third as Kuno & Shibatani 1990. Perhaps serendipity, perhaps the venue account for the high proportion of important papers brought together at the Stanford workshop and collected in this volume. Taken together, the papers provide a good overview of current issues in theoretical syntactic research based on Japanese data: the VP-internal subject hypothesis (described below), the syntax/morphology interface , local and long distance dependencies as exemplified by quantifier float and scrambling, and others. In this regard the editor has accomplished a worthy task. In other respects, editing of this collection of camera-ready manuscripts is minimal, but sufficient to render the title ofthe paper by Shigeru Miyagawa, 'Predication and Numeral Quantifier', as 'Prediction and numeral quantifier'. Two contributions, by Sige-Yuki Kuroda and Masayoshi Shibatan i & Taro Kageyama, are of particular significance even to readers outside the Japanese field. Kuroda's paper, 'Whether we agree or not: A comparative syntax of English and Japanese', is cited by such researchers as Koopman & Sportiche (1988) as one of the first thoroughly articulated versions of the VP-internal subject hypothesis. Kuroda proposes that subjects originate universally in a position internal to the verb phrase; in languages like English, with a morphological requirement that the subject agree with the tensed verb, the subject must move from this position to one where it may trigger agreement. This position is external to the VP (as evidenced in VP Ellipsis contexts), and in Kuroda's and most subsequent versions of the hypothesis it is identified with the specifier of the phrasal projection headed by the tensed verb, I[nflection] P[hrase]. In languages without a morphological agreement requirement, such as Japanese, the subject may remain internal to the VP. Kuroda's version of the VP-internal subject hypothesis for Japanese differs from the roughly contemporary versions of this hypothesis proposed by Kitagawa 1986 and Fukui 1986. Kuroda gives sentences in Japanese and English the same phrase-structural analyses at all levels of representation: both are IPs with specifier positions and VP complements. They differ only in that the IP specifier position is obligatorily filled in English but optionally filled in Japanese . There is thus a potential transition between Kuroda's proposal and subsequent work such as Masunaga 1987, Diesing 1988, Tateishi 1989, and Ueyama 1989. These scholars identify both VP-internal and VP-external positions for the surface subject in Japanese: the nominative-marked VP-internal subject is identified with the 'thetic' subject of Kuroda 1965, 1972 or the 'neutral description ' pattern of Kuno 1973, while nominative-marked VP-external subjects are identified with Kuroda's 'categorial' or Kuno's 'exhaustive-listing' subject. 212LANGUAGE, VOLUME 68, NUMBER 1 (1992) Shibatani & Kageyama ('Readjustment and compound formation') distinguish a class of what they call 'post-syntactic compounds' in Japanese, such as in la, from lexical compounds, as in lb: (1) a. kazoku to [[Yooroppa: ryokoo] noorí] family with Europe travel genitive occasion 'the occasion of traveling in Europe with the family' (S&K: 223) b. kazoku to no[Yooroppa: ryokoo] (no family with genitive Europe travel genitive ori) occasion '(the occasion of) European travel with the family' (222; my gloss) Shibatani & Kageyama enumerate a number of crucial properties that support the hypothesis that compounds like the one in la have a clausal source. These include the fact that postsyntactic compounds license arguments and adjuncts of the head (ryokoo 'travel' in 1) in the case- or postposition-marked form they appear in with verbal predicates; and the fact that the accentual characteristics of the derived noun [Yooroppa: ryokoo] (where ':' indicates the juncture characteristic of postsyntactic compounds) are distinct from those of lexical compounds : the latter may contain at most one accentual peak, while the postsyntactic compound in la retains the pitch pattern...