Skip to Main content Skip to Navigation

Suites de Consonnes en Berbère : Phonétique et Phonologie

Abstract : My thesis is in phonetics and phonology, with particular interest in experimental approaches to phonology. One of my main interests was in the phonetics and phonology of complex consonant clusters. I mainly work on a language, Tashlhiyt Berber spoken in the Southern part of Morocco, which presents some typologically rare features of the utmost phonetic theoretical interest. One of these features is a distinction between simple (short) and geminate (long) consonants in different prosodic positions (including absolute initial and final positions). This phenomenon is quite unusual and raises the question of how such segments are phonetically implemented and how they are phonologically represented and syllabified. Another phenomenon is a peculiarity of Tashlhiyt Berber. In this language, words may consist at the underlying level of consonants only (e.g. /tssrglttnt/ "you closed them"). These words may even consist of voiceless obstruents only (e.g. /tfktstt/ "you gave it"). These underlying voiceless words raise two main questions: (1) Do these words contain some kind of vowels which may occupy the nucleus of the syllable? (2) What kind of laryngeal coordination patterns are produced during the realisation of these forms?

Tashlhiyt Berber has been proposed as a language in which any consonant, even a voiceless stop, can act as a syllable peak (Dell and Elmedlaoui 1985, 1988, 2002, Prince and Smolensky 1993, Clements 1997). The most striking and controversial examples, taken as arguments in favour of this analysis, involve large series of words composed only of non-sonorant segments. This claim is challenged by different authors who argue that the alleged consonant-only sequences are actually pronounced with schwa vowels in the context of the syllabic consonants. The goal of this part of my thesis is to determine whether, in addition to /a/, /i/, and /u/, there is a fourth vocalic segment at the level of phonetic and phonological representations that can act as a syllable peak. Different types of arguments, including acoustic, fiberscopic, and photoelectroglottographic data, are presented. They converge in support of the claim that voiceless, vowel-less words exist in Tashlhiyt Berber. This provides a compelling argument to the view that in this language a consonant sequence of the type /tk/ is a well-formed syllable.

Concerning the laryngeal behavior during the production of voiceless words, results of this thesis clearly show that the glottis, even in completely voiceless utterances, does not simply remain open, but rather that the glottal aperture is continuously modulated. This laryngeal modulation is quite systematically related to the phonetic nature of the individual segments present in the sequence: segments produced with a high rate of oral airflow are produced with a separate laryngeal-opening gesture. These results, which is in general agreement with those obtained from some Germanic languages, gives a compelling demonstration of how intimately laryngeal and oral articulations are linked.

Within CV phonology, geminates, including post-lexical ones, are represented as single melodic units associated to two prosodic positions. This thesis examines the way these autosegmental representations are reflected in the phonetic details of speech production and questions the phonological relevance of these correlates. In particular, it investigates two questions: what are the acoustic and articulatory differences between singletons and lexical geminates; and are there any acoustic differences between the different types of geminates? Tashlhiyt Berber serves as an excellent test case for these issues. It has contrastive singleton and lexical geminate consonants in all positions. In addition, it presents two types of phonologically derived geminates: concatenated and assimilated ones. Results of this thesis show that the primary correlate that distinguishes singletons from lexical geminates is duration, even for voiceless stops after pause. This primary correlate is enhanced by additional correlates, which may be crucial to the perception of absolute initial and final geminates. The three types of geminates all show the same temporal characteristics, which supports their receiving the same timing representation, but there are additional phonetic characteristics on which they differ. While assimilated geminates, like underlying ones, are enhanced by additional acoustic attributes, concatenated geminates are not. Implications of these results for the general issue of geminate behaviour are discussed, with particular attention to geminate Ambiguity and geminate Inalterability.
Document type :
Complete list of metadatas
Contributor : Gwénaëlle Lo Bue <>
Submitted on : Thursday, April 26, 2007 - 12:39:39 PM
Last modification on : Thursday, January 11, 2018 - 6:17:29 AM
Long-term archiving on: : Tuesday, April 6, 2010 - 10:58:13 PM


  • HAL Id : tel-00143619, version 1



Rachid Ridouane. Suites de Consonnes en Berbère : Phonétique et Phonologie. Linguistique. Université de la Sorbonne nouvelle - Paris III, 2003. Français. ⟨tel-00143619⟩



Record views


Files downloads