Multiple response free-word association and the syntagmatic-paradigmatic shift in Japanese adults learning English as a second language

Research has shown that English-speaking adults tend to give response associates of the same grammatical and semantic class as the stimulus word on a word-association test, whereas children typically do not, instead responding syntactically (Thumb and Marbe, 1901; Esper, 1918; Deese, 1962; Fillenbaum and Jones, 1965; Entwisle, 1966). This pattern of responding seems to hold for many languages, including French, Italian, Spanish, Polish and German. This shift from syntactic responding to same form-class responding is often referred to as the syntagmatic-paradigmatic shift. Moran and Murakawa (1968) and Moran (1973) found that Japanese adults responding in Japanese to word-association stimuli respond syntactically, that is, they seem not to experience the S-P shift. Two main hypotheses were tested. 1. That Japanese adults beginning to learn English would give predominantly syntagmatic responses to nouns, verbs, and adjectives in English, and thus differ significantly from native-English speakers. 2. That advanced Japanese students of English would give fewer syntagmatic responses in English than the beginner group and more closely parallel native-English speakers. Two subsidiary hypotheses tested were: 3. That the absolute count of responses to stimuli would correlate with scores obtained on a test of language proficiency by the Japanese subjects. 4. That the primary responses of the Japanese advanced group would more closely resemble the native-English responses than would those of the Japanese beginners. A timed multiple-response free-word association test comprising eight nouns, eight verbs, and eight adjectives was administered to forty adult native-English university students and forty-seven Japanese ESL students. The Japanese students were also given the University of British Columbia Language Institute Placement test. On the basis of the results of this language test, the top fifteen scorers were assigned to the Japanese advanced group, while the bottom fifteen scorers were assigned to the Japanese beginner group. The scores correlated well with the language instructor's own tests of language ability. The word-association tests were scored by two independent markers, and mean paradigmatic response tables were compiled. Analyses of variance and Pearson's product moment correlations were performed on the appropriate data. Results partially supported the hypothesis that Japanese beginners would respond syntagmatically to nouns, adjectives and verbs. Because this group responded paradigmatically to nouns the conclusion reached was that they were paralleling native-English-speaker development. There was no statistically significant difference in paradigmatic responding between the native-English group and the Japanese advanced group, the conclusion being that the more fluent a foreign student becomes in English, the more paradigmatic responses will be given. The absolute count of responses correlated overall with scores on the language test, but in isolation the Japanese beginner group responses did not correlate with the language-test score. It was thought that the reason for the non-correlation was a sampling error. The total frequency of the three most frequent responses for nouns was identical between the Japanese groups, but for all form-classes the Japanese advanced group was much closer to the native-English group. This convergence of commonality is taken as evidence that idiodynamic sets are constrained by the language being used.Education, Faculty ofLanguage and Literacy Education (LLED), Department ofGraduat