Rules and schemas in the development and use of the English past tense

RULES AND SCHEMAS IN THE DEVELOPMENT AND USE OF THE ENGLISH PAST TENSE Joan L. BybeeDan I. Slobin State University of New York,University of California, BuffaloBerkeley Consistent error patterns in English past-tense forms are reported for three age groups: preschoolers, 8-10-year-olds, and adults. It is argued that, although irregular forms are rote-learned, speakers make generalizations about such forms. Such a generalization is defined as a schema which describes general phonological properties of a morphological class, and is used in organizing and accessing the lexicon. Schemas for the English past tense develop and change with age, yielding implications for both acquisitional and diachronic theory.* 1. English verbal morphology is rather restricted, compared to that of a fullfledged inflectional language, since it offers only four inflectional morphemes: the 3rd singular present, the past tense, the past participle, and the progressive. Thus it provides no opportunity to study the complex interaction between intersecting inflectional categories within a paradigm, such as person, number, mood, and tense. It does, however, provide the opportunity for a study of a different sort: although English has a demonstrably productive process of suffixation for past-tense formation, in the form of -ed, it also has many irregular verbs whose past tense is formed in some cases without suffixation, and in others with changes of vowel (or, less commonly, consonant) in the stem. These irregular verbs are relatively few in number. Bloch 1947 identifies about 200 (several of which are archaic), but thousands of verbs form their past tense by adding an allomorph of -ed. Although irregular verbs are relatively insignificant as to type frequency, the picture is quite different when token frequency is considered: these are among the most frequent verbs of the language . Of the 30 most frequent past-tense forms (Kucera Zell Greenberg tested the preschool children and analysed the results; Amy Strage and Ann Eisenberg tested the adults; Mr. B. Harris and his third-grade class at Cornell Elementary School, Albany, CA, played the 'verb game' with us; Tanya Renner, Carol Lynn Moder, and BongHee Choi helped with the statistics; and Brody Hooper let us use his verb forms and his friends. While this research was being conducted, loan Bybee held a Postdoctoral Research Training Fellowship from the Social Science Research Council. This research was also supported by the University of California Committee on Research and the Institute of Human Learning, University of California, Berkeley. 265 266LANGUAGE, VOLUME 58, NUMBER 2 (1982) The difficulty presented by these irregular verbs, both for the learner and for the linguist, is that so many different irregularities occur. Though most irregulars are characterized by a vowel change, a number of different changes occur, e.g. stick/stuck, sing/sang, keep/kept, grow/grew. In addition, some verbs add a t or d; and some have consonant changes as well, e.g. makelmade. Jespersen 1942 lists ten classes of irregular past-tense formations, and dozens of sub-classes, while Bloch comes up with twenty conjugation types. Such variety increases the difficulty of the learning task, but at the same time provides us with an opportunity to investigate the interaction of rule and rote-learning in morphology. It is clear that some rote-learning is necessary in acquiring irregular past-tense forms: the pair go/went could not be learned in any other way. What is not so clear is whether speakers must learn all irregulars by rote, or whether...