Context, coarticulation effects and undershoot in vowel transcription

Consistent with the view that vowel sounds are interpreted by the ear as contextually coherent linguistic targets, the decision was taken to represent familiar-sounding vowels and diphthongs as they were perceived/heard in whole-word citation forms. The ear always evaluates speech sounds in context and automatically compensates for coarticulation effects and articulatory undershoot, hearing the intended target, rather than the ‘underachieved’ peak in the attained formant trajectory.

For example, in the Norfuk vowel cluster (describable as a diphthong followed by a short vowel or as a triphthong) of the word fire, the second element is perceived as a high front vowel [i] or [ɪ]. But if one attends only to the central region of the vowel cluster, isolated from context, this segment has the auditory quality of a low or mid-low front or central vowel [æ] - [ə]. Clearly, this is a case of articulatory undershoot of the off-glide target of the diphthong. Our speech perception mechanism automatically compensates for articulatory undershoot when listening to the vowel in whole-word context. In so doing, tacit phonetic and phonological knowledge of the listener is applied to the perception of the auditory stimulus. A more stable percept is achieved by judging vowel quality in whole word contexts, but at the possible cost of undue contamination of phonetic judgements by phonological expectations from the listener’s native language.