A language-specific markedness constraint: velar elision in Japanese

Kevin Heffernan


Constraint-based phonological theory claims that markedness constraints are universal. In this paper, I argue against this by demonstrating the need for a language-specific markedness constraint to account for the apparent k~i alternation in the Japanese verb paradigm: kaku ‘to write’, but kaita ‘wrote’. Previous accounts have claimed that either the i is epenthetic, or that /i/ is transformed into [k]. However, these possibilities are ruled out because there are always other candidates that better satisfy the markedness and faithfulness constraints. The only other possibility is to posit a VELAR ELISION constraint that deletes velar stops when followed by a high front vowel. I draw on both synchronic and diachronic evidence to argue that such a constraint is active in Japanese. At the same time, the lack of a perceptual and physiological link between velars and the high front vowel suggests that the constraint is specific to Japanese.


Japanese; Velar Elision; Markedness Constraint; Epenthesis

Full Text:


Copyright (c)