Abstract
This paper deals with the perceptual salience and intelligibility of reduced segments, both vowels and consonants, in casual Czech. It focuses on a specific type of reduction in which two neighboring sounds (e.g. a VC sequence) are pronounced simultaneously, thereby giving rise to an alternative segment which does not occur in standard Czech phonology, e.g. a nasalized vowel.
These segments constitute a category of parallel articulation as phonetic features of both of the segments are realized at the same time. We hypothesize that parallel-articulated segments, despite having originated as a consequence of purely phonetic processes, do in fact act distinctively in the context of casual speech.
For this purpose, an experiment was designed in which the target segments were used in sentence pairs differing only in the parallel-articulated item, and the capability of listeners to identify the intended meaning of each sentence was tested. It showed that listeners do in fact treat the target segments as distinctive which suggests that these sounds might function as phonologically autonomous segments in the everyday communication of Czech speakers.
In addition, an important role in the segment identification is also to be assigned to non-phonetic factors, such as the frequency and expectability of a particular word.