NOTES ON THE CHOMSKY-HALLE PHONOLOGICAL FEATURE SYSTEM
Traditional categories and approaches
There are various ways in which languages have been transcribed. For example, one may symbolize one aspect of a contrast (e.g. length) or some other quality. One may show only underlying phonemes, or only some allophonic differences. Traditionally, a broad transcription designates one that uses a simple set of symbols, whereas a narrow transcription exhibits more phonetic detail. Diacritics increase precision, such as for indicating voicelessness, or a dental rather than alveolar sound.
Conventionally, transcription has two aspects, viz. the text itself and its interpretation. One set of conventions ascribes general phonetic values to symbols. Ladefoged (1993) thus regards symbols as approximate specifications of the articulations involved. The other set of conventions is constituted of the rules which specify allophones occurring in different circumstances. A systematic phonetic transcription shows all the rule-governed alternations among the sounds.
Rather than describing the sounds of English in this traditional manner, one can also specify the features of which they are composed. In phonology, a feature--or, more properly, a distinctive feature--"is the smallest contrastive unit proposed to explain how the sound system of a language is organized" (Crystal, 1992, p. 107). Voicing, tongue height, and lip rounding are examples of phonological distinctive features.
Sets of features have been used to classify the sounds of English into, e.g., voiced (+ voice) and voiceless (- voice),
coronal (+ anterior) and coronal (- anterior), nasal (+ nasal) and nasal (- nasal), lateral (+ lateral) and lateral (-lateral), sibilant (+ sibilant) and sibilant (- sibilant), back (+ back) and back (- back), syllabic (+ syllabic) and (- syllabic), stricture (stop, fricative, approximant), height (maximum, 4 height, 3 height, 2 hei...