Irregular morphology refers to word forms — typically inflections and derivations — that cannot be produced by applying the regular, productive grammatical rules of a language. They are exceptions to the pattern: the English past tense of “go” is “went” (not “goed”), the plural of “mouse” is “mice” (not “mouses”), the Japanese verb する (suru) conjugates unpredictably relative to most other verbs. Irregular forms must be stored and retrieved from memory individually rather than computed by rule, making them a significant challenge in both L1 acquisition and SLA — particularly because they are often among the most frequently used words in the lexicon.
In-Depth Explanation
What makes a form irregular:
A morphological form is irregular when it cannot be derived by applying the productive rules available in the grammar to the base form. In English:
| Type | Regular form | Irregular form |
|---|---|---|
| Past tense | walked, played, talked | went (go), ran (run), sang (sing), broke (break) |
| Plural | cats, dogs, houses | mice (mouse), geese (goose), children (child), oxen (ox) |
| Comparative | taller, wider | better (good), worse (bad) |
| Participial | played, talked | done (do), gone (go), been (be) |
Why irregular forms survive:
The dominant explanation is frequency: irregular forms are so common in everyday speech that the memorized forms remain accessible and are not gradually replaced by analogy to regular patterns. The word “go” appears thousands of times more frequently in everyday speech than, say, “stride” (whose irregular past “strode” is already declining in favor of “strided” in informal use). Frequency maintains irregular forms through constant practice; low-frequency irregular forms gradually yield to regular analogy over time.
Suppletive forms:
The most extreme irregular morphology is suppletion — where the form for a grammatical category is etymologically completely unrelated to the base: go/went, be/was/were, good/better/best. These represent the merger of two or more originally distinct lexical items into a single paradigm.
SLA and irregular morphology:
L2 learners show characteristic error patterns with irregular morphology:
- Initial correct use (rote memorization): In early stages, learners may produce frequent irregular forms correctly because they are encountered often and memorized as whole-form chunks (“went,” “said”).
- U-shaped development (overgeneralization): As learners acquire the productive rules of the language (e.g., “-ed for past tense”), they often begin applying them to irregular verbs — producing forms like “goed,” “runned,” “wented.” This apparent regression is actually evidence of rule learning.
- Gradual differentiation: With continued input and explicit learning, learners sort lexical items into the rule-governed and the exceptional, eventually storing irregular forms correctly.
The U-shaped development pattern is well-documented in both L1 and L2 morphological acquisition.
Japanese irregular morphology:
Japanese has relatively regular verb morphology for Class 1 (godan-verbs) and Class 2 (ichidan-verbs) verbs but two completely irregular verbs: する (suru, to do) and くる (kuru, to come). These verbs produce highly frequent forms that must be memorized individually:
- する: します, して, しない, すれば, しよう, させる, される, etc.
- くる: きます, きて, こない, くれば, こよう, こさせる, こられる, etc.
Additionally, Japanese i-adjectives and na-adjectives have different conjugation patterns, and the copula だ/です has irregular forms across registers and tenses.
History
Morphological irregularity has been studied in the context of the “dual mechanism model” (Pinker & Prince, 1988; Pinker & Ullman, 2002), which proposed that regular morphology is computed by a rule-based system and irregular morphology is stored in associative memory (lexicon). This was challenged by single-mechanism connectionist models (Rumelhart & McClelland, 1986) that simulated both regular and irregular past tense acquisition in a single network without separate rules. The debate between dual-mechanism and single-mechanism accounts has driven decades of psycholinguistics research.
Common Misconceptions
- “Irregular forms are just mistakes or historical accidents.” Many irregular forms are systematic: English strong verbs follow predictable vowel shift patterns (sing/sang/sung, ring/rang/rung, drink/drank/drunk). They are “irregular” relative to the weak (regular) verb pattern but have their own internal regularity.
- “The solution is just to memorize them all.” For high-frequency irregulars, rote memorization is indeed part of the answer. For lower-frequency items, pattern recognition within irregular classes is more efficient.
- “Overgeneralization errors mean the learner is regressing.” U-shaped development is a marker of internalization of productive rules — a positive sign, not a regression.
Practical Application
For Japanese learners:
- Memorize the conjugation tables for する and くる early and comprehensively — these verbs appear constantly. Every verb compound with する (studying = 勉強する, cooking = 料理する) will produce all these forms.
- For suru-compound verbs specifically, learn the full conjugation of する once and it applies across the category (勉強する → 勉強します、勉強して、勉強しない, etc.).
- Use high-frequency irregular forms in sentence-context Anki cards rather than isolated conjugation drills to build contextually appropriate recall.
Related Terms
Sources
- Pinker, S. & Ullman, M.T. (2002). “The past and future of the past tense.” Trends in Cognitive Sciences 6(11): 456–463.01990-3) — dual mechanism model summary.
- Rumelhart, D.E. & McClelland, J.L. (1986). “On learning the past tenses of English verbs.” In McClelland & Rumelhart (eds.), Parallel Distributed Processing, Vol. 2. MIT Press. — connectionist alternative model.
- Marcus, G.F. et al. (1992). “Overregularization in language acquisition.” Monographs of the Society for Research in Child Development 57(4). — U-shaped acquisition documentation in L1.