Phonology: case studies
By Scott Myers and Megan Crowhurst
Department of Linguistics at the University of Texas
Derivational analysis of the palatalization pattern
In a derivational framework, the rules of analysis for a complementary distribution, like the one involved in Kinyarwanda palatalization, are the following:
- The default class of sounds are the phonemes. Only the default class occurs in underlying representations.
- The restricted class of sounds are not phonemes, but are created only by a phonological rule.
- There must be a rule changing the default class to the restricted class, in the context in which the restricted class occurs.
In the case of Kinyarwanda postalveolars, the default class is the velars, so these are the phonemes: /k,g/. The underlying representations of all words with postalveolar stops must have velar stops: /gukata, kugenda, umugí/. The palatals are the restricted class, so they must be absent from underlying representations, and created only by a phonological rule. This rule must change the default sounds (the velars) into the restricted sounds (the palatals) in the context where the latter occur (before a front vowel). In formal notation using distinctive features, the following is a rule that accomplishes this:
[-son, +back] → [-back] / ___ [-cons, -back]
To the left of the arrow is the target of the rule, which defines the class of sounds that are subject to the operation. In this case that is the class of back (i.e. velar) consonants. To the right of the arrow is the change, which defines how the output of the rule differs from the input, i.e. how the restricted class of palatal stops differs in features from the default class of velar stops. In this case, the palatal stops are identical to the velar stops except that they are [-back]. To the right of the slash is the context. The blank is the focus bar, which represents the location of the target. The focus bar is before [-cons, -back], so the rule applies if a back obstruent occurs before a front vocoid. It changes that back obstruent into a front one.
The following derivations illustrate how this analysis would apply to representative forms:
| Underlying representation | /gukata/ | /kugenda/ | /umugí/ |
| Rule | -------- | /kuɟenda/ | /umuɟí/ |
| Surface representation | [gukata] | [kuɟenda] | [umuɟí] |
We start in underlying representations with velar stops and no palatal stops. The phonological rule changes velar stops to palatals only if they are before a front vowel. Elsewhere, the stops remain unchanged, so they remain velars. The result is the observed pattern of distribution: palatal stops occurring only before front vowels and velar stops only elsewhere. The rule only changes the place of articulation of the stop, so all the other properties remain unchanged. The voiced phoneme /g/ is changed before a front vowel to the voiced stop [ɟ], and the voiceless stop /k/ is changed to the voiceless stop [c].
Reference
Chomsky, N. and M. Halle (1968). The Sound Pattern of English. Harper and Row, New York.
- [son] is an abbreviation for [sonorant].
- [-son] picks out the set of obstruents, produced with an oral constriction great enough that there is turbulent airflow at normal airflow rates.
- [+son] picks out the set of sonorants, produced without such turbulent airflow.
- [+back] denotes the set of back sounds, with a constriction made with the tongue body behind the hard palate. It includes back and central vowels and glides, velars, uvulars, and pharyngeals.
- [-back] denotes the complement set. It includes front vowels and glides, palatals, and sounds that don't involve a tongue body constriction.
- [cons] is an abbreviation for [consonantal].
- [-cons] denotes the set of sounds made without significant constriction in the oral cavity, i.e. vowels and glides.
- [+cons] denotes the set of sounds made with significant oral constriction,. i.e. obstruents, nasals, and liquids.