Voice coarticulation as a function of place and manner of articulation in Russian

The paper reports some new data based on experimental study in voice coarticulation of voiced and voiceless obstruents adjacent to sonorants as a function of place and manner of articulation of these consonants in Standard Modern Russian. The results of the first experiment based on the data collected from fifty-seven participants show that the phonetic realization of postpositive particles b and ž in the position before the next word’s initial sonorant depends on the degree of similarity of the consonantal segments in question: the more phonetic features they have in common, the lower is the probability of the obstruent’s word-final devoicing in the external sandhi position. The results of the second experiment have shown that in word-internal [sonorant + obstruent + sonorant] consonantal clusters the voice coarticulation of the obsruent is observed; it may be determined by the preceding homorganic nasal consonant. This voice accommodation does not lead nevertheless to the voiced/voiceless phonemes’ neutralization, since the perceptual analysis demonstrates that the contrast in question is still maintained — by means of phonetic parameters other than voice (phonation itself), such as closure duration, burst duration (being significantly higher in underlyingly voiceless stops) and relative overall intensity (being noticeably higher in underlyingly voiced obstruents) etc. The paper finally argues that the phenomenon discussed should be phonologically treated as a sort of so-called “incomplete neutralization”.