Dependency Phonology

I. Representatives:

     Luccien Tesniere

     John Anderson


II. Theory:

    The idea of dependency comes from syntactic domain, referring to the head-modifier relation, expressing the asymmetry of constituent. 

     "Dependency phonology" is a theory of the representation of phonological segments and sequences. Its original statement antedated much of the work in other branches of "nonlinear phonology." The theory takes its name from its utilization of the relation of dependency. In dependency phonology, an unstressed syllabic segment is said to be dependent on its stressed sister. The dependency relation has a wider application than the metrical strong-weak relation does. It is employed in the representation of the internal structure of the segment: the phonological features of which a segment is made up can also show dependency relations with each other. The extension of the dependency relation to the segmental domain distinguishes dependency phonology from other nonlinear approaches.

(A) Suprasegmental Structure:

Within the suprasegmental domain, the hierarchical dependency relation is phonetically realized as relative prominence. Stressed syllabics are clearly more prominent than unstressed syllabics. Within the syllable relative prominence is manifested as relative sonority. Vowels are inherently more sonorous than consonants. Consonants are dependent on vowels within the syllable. Similarly, sonorant consonants like liquids and nasals are inherently more sonorous than obstruents such as fricatives and stops. In a cluster, an obstruent consonant is dependent on a sonorant or sonorant governs the obstruent. These considerations lead to a representation of the English word clomp


k  l a  m p                                    

     The dependency relation is represented by vertical placement with the governor of any constituent being placed higher than its dependent. In the example clomp, /p/ is dependent on /m/, which is governor of the coda /mp/. The vowel /a/ is governor of two constructions, the rhyme /amp/, in which it governs /m/, which in turn is the governor of the coda /mp/, and also the syllable /klamp/ as a whole, in which /a/ governs /l/, the governor of the onset /kl/.

(B) Segmental Structure:

The phonological features used in dependency phonology are grouped together in a number of "gestures." The concept of gesture allows phonological rules to refer to well-defined subgroupings of features within the matrix characterizing the segment. Rules involving syllabification and other sonority-related processes are concerned only with properties such as consonantality and sonority. Therefore, they make reference only to features within the categorical or major-class gesture. Those involving phenomena such as vowel harmony and assimilation of place are concerned with properties such as backness and labiality.

The features or components used in dependency phonology are all single valued. The vowel space is treated as triangular, and can be characterized by the interaction of a number of these features. They are including "frontness" (represented as /i/), "roundness" (represented as /u/), and "sonority" or "lowness" (represented as /a/). Low vowels have greater inherent sonority than high vowels. Other features are also needed for a comprehensive account of the representation of the vowel space. As the features are singled valued, vowels contain only those features which characterize the properties that they bear: a high front vowel /i/, will be characterized by the frontness feature alone, and will contain neither /u/(because /i/ is not round) nor /a/(because /i/ is not low). The three features /i/, /u/, /and /a/ characterize the three corners of the vowel triangle. When they occur alone in the representation of a vowel, represent the three most basic vowels. Combinations of the three features represent other vowels. Thus a front mid vowel will contain both /i/ and /a/, a back mid vowel /u/ and /a/, and a front rounded vowel /i/ and /u/. Evidence that these are appropriate representations can be found from processes involving "fusion' and "fission." Fusion processes are those in which a diphthong undergoes monophthongization; fission processes involve the reverse. Typically, the fusion of a diphthong /ai/ or /au/ yields a midvowel /e/ or /o/ and fission the reverse.

Dependency phonology is further distinguished from other nonlinear approaches in the way in which it utilizes single-valued features and the dependency relation for the representation of major classes of segments as well as of different manners of articulation. In dependency phonology, the various classes involved vowels, liquids, nasals, voiced fricatives, voiceless fricatives, voiced stops and voiceless stops are assumed to form a group, by virtue of the fact that they are typically involved in processes governed by relative sonority. The order of segments in a syllable is canonically determined by relative sonority, and the most sonorous segment in a syllable, normally the vowel, is the syllabic. Relative sonority also plays a role in intervocalic "weakening" processes.

Sonority Hierarchy (moving from left to right along the sonority hierarchy):

C        C      V,C        V,C        V        V      V




             V                  V        V,C       C

Voiceless   voiced   voiceless     voiced     liquids    nasals   vowels                       

  stops     stops   fricatives     fricatives          


III. Reference:

 Anderson, -John; Durand, -Jaxques. 1986. Dependency Phonology. In: Durand

  Jacques (ed.). Dependency and Non-Linear Phonology. London: Croom Helm. 1-


Asher, R.E.(ed). 1994. The Encyclopedia of Language and Linguistics. Pergamon

  Press: Great Britain.


IV. Relative websites


new¡½The Leiden Centre for Linguistics-Course syllabus: Phonology

                dependency phonology