rules that specify allophonic detail are general and may be applicable to several
phonemes.
As it is now stated, rule 1 is not precisely in the form in which phonological
rules must be written. Two changes are needed: first, the specifications + (plus)
and - (minus) must be added to the features; and second, a more precise and
empirically justified set of distinctive features must be employed. For justification
of the intrinsic content of a more appropriate set of distinctive features, see
Linguistics, pp. 1 10- 1 18.
We have already tacitly assumed that the features that make up phonemes are
binary. That is, each one can have two values: + and -. For example, the feature
of voicing can appear either as voiced ([+voiced]) or as voiceless ([-voiced]).
There are two reasons for claiming that distinctive features are binary. First,
people perceive features categorically-as being either present or not present-and
not as a continuum. If a voiced sound-for example, b-is produced with some
degree of nasalization, listeners perceive either m or b-they do not perceive some
intermediate third sound. In experiments with synthetic speech, for example,
subjects will hear either mad or bad, depending on the amount of nasality that was
supplied in synthesizing the initial consonant. Second, only the absolute values +
and - are needed for the proper statement of phonological rules. For example,
one never has to state that a phonological rule is applicable if m has 3 degrees of
nasality, n has 2 degrees of nasality, g has 4 degrees of nasality, and so forth. If
the class of nasal phonemes participates in a rule, only the feature [+nasal] (or
[-nasal]) is needed to specify that class.
One other point needs to be made before we state rule 1 in its final form. The
consonants in part D of rule 1 all share the property of being voiced; that is, they
all have the feature [+voiced]. These consonants thus constitute a natural class of
phonemes that can be defined by a small number of distinctive features. (See
Linguistics, pp. 12 1 - 124, for additional discussion and motivation of the notion
"natural class. ") Several problems in this workbook (exercises 3.3-3.6) require a
phonological rule to be stated in terms of the distinctive features that dejne a
natural class of phonemes. Your task in these cases is to find a set of features that
will include all of the phonemes in the class and exclude all of the other phonemes
in the language. A chart listing the distinctive features of all the phonemes needed
for the exercises has been included in appendix 4. (For a description of the
features themselves, see Linguistics, pp. 1 14-1 1 8.)
To return to rule 2: The phonemes that participate in this rule can be found in
appendix 4. It is similar, then, to the intermediate stages you will go through in
formulating your rules in exercises 3.3-3.6. In rule 2 the feature that uniquely
specifies all of the sounds that undergo vowel lengthening is the feature [+syllabic].
The features that uniquely specify the set of phonemes to the right of the focus
bar (part D of the rule) are [+consonantal] and [+voiced]. The feature that
specifies part B of the rule is [+long]. (Since there is some question concerning
how the feature [+long] is to be represented in phonological theory, it has not
been listed as a feature in appendix 4.)
The final form of the rule can now be given as follows:
- English Vowel-Lengthening Rule (final form)