Perception of Consonant Clusters and Variable Gap Time

In every case in which measurements of labial-velar stops [kp, gb] have been made, it has been found that the labial and velar gestures are not strictly simultaneous, but rather that the velar gesture slightly precedes the labial one (thus [kp] and not [pk]). One possible explanation for this is that [kp] is more perceptually salient than [pk]. This paper reports an attempt to test this hypothesis by observing listeners' identifications of [apka] and [akpa] with variable gap times inserted between the consonantal onset and release. The results showed that [apka] was more readily identified than [akpa], effectively showing that perceptual salience cannot be invoked to explain the ordering of velar and labial gestures in labial-velar stops