Difference between revisions of "Theoretical Treatments of Epenthesis"

Revision as of 18:07, 25 September 2009

Epenthesis in Scottish Gaelic primarily refers to vocalic epenthesis (also known as Svarabhakti), which is generally understood as the insertion of a vowel between a consonantal sequence. Vocalic epenthesis is closely related to syllabification and is subject to stringent phonological conditions on the part of the surrounding consonants. The locality, prosody and identity of Svarabhakti vowels are nevertheless predictable. Researchers however debate the phonological mechanisms that render those characteristics. Two such accounts are exposed below.

Theoretical Treatments

Autosegmental Phonology

Adopting an autosegmental framework for his analysis, Clements (1986) offers a formal treatment of syllabification and vowel epenthesis in the Barra dialect of Gaelic. His argument pertains to the non-contrastiveness of syllabification at the level of underlying representation. Two main types of surface syllabification are delineated:

- Type A relates to the syllabification of vocalic sequences -- VV [duən] poem vs. V.V [du.an] hook

Clements reports an observation first made by Borgstrøm (1937) that "the break in the tension [occurring at the location of the dot – GNC] is a phenomenon which occurs in positions where a consonant could be expected, a kind of ‘consonant without oral articulation’, comparable to h" (150). This remark leads Clements to assume that the underlying contrast of syllabification is really one of /VV/ vs. /VCV/ where C represents a silent consonant devoid of any segmental features except for syllabicity (i.e. [-syllabic]). Clements conjectures that such resulting hiatus might be the reflex of a historical loss of the intervocalic consonant. Illustrative examples are provided below (adopted from Clements 1986: 333):

File:VV/V.V.jpg

- Type B relates to the syllabification of an intervocalic consonant -- VC.V [ar.an] bread vs. V.CV [a.ram] army

Clements identifies three phonological environments influencing the syllabification of this intervocalic consonant:

(1) V:.CV [mo:.ran] much -- C assumes the onset position of the second syllable when directly preceding a long vowel.

(2) V̆C.V [ar.an] bread -- C assumes the coda position of the first syllable when directly preceding a short vowel.

(3) ˈV̆.CˈV̆ [ma.rav] dead -- C assumes the onset position of the second syllable when directly preceding and following short, stressed vowels.

Borgstrøm (1940) reports that “[n]ative speakers are described as distinguishing between the two types of words [i.e. (2) and (3)] in deliberate speech, syllabifying faNak crow as [faN.ak], while declaring that in ʃaLak [ʃa.Lak] hunting the L and the k are inseparable” (153). Clements proposes that this difference of syllabification is the result of vocalic epenthesis, which "occurs before the sequences [hp, çk', xk] only if preceded by a nonlenited plain consonant , i.e. /N, L, R/." (1986: 327).

Clements also predicts the nature of the epenthetic vowel from the surrounding phonological environment. Borgstrøm (1937) mentioned that out of the 78 occurrences of epenthetic vowels, 52 of them where identical to the preceding vowel, while 26 were not. Relying on an autosegmental analysis where nonlabial consonants are specified for the feature [back] and labial consonants are not, Clements shows that the featural composition of the epenthetic vowel is entirely supplied by the preceding CV sequence via a rightward spreading of features. Nonlabial consonants, specified for the [back] feature, are opaque to the featural spread, while labial consonants are transparent to the process. The following schematization illustrates the proposed mechanism adopted from Clements 1986: 331).

File:Feature spreading (epenthetic vowel).jpg

Articulatory Phonology

Bosch (1995) adopts a gestural approach to the treatment of syllabification and vowel epenthesis in the Barra dialect. Her argument crucially rests on a gradient (rather than discreet) approach to syllable constituency, where the syllable formed by the epenthetic vowel is an extension of the original syllable.

Unlike previous accounts, Bosch's analysis seeks to arrive at a justification for the particular prosodic characteristics of the epenthetic vowel, namely the fact that both the original and epenthetic vowels have equal stress and pitch level. Espousing the theoretical tools made available by articulatory phonology, she claims that "epenthesis can be viewed as the result of a premature articulation of the sonorant gesture, which is itself superimposed upon the preceding vocalic gesture" (10). The following gestural representation shows epenthesis as a readjustment in timing (adopted from Bosch 1995:11).

[[File:Gesture [aram].jpg]]

As a result, epenthesis is no longer thought of as the consequence of an additional tongue body (vocalic) gesture, nor as an additional vocalic slot on the timing tier, nor as an additional word-internal moraic position. The outcome is that there is only one long syllable carried out by the original tongue body gesture that retains the original stress and pitch level. Such perspective has the benefit of accounting for the fact that putative disyllabic words composed of an epenthetic vowel are perceived as 'shorter than' ordinary syllables (Bosch 1995:11).

References

- Clements, G.N. (1986). Syllabification and epenthesis in the Barra dialect of Gaelic. In K. Bogers, H. van der Hulst & M. Mous (eds.), The Phonological Representation of Suprasegmentals: Studies on African Languages Offered to John M. Steward on His 60th Birthday, 317-336. Dordrecht: Foris.

- Bosch, A. (1995). A gestural analysis of epenthesis in Scottish Gaelic. Ms., University of Kentucky.