Visual and auditory perception of epenthetic glides

The purpose of this contribution is to improve our knowledge about the time course of visual and auditory perception with regard to the representation of sound types as different in their phenomenological format as vowels and glides. Our results on the perception of Vowel-to-Vowel gesture via the production of epenthetic glides in between – according to our 2-COMP-Vowel Model – allow us to conceive of the timevarying vs. stationary representational issue as linked to the underlying control for moving phases: (i) while the stationary (plateau) or climax phase of a vowel can be shown as truly representative for this segment, (ii) and whereas motion in the time-varying on-gliding phase can be shown to be informative only when shape is incomplete, (iii) motion in the off-gliding phase of the same vowel can reveal itself as misleading, in the sense that it could prime an erroneous candidate for the following vowel, up to the end of the transitional glide in V-to-V. Thus, contrary to the dynamic specification theory for vowels, the moving on-gliding and off-gliding phases can be, respectively, less informative than stationary parts, as shown before, and even truly misleading as shown here. (iiii) Moreover the same off-gliding phase can be recovered as a true controlled glide under certain language-based constraints (as exemplified by the powereffect). Finally in line with our caveat against the claim that "all is dynamics in speech", we will briefly mention recent computational modeling and neural data which support our hypothesis, especially the snapshot neuronal computation which fits with recent brain imaging data of stilled and moving speaking mouths.