Ellipsis and Reference Tracking in Japanese (review)

This book is a very thorough study of the phenomenon of ellipsis in Japanese. Ellipsis is pervasive in the language, with up to 80% of all subjects being elided, depending on the genre. While this fact has attracted a lot of interest and sparked a lot of research, however, most studies so far have been rather limited in their scope, confining themselves to particular levels of analysis and particular factors that lead to ellipsis. Shigeko Nariyama sets out to take research in this field to a new level by ‘integrating all of the subfields that are essential in fully describing the complex mechanisms of ellipsis’ (356). Her goal is to ‘present a detailed description of linguistic mechanisms for ellipsis resolution, and … to formulate an algorithm based on these mechanisms that shows how referents of ellipted arguments are identified’ (6).

The first two of the seven chapters in this book provide an introduction to the area of study and to previous research. As throughout the book, the presentation of previous research is comprehensive and thoughtful, taking a balanced view of different approaches. Reference tracking is treated here as well, since ellipsis in text is always a problem of reference tracking. The core of the book, elaborating on the linguistic devices at work in ellipsis and reference tracking in Japanese, runs from Ch. 3 through Ch. 6. Ch. 3 deals with predicate devices, verbal semantics, inverse verbs, honorification, epistemic morphemes, and the like, which all function to identify referents. Ch. 4 introduces a ‘principle of direct alignment’, claiming that, similar to Algonquian languages for example, sentences in Japanese obey a person/animacy hierarchy in which the subject always has to be higher on the hierarchy than nonsubjects. Ch. 5 presents an additional ‘principle of argument ellipsis’ which states that ‘[t]he higher an argument is in terms of the Person/animacy hierarchy and Discourse salience, the more prone it is to ellipsis, unless focus is placed on it’ (242). Ch. 6 elaborates on discourse devices, focusing on the problem of subject/continuity in discourse and its relationship to ellipsis. Ellipsis here is conceptualized as ‘the unmarked representation of [referential] sameness’ (280). In Ch. 7, N comes up with an algorithm for ellipsis and referent identification by integrating the predicate devices, sentence devices, and discourse devices discussed in the previous chapters. This algorithm is tested on a short stretch of written discourse.

As stated before, N’s discussion of previous research is fully adequate, and she adds a new dimension to the study of the topic by carefully integrating what is already known from more particularized studies about the role of predicate devices and discourse devices in reference tracking. This alone makes N’s book worthy of publication. Her most original contribution to the topic, however, the direct alignment hypothesis, although not implausible, would gain significantly if it were supported by better evidence. Atypical for her book, N here relies on her grammatical judgment of made-up sentences with little support from textual data, and some of her judgments might fail to convince the critical reader. Thus at face value it is hard to see what is wrong with sentences like Taroo-ga watashi-o mitsuke-ta ‘Taroo found me’ (marked with two question marks on p. 179 but with only one on p. 226) or English Taroo heard it from me (180). Another drawback is the inconsequential and sometimes erroneous morpheme transcription of the Japanese sentences. Thus we find osowa-rare-ta ‘was taught’ instead of osowar-are-ta (189), iki ma-sen ‘go not’ instead of iki-mas-en (146), omot-teiru (omot-te i-ru ‘be thinking’, 143) but kangaeteiru (kangae-te i-ru ‘be thinking’, 159), and choosen shi (‘challenge’, 274) but kitaishi (‘expect’, 261), to give just a few examples. Fortunately, these technical flaws do not affect the core issues. On the whole, N delivers what she promises, and the merits of this book clearly outweigh its shortcomings. N...