This paper reports results from three experiments on the English vocabularies of Hiroshima University students and their psycholinguistic features. In Experiment 1, the sizes of receptive vocabularies were estimated by means of the regression equations for correct responses and word frequencies: the number of English lexical items which average Hiroshima University students knew was estimated to be 3,293; those which upper-level and lower-level students knew were estimated to be 5,652 and 2,473, respectively.
In Experiment 2, two types of receptive vocabulary test were constructed: a retrieval test and a recognition test. The former was of the same kind as given in Experiment 1, where the subject attempted to retrieve the lexical meaning of an English word and wrote it in Japanese. The latter was a multiple-choice test in which (a) a target, (b) a semantically related distractor, (c) a phonologically and/or orthographically similar distractor, (d) a distractor semantically related to distractor c were the choices. Main results were (1) mean percent correct was significantly higher on the recognition test than on the retrieval test even though the scores for the latter were adjused, and (2) both semantic and 'phonological and/or orthographic' errors were observed in both tests. The architecture and psycholinguistic nature of the mental lexicons in students were discussed.
Experiment 3 was concerned with possible determiners of learnability of English words and with a further issue about semantic and 'phonological and/or orthographic' errors. The familiarity of the Japanese equivalents of the English targets was significantly correlated with the mean percent correct responses. The mean semantic error rate was not as high as that in Experiment 2.
In light of the outcome of the three experiments, theoretical and practical implications for English language education focusing on Hiroshima University students are briefly discussed.