PublisherUniversity of Arizona Linguistics Circle
AbstractIn this paper, I argue that the rhythm rule phenomenon in English is best treated in terms of a theory incorporating the notion "stress clash" (Hammond, 1988), rather than the notion "eurhythmy" (Hayes, 1984). There are three central arguments. First, it is argued that the eurhythmy theory is intrinsically undesirable as it requires a theory of universal grammar that countenances arbitrary counting. Second, it is shown that the eurhythmy theory makes incorrect predictions about the behavior of words with initial stressless syllables. Third, it is shown that the clash -based theory, as opposed to the eurhythmy theory, generalizes nicely to account for the Montana cowboy phenomenon. The organization of this paper is as follows. First, I review the traditional clash -based account of Liberman and Prince (1977). I go on to review the eurhythmy account of Hayes (1984). This includes three central claims /effects: the quadrisyllabic rule, the disyllabic rule, and the phrasal rule. It is next shown that each of these effects can be achieved with independently required principles and machinery and that there is no need for a specific theory of eurhythmy. The following notation will be used in this paper. An acute accent will denote the strongest stress in a domain; a circumflex marks an intermediate stress; a grave indicates less stress; and an unmarked vowel indicates even less or no stress.