Feature model

This paragraph will take us back to both consonant and vowel systemic comparisons between the two languages, or any number of contrasted languages, for that matter. The method consists in comparing distinctive features only, instead of comparing long lists of phonemes. Thus, for instance, we could just start comparing the English and the Russian consonant systems by juxta­posing the two distinctive features: alveolar vs. dental, or fortis vs. lenis.

In phonology the notion of distinctive features describes sound units as actually composed of simultaneously occurring features, a bundle of distinctive features. The features were drawn from a universal set of distinctive features. Thus the same features were used for classifying vowels and consonants. The features are binary: vowels are marked, for instance, as either [+high] or [—high], [+low] or [—low]. One of the principles is that the system should be economical. Mid-open (or mid-high) vowels were marked [-high] and [—low], which put them in a separate class (without using additional labels).

Any sound could thus be matched with any other sound which exhibited similar properties simply by specifying a rule involving that specific feature or a subset of features. For example, Chomsky and Halle's "Sound Pattern of English" (1968) has a rule which deals only with those sounds in English which have the feature of [+coronal]; this involves only those sounds (dental, alveolar and palato-alveolar) which involve the tongue tip and blade (the crown, or "corona"), thus excluding those which are [—coronal] (labial and velar). This grouping can obviously be neatly and accurately specified with just one feature.

Distinctive feature model makes the description of a particular sound or a group of sounds more economical. It was aimed at discovering elements universal to all languages. Features are usually considered to be binary, for instance: [+high], [—high], [+voiced], [—voiced]. R. Jakobson, G. Fant and M. Halle established 12 binary distinctive oppositions, with the help of which, as they claim, it is possible to classify the phonemes of any language. For the English language the following 9 oppositions are sufficient: vocalic - non-vocalic, consonantal — non-consonantal, compact — diffuse, grave — acute, flat — plain, nasal — oral, tense - lax, dis­continuous — continuant, strident — mellow.

In phonetics and phonology the features are based either on articulatory analysis (cf. IPA — International Phonetic Alphabet) or on acoustic details of the sounds produced by speakers, the analysis of the spectra of sounds (Jakobson, Fant and Halle). The feature model associated with transformational generative grammar is a mixture of acoustic and articulatory details. It is less accepted in general use than IPA. Generative phonology continues to develop as a purely American phenomenon (see "Dynamic models" below).

Дата добавления: 2014-01-11