Johns Hopkins University – Homewood Campus – (410-516-5250/office phone)
Thursday, March 8, 2007
3:45 p.m.
Florida
Atlantic University
Department of Psychology
What we know about what we have never heard: Evidence from perceptual illusions
Are speakers equipped with grammatical preferences regarding linguistic structures that they have never encountered? Optimality Theory (Prince & Smolensky, 1993/2004) postulates the existence of a universal set of grammatical markedness constraints--constraints that express (dis)preferences for certain linguistic structures (e.g., structure A>B). By hypothesis, these constraints form part of the grammars of all speakers, irrespective of whether the structures under consideration (e.g., A, B) are present in their language. The following research tests this hypothesis. As a case study, we examine sonority-related restrictions on onset structure. Typological evidence suggests that onsets of rising sonority are preferred to onsets with sonority plateaus, which, in turn, are preferred to onsets of falling sonority (e.g., bnif>bdif>lbif). We demonstrate that such preferences modulate the perception of unattested onsets by English speakers: Universally-dispreferred onsets are more likely to be misperceived as disyllabic (e.g., lbif'lebif) compared to onsets that are relatively preferred (e.g., bdif). The perceptual illusions of English speakers are inexplicable by several aspects of linguistic experience, including the statistical properties of the English lexicon and experience with reduced forms in fast speech (e.g., benath'b'neath). Likewise, such misperceptions are not due to stimulus artifacts (the same stimuli are perceived accurately by speakers of Russian--a language that tolerates such clusters), or a failure to encode the phonetic form of the stimulus (similar misperceptions are found even with printed materials). These observations suggest that the misperception of marked onsets reflects grammatical preferences. The generality of such preferences across segment classes (e.g., for nasal onsets, mlif>mnif>mdif) further indicates that the relevant grammatical knowledge specifically concerns sonority profile, not merely segment co-occurrence. We conclude that the grammars of all speakers might include universal markedness constraints on sonority profile.
Visit our website: http://www.cogsci.jhu.edu (menu header: Events)