[Author Login]
Title:Production, perception, and emergent phonotactic patterns: A case of contrastive palatalization
Authors:Alexei Kochetov
Abstract:The goal of this thesis is to answer two related questions: (i) to what
extent can limitations of human speech production and perception explain
cross-linguistic positional markedness asymmetries? and (ii) is it
logically necessary to attribute these scales to Universal Grammar, as
is commonly assumed (e.g., Prince & Smolensky 1993)?

In order to answer these questions, a case study involving the
distribution of the plain-palatalized contrast in labial and coronal
stops was carried out. A typological survey of languages with
contrastive palatalization shows that the distinction between plain and
palatalized segments is most often maintained in the syllable onset
position and most commonly neutralized in the preconsonantal coda
environment. Palatalized labials are more susceptible to neutralization
than the palatalized coronals. The most common outcome of neutralization
is a plain segment.

In a number of articulatory, acoustic, and perceptual experiments I
investigate Russian plain and palatalized stops in cross word-boundary
sequences. The articulatory study reveals that the environments that
generally induce neutralization exhibit the most variability in the
magnitude and timing of the tongue body gesture, the articulatory
correlate of the plain-palatalized contrast. They also show that that
the effect of environment is different for palatalized labial and
coronals. In addition, the variable overlap of primary gestures has
important acoustic consequences: it results in the lack of acoustic
release burst of the first consonant and less distinct vocalic

The perceptual findings from both native and non-native subjects under
several conditions demonstrate that listeners reliably distinguish the
contrast in the contexts when the respective gestures are stable (as in
syllable onset), and fail to hear it in the environments that induce
gestural variability (syllable coda, especially before consonants).

The results of perception, taken as derived scales, serve as input to a
hypothetical learner constructing language-particular grammars.
Crucially, the learner is not equipped with positional markedness
scales. Limitations on what can or cannot be recovered severely restrict
the learning path, ultimately resulting in a limited set of possible
grammars that correspond to the attested cross-linguistic patterns of
contrastive palatalization.
Article:This article has been withdrawn.