Position and place effects in Russian word-initial and word-medial stop clusters

Abstract

Studies of inter-gestural timing have shown that (i) word-initial obstruent clusters tend to exhibit less gestural overlap than word-medial or word-boundary clusters, and (ii) the degree of overlap is further affected by the place of articulation of the obstruents (Byrd, 1996; Chitoran, Goldstein, and Byrd, 2002). Both findings have been attributed to perceptual recoverability considerations. This paper presents results of a magnetic articulometer (EMMA) study of Russian word-initial and word-medial stop clusters (e.g., [pt]ashka little bird versus la[pt]a bat). Data collected from 3 native speakers of Russian show that clusters with coronals and dorsals as C1 ([tk], [kt], [kp], [tjm]/[djb]) exhibit less overlap word-initially than word-medially, while the cluster with the labial as C1 ([pt]) does not exhibit the same timing pattern. The findings are interpreted as providing additional support for the role of perceptual recoverability in intergestural timing. First, less overlap in word-initial clusters, compared to word-medial clusters, ensures better place recoverability of C1 (cf. Chitoran et al., 2002). Second, unreleased labials are more perceptually robust than unreleased coronals (Byrd, 1992; Surprenant and Goldstein 1998) and dorsals (Wright, 2001; Kochetov and So, 2005), and thus do not require the same degree of overlap. [Work supported by SSHRC.]