Palatalization in Japanese mimetics: Response to Mester and Itô

DISCUSSION NOTE Palatalization in Japanese mimetics: Response to Mester and Ito* Lawrence SchourupIkuhiro Tamori Osaka Women's UniversityKobe University of Commerce In an article in this journal Mester & Itô (1989) argue that patterning in the distribution of a palatalization morpheme occurring in Japanese mimetic forms constitutes strong evidence for the phonological theory of Restricted Underspecification . Their position is tenable, however, only if they can successfully show that mimetic palatalization is in fact a morpheme. We do not feel that they have shown this convincingly. M&I posit that mimetic roots acquire nonautomatic palatalization through lexical association ofan autosegmental morpheme tt with unpalatalized mimetic roots. As evidence for the existence of p they cite forms in which palatalization is claimed to coincide with the meaning element 'uncontrolledness'. The forms in question fall into two groups: (1) palatalized (hereafter P) forms which have a nonpalatalized (hereafter N) counterpart (e.g. zyabu-zyabu and zabu-zabu 'splashing') and (2) forms (e.g. uzya-uzya 'swarming') with no surface-occurring N counterpart (*uza-uza), but in which palatalization is nevertheless said to correspond to the meaning 'uncontrolledness'. M&I do not find the absence of a corresponding N form in cases of type (2) to be problematic; nor are they concerned about N forms which lack a P counterpart (e.g. paku-paku 'munching ', *pyaku-pyaku), or about cases where the semantic relationship between P and N forms is 'rather opaque' (e.g. horo-horo 'weeping elegantlyV/i^orohyoro 'looking thin and weak'; but see our Appendix for alternative glosses). Rather, they interpret such asymmetries as further justification for their claim (269): 'Idiosyncratic behavior as in [such examples], the hallmark offorms derived in an early lexical level (see e.g. Kiparsky 1982), confirms our hypothesis that palatalized mimetics are the results of a genuine lexical process and not of a 'surfacy' (say, postlexical) paralinguistic phonetic modification process.' M&I do not state how many forms are involved in the regularities and idiosyncratic behavior they refer to. A generous reading of the above passage will take it for granted that asymmetric forms are relatively few in number and that forms which clearly demonstrate the regularity are at least plentiful; this is further suggested by the statement (269) that 'Not all mimetics have palatalized counterparts, and some have no nonpalatalized source' (emphasis ours). To be weighted fairly, however, such statements must be backed up by some indication of the actual numbers of forms involved. * We wish to thank Hisao Kakehi, Masayoshi Shibatani, Timothy J. Vance, Arnold M. Zwicky, and two anonymous reviewers for useful comments on an earlier draft. Responsibility for errors is entirely our own. 139 140LANGUAGE, VOLUME 68, NUMBER 1 (1992) The following questions all seem to us to be relevant, in varying degrees, to an evaluation of the claim that nonautomatic palatalization in Japanese mimetics is produced by mapping a palatalization morpheme onto N roots: (i) How many P/N pairs are there? (ii) In how many of these pairs is there a meaning relationship, between the P and N members, attributable to the presence vs. absence of a morpheme meaning 'uncontrolledness'? (iii) How many N forms lack P counterparts? (iv) How many P forms lack N counterparts? While there are no generally accepted criteria by which to establish the existence of a lexical process, it seems clear that if the number of forms is high for both questions (i) and (ii) and low for questions (iii) and (iv), this will be of some comfort to proponents of a morphemic view of Japanese mimetic palatalization . The answers to these questions are, however, all problematic for the analysis proposed by M&I. To answer questions (i)-(iv), it is necessary to consider as many as possible of the mimetic forms in Japanese. Three fairly extensive dictionaries of Japanese mimetic words have been published: Amanuma 1973, Asano 1978, and Mito, Kakehi, et al. 1981. In addition, a thesaurus of Japanese mimetics has recently appeared (Chang 1990). ' Of these works Asano 1978 is the most comprehensive , but a number of common forms and usages are omitted and definitions are often given only in broad outline. In 1980, construction of a database to serve as a resource for a more comprehensive dictionary of Japanese mimetics (Kakehi et al. 1992...