Rhetorical stress in Spanish

Stress systems with rhythmically alternating secondary stresses have played a major role in the development of metrical stress theory (Halle and Vergnaud 1987). Hayes (1995: 33), in fact, focuses primarily on such rhythmic systems for his theoretical proposal, stating that “these are the systems that arguably are of the greatest interest from a metrical perspective”. It is somewhat perplexing, then, that, in the case of several languages that have been taken to possess such rhythmic secondary stress systems, phonetic investigation has failed to find evidence for them. Thus, Dogil and Williams (1999) did not find phonetic evidence for rhythmic secondary stress in Polish or Spanish, and Arvaniti (1992, 1994) also failed to find evidence for rhythmic stress in Greek. Further experimental work on Spanish secondary stress has also reported lack of acoustic evidence (Prieto and van Santen 1996; Díaz-Campos 2000). Polish, Spanish, and Greek are languages with contrastive primary stress, whose position in individual words is not in question and where, in addition, some phonologists have postulated the presence of rhythmic, alternating secondary stresses on the basis of their intuition. The fact that no empirical evidence has been found so far to support these intuitions raises serious concerns regarding the nature of the evidence that has been used in typological and theoretical work. This can be seen as particularly disturbing because, among those languages that have been claimed to possess rhythmically alternating stress, Spanish, Polish, and Greek have large numbers of speakers and are easily accessible, but, as de Lacy (this volume) points out, for a majority of languages in stress databases we have a single source. If, for languages like Spanish, Polish, and Greek, experimental work has failed to find evidence for rhythmic stress, how much faith can we have in the correctness of the description of the stress systems of less accessible and more poorly documented languages? (de Lacy raises this question also for primary stress in languages where presumably it is not phonologically contrastive.)