This is an electropalatographic investigation of coarticulation for heterosyllabic stop consonant clusters in American English and Catalan VCCV sequences. The heterorganic clusters under analysis were [tk], [kt], [tp], [pt], [kp], [pk]. Evidence for gestural overlap between the two adjacent consonants in the cluster is found quite systematically about the closure midpoint and, less so, at C1 onset and at C2 offset. Overlap occurs between constrained and unconstrained lingual regions (e.g., [p] is produced with more alveolar contact than usual when preceded by [t]) and gives rise to blending between tongue front and tongue dorsum activity during the production of lingual clusters [kt] and [tk]. Clusters are equally sensitive to vowel-dependent effects at all moments in time during the closure period. Such effects are quite large for clusters involving lingual [t] or [k] and non-lingual [p] and quite small for clusters made of the two former lingual stop consonants. These data on consonant- and vowel-dependent coarticulatory effects suggest that stop clusters are produced as highly cohesive production units. A constraint for anticipatory vowel-dependent effects to occur at the cluster midpoint but not so at cluster onset can be taken in support of a time-locked model of coarticulation. Speaker-dependent trends were also observed. © 1993.
- gestural blending
- intergestural cohesiveness