Lexical strata as a part of grammar have been assumed to account for various phonotactic (un)markedness in certain groups of lexical items. In particular, the reranking hypothesis of Ito and Mester (1995a,b) and the indexed FAITH hypothesis of Fukazawa (1998) had triggered disputes and discussions among scholars on the fundamental organization of phonological grammar. This article will examine approaches to the sublexicon, and show problems with the aforementioned studies. In particular, the emergence of the unmarked patterns in the marked vocabulary classes presents a serious problem.
The aim of this article is to emphasize the necessity for further investigation into the relationship between Sino-Japanese and Yamato, and to present a couple of topics on the historical phonology of Japanese. One of the main questions that deserves to be challenged is how the indigenous structure extended to SJ word formation, such as geminations, of which the earlier stage can be hardly attested by historical records. Another topic concerns vowel coalescence. The interesting interaction between SJ word types and Yamato accounts for these sound changes.
This paper surveys phonetic and phonological studies dealing with the question of why Japanese speakers 'hear' a geminated consonant in a certain foreign sound string. Phonetic studies reveal that not only the length but also some other factors might contribute to the perception of a geminated consonant. Phonological studies claim that at stem edge position systematic gemination occurs to preserve coda moraicity of the input string while the occurrence of the one inside the word might be related to the Japanese foot structure. We argue that crucial to the issue is how much the system can predict systematic non-occurrence of the geminated consonants as seen in the word 'dokutaa' (doctor).
This paper describes Japanese loanword accentuation, and proposes a strategy to assign accent types of these words. About four hundred sixty unitary loanwords are prepared having more than 5 morae. Nine native speakers, with a Tokyo dialect, are instructed to assign the most separable intra-word juncture and accent location of the words. The experimental result shows that Japanese loanword accentuation is closely related to the word formation quasi-structure and the metrical structure formed by compounded syllables in the words. It is clarified that accentuation rules of unitary words with phonological juncture parallel Japanese compound accent rules. Moreover, an additional new accenting rule is also proposed based on the Japanese metrical structure. The loanword accent under such a metrical structure is placed at the penultimate position of the two-mora syllable unit (foot).
This study aims to clarify when and how compound words are abbreviated to three-mora words. It turns out that a word-final long vowel tends to be maintained when it forms a morpheme on its own. When the final syllable does not constitute a morpheme, the long vowel /aː/ is shortened (e.g. sutoreːto-paːma → sutopa) more frequently than the long vowel /uː/ (e.g. roNgu-buːtsu → roNbuː). In addition, three-mora abbreviations with their final vowels shortened are more likely to be medially accented than other three mora abbreviations in the Kinki dialect. Moreover, a compound word is more frequently abbreviated to a single component when the component consists of one heavy syllable followed by one light syllable than when it contains four moras or one light syllable followed by one heavy syllable (e.g. tisshu-peːpaː → tisshu; sunoː-boːdo → boːdo).