Questions: Allomorphy and Phonologically-Conditioned Alternation

Question 1

An English speaker correctly produces 'dogs' [dɒgz], 'cats' [kæts], and 'buses' [bʌsɪz] without being explicitly taught these forms. A linguist analyzing these plural forms should conclude:

• A. These are three distinct plural morphemes with slightly different grammatical meanings
• B. These are three allomorphs of a single plural morpheme, with the choice conditioned by the phonological environment of the stem — voiced non-sibilant → [z], voiceless non-sibilant → [s], sibilant → [əz]
• C. These forms must be stored individually as lexical entries because the variation is unpredictable
• D. These are morphologically-conditioned allomorphs, like man/men, that require lexical class specifications

Answer: (B)

The distribution of [z], [s], and [əz] is entirely predictable from the phonological environment of the stem — it is phonologically-conditioned allomorphy. Positing three separate morphemes misses the generalization that they are in complementary distribution and their distribution follows from the phonology of the language. A single underlying form plus a phonological rule (or set of constraints) explains the three forms with less machinery than three separate entries. Native speakers extend this pattern automatically to novel words (e.g., 'blicks' → [blɪks]), proving the rule is productive.

Question 2

Which of the following is the clearest example of suppletive allomorphy?

• A. The English plural alternation: [z] in 'dogs,' [s] in 'cats,' [əz] in 'buses'
• B. The English present/past alternation in 'go' / 'went'
• C. The [f]/[v] alternation in 'leaf' / 'leaves'
• D. The vowel change in 'sing' / 'sang'

Answer: (B)

'Go' and 'went' are the canonical example of suppletive allomorphy: the past tense form is historically from an entirely different verb (*wend*), and no phonological rule derives 'went' from 'go.' The relationship must be stored as an arbitrary lexical pairing. By contrast, the plural alternation ([z]/[s]/[əz]) follows a productive phonological rule; 'leaf/leaves' reflects a semi-regular morphological class; 'sing/sang' follows an ablaut pattern shared by other strong verbs. Suppletive allomorphs are the residue that cannot be derived by rule — pure lexical accidents of history.

Question 3

Suppletive allomorphs like 'go/went' and 'good/better/best' must be stored as separate lexical entries in the grammar because no productive phonological or morphological rule can derive one form from the other.

• T. True
• F. False

Answer: True

Correct. The goal of morphophonological analysis is to reduce surface alternations to a single underlying form plus rules wherever possible. When no rule captures the alternation — because the relationship is historically arbitrary — the grammar has no choice but to list the forms separately. Suppletive forms are identified by the diagnostic that native speakers do not extend them to new words: you cannot form the plural of a new noun using vowel change (*'mouses' → [maɪs]?), and children must learn 'went' explicitly rather than deriving it from 'go.'

Question 4

Because [z], [s], and [əz] are three different surface forms, they represent three separate plural morphemes in the grammar of English.

• T. True
• F. False

Answer: False

This is the key mistake in analyzing allomorphy. Having different surface forms does not imply different underlying morphemes. The three forms are in complementary distribution (each appears in a distinct, non-overlapping environment) and collectively cover all plural contexts — the hallmark of allomorphs of a single morpheme. Positing three morphemes makes the grammar redundant: it lists three entries where one plus a rule would suffice, and it fails to capture the generalization that all three mark plurality. Morpheme identity is determined by function and distribution, not surface form.

Question 5

Why do linguists prefer to analyze the English plural [z/s/əz] as one morpheme with phonologically-conditioned allomorphs rather than as three separate morphemes? What analytical principle motivates this choice?

Model answer: The three-morpheme analysis duplicates information: it lists three separate entries that all mean 'plural,' without explaining why they appear in mutually exclusive environments. The one-morpheme analysis captures the generalization that the alternation is entirely predictable from the phonological context of the stem — voiced non-sibilant, voiceless non-sibilant, sibilant — using a single rule rather than arbitrary listing. The guiding principle is economy: prefer the analysis that captures alternations as productive rules over one that treats them as coincidental. Suppletive allomorphs are the exceptions that cannot be derived by rule; phonologically-conditioned allomorphs are the norm that reveals the language's phonological grammar.

This is the core methodological principle of morphophonology: reduce surface complexity to underlying simplicity through rules. The morpheme inventory should be as small as possible while still accounting for the data.