Publishing Partner: Cambridge University Press CUP Extra Publisher Login
amazon logo
More Info

New from Oxford University Press!


Cognitive Literary Science

Edited by Michael Burke and Emily T. Troscianko

Cognitive Literary Science "Brings together researchers in cognitive-scientific fields and with literary backgrounds for a comprehensive look at cognition and literature."

New from Cambridge University Press!


Intonation and Prosodic Structure

By Caroline Féry

Intonation and Prosodic Structure "provides a state-of-the-art survey of intonation and prosodic structure."

E-mail this page 1

Dissertation Information

Title: Deriving Natural Classes: The Phonology and Typology of Post-Velar Consonants Add Dissertation
Author: John Sylak-Glassman Update Dissertation
Email: click here to access email
Institution: University of California, Berkeley, Department of Linguistics
Completed in: 2014
Linguistic Subfield(s): Phonetics; Phonology; Typology;
Director(s): Andrew Garrett
Sharon Inkelas
Darya Kavitskaya
Keith Johnson

Abstract: In this dissertation, I propose a new method of deriving natural classes that is motivated by
the phonological patterning of post-velar consonants (uvulars, pharyngeals, epiglottals, and
glottals). These data come from a survey of the phonemic inventories, phonological
processes, and distributional constraints in 291 languages.
The post-velar consonants have been claimed to constitute an innate natural class, the
gutturals (McCarthy 1994). However, no single phonetic property has been shown to
characterize every post-velar consonant. Using data from P-base (Mielke 2008), I show that
the phonological patterning of the post-velar consonants is conditioned by the presence of a
pharyngeal consonant, and argue more generally that natural classes can be derived from
phonetic connections that link specific subsets of phonemes. Phonological entailments
(Burzio 2002a,b; Wayment 2009) are used to model these connections. Entailments are
derived from the co-occurrence of features within a single phoneme, and state that if one
element of representation (p) is present, then so is another (q). Entailments are central to
deriving natural classes, and function as a source of explanation for why phonemes are able
to pattern together.
Because natural classes are proposed to be derived rather than representationally specified,
I propose that formal representations are responsible for capturing phonemic contrast and
phonetic detail that is essential for accurately describing phonological processes and
distributional constraints. Following this proposal, I present a new formal representation of
the post-velar consonants. Traditional phonological features are associated with the
phoneme itself and are motivated by the phonemic contrasts discovered through the
typological survey. In addition, phonetic subfeatures are associated with language-specific
allophones of phonemes and are motivated by phonological processes that are influenced
by non-contrastive phonetic properties.
Natural classes are derived as feature classes (symbolized ʗ; Padgett 1995, 2002), which
are sets of feature bundles combined in a union relationship. Their composition is derived
using a new type of Optimality Theory constraint, ASSOCIATE(ʗi, p ↔ q), which uses
entailments to require the feature bundles represented by p and q to be present in a feature
class, ʗi. I argue that the entailments in these constraints establish a surface
correspondence relationship between the feature bundles that they require to be present,
and that regulating this correspondence relationship is key to determining the composition of
feature classes within a language. MAX-FFʗ requires feature bundles to be in
correspondence, which eliminates feature bundles that are not required by ASSOCIATE
constraints. DEP-FFʗ militates against feature bundles being in correspondence, and its
ranking with respect to the ASSOCIATE constraints determines which feature bundles are
included in a feature class. The feature bundles required by ASSOCIATE constraints ranked
above DEP-FFʗ are included in a feature class, while those that are required by ASSOCIATE
constraints ranked below DEP-FFʗ are not. The dissertation concludes by demonstrating
that this proposed system can derive natural classes composed of post-velar consonants,
including the guttural natural class in Arabic.