Abstract
The 'motor theory of speech perception' by Liberman et al. (1967) is a popular neurolinguistic theory that attempts to explain how speech sounds (phonemes) are perceived by the human brain. It has been shown that one and the same phoneme can have different acoustic realizations, depending on the sound context it is embedded in.
The phoneme /g/ has different acoustic characteristics in the words 'geese' and 'goose' since it anticipates features of the following (i) and (u) sounds, respectively. This process of assimilating adjacent speech sounds is known as co-articulation, and it leads to changes in the acoustic nature of sounds.
Therefore, when co-articulation occurs, acoustic description cannot solely define a phoneme. The 'motor theory' proposes that the articulatory gesture (vocal tract gesture) that underlies every speech sound is the actual percept of speech since it is of invariant and stable nature.
In this paper, song vocalizations of three gibbons (including the species Nomascus leucogenys, Hylobates muelleri, and H. agilis) are analyzed. The results provide first indications that co-articulation occurs in the calls of each species.
This suggests that the acoustic signal itself is unstable and does not function as the only determinant of sounds. If this is correct, gibbons - in a way similar to humans - may perceive sounds according to their articulatory gestures rather than solely their acoustic features.