A phonological rule causes /p/, /t/, and /k/ to become voiced before a voiced consonant. Using feature notation, how should this rule be stated?
AList all three sounds explicitly: /p/, /t/, /k/ → voiced / _[+voice]
BChange [−voice] to [+voice] for all segments specified as [−voice, −continuant] (voiceless stops)
CAdd [+voice] to all consonants in the environment of a voiced segment
DUse the feature [+labial] to capture the class of sounds that undergo this change
The rule applies to a natural class — voiceless stops, defined by [−voice, −continuant] — so the rule should reference those features, not list individual sounds. Feature notation reveals that only one feature value changes (voice: − → +) while all other feature values remain constant. This is both more general and more explanatory than a list. Option A is how rules were stated before feature theory; option C is too broad (it would apply to all consonants, including fricatives); option D is wrong because [+labial] only captures /p/, not /t/ and /k/.
Question 2 Multiple Choice
A linguist proposes that /m/, /n/, and /ŋ/ form a natural class. Which feature best defines this class?
A[+labial] — all are produced with the lips
B[+voice] — all are voiced consonants
C[+nasal] — all allow air to flow through the nasal cavity
D[−continuant] — all are produced with complete oral closure
The feature [+nasal] uniquely defines the class {m, n, ŋ} — all three are nasal consonants in which the velum is lowered and air flows through the nose. [+labial] applies only to /m/; [+voice] includes many other consonants (b, d, g, v, z, etc.); [−continuant] includes all oral stops as well. A natural class must be the most restrictive (smallest) set of features that picks out exactly the sounds that behave together. The [+nasal] specification is both necessary and sufficient to identify this class.
Question 3 True / False
Feature matrices reveal not just what changes in a phonological process but also what remains constant — and this is as analytically important as identifying the change.
TTrue
FFalse
Answer: True
When a rule changes /t/ to /d/, the feature matrix shows that only [voice] changes from − to +. Every other feature — place of articulation, manner, nasality — remains identical. This is crucial: it shows the process is minimal (only one property is affected) and that the segment's identity is largely preserved. Symbol-based notation (/t/ → /d/) hides this; matrix notation makes it explicit. Identifying what stays constant is part of what distinguishes a voicing rule from an entirely different process.
Question 4 True / False
Feature matrices are simply a more convenient notation for writing out individual sounds — any phonological rule that can be written with features could just as well be written by listing the affected segments individually.
TTrue
FFalse
Answer: False
Feature matrices do more than provide notation — they reveal the underlying structure of natural classes and phonological processes. A list of affected sounds (/p/, /t/, /k/ → voiced) captures the facts but offers no explanation for why these sounds behave together or what changes. Feature notation shows that the sounds form a natural class (voiceless stops), that only one property changes (voicing), and that the rule generalizes to any new voiceless stop in the language. A list cannot predict behavior for novel sounds or across languages, while features can.
Question 5 Short Answer
Why can't phonological rules be stated simply as lists of individual sounds? What does the feature matrix representation reveal that a list of symbols cannot?
Think about your answer, then reveal below.
Model answer: A list of sounds captures which segments are affected but cannot explain why those segments pattern together or predict what other sounds should behave the same way. Feature matrices reveal the natural class membership — the shared features that define which sounds undergo the rule — and show exactly which property changes while others stay constant. This makes rules general, predictive, and explanatory rather than merely descriptive.
The point of feature theory is that phonological generalizations are not arbitrary lists — they reflect systematic properties of the articulatory-acoustic space. When a rule voices voiceless stops before a voiced consonant, the sounds affected are precisely those sharing [−voice, −continuant], not an arbitrary assortment. Feature notation encodes this systematicity. It also makes cross-linguistic comparison possible: the same rule, written in features, can be compared across languages, revealing universal tendencies in phonological processes.