Abstract:
This dissertation examines the nature of the high-frequency English words that have been adopted into Japanese and their effect upon L2 English vocabulary acquisition. The findings include:
(i) High-frequency English vocabulary is of great use to learners in an L2 situation, as it provides good coverage of even unsimplified texts.
(ii) It appears that just 2,000 to 3,000 word families provide adequate coverage of general, unsimplified texts.
(iii) The BNC 3000 (Nation, forthcoming), a frequency list based on the British National Corpus, is an appropriate fequency list on which to base other research.
(iv) Although loanwords are widespread and deeply integrated into Japanese, questions remain regarding their comprehensibility to Japanese.
(v) The "paradox of cognates" -- that cognates appear to both hinder and aid acquisition -- can be resolved by seeing the role of transfer errors from a developmental perspective. Regardless of semantic overlap, formal similarities between L1 and L2 provide a "lexical peg" which tends to propel acquisition forward.
(vi) There is substantial empirical evidence that English loanwords in Japanese assist the acquisition of their related borrowed words in English. Moreover, loanwords were found in disproportionate numbers in learner production.
(vii) A thorough examination of the comprehensibility of two dictionaries of "common" loanwords showed that researchers needing a fairly complete listing of today's common gairaigo can not depend upon a single source of data.
(viii) Daulton's (1998) findings concerning the overlap between common loanwords in Japanese and high-frequency English, vocabulary are unreliable because of lack of validity in his list of loanwords.
(ix) While there is a profusion of everyday loanwords in the Japanese media, there are likewise a large number of poorly understood items. Among these can be found various nonce borrowings and idiosyncratic usages, which are a propelling force in the growing lexicon.
(x) Using four types of lexical data including a newspaper corpus, it was found that 45.3 percent of BNC 3000 word families correspond to loanwords whose comprehensibility was confirmed with Japanese university students.
(xi) Japanese learners are able to generalize their cognate-based knowledge of a word family member to the other inflected forms and derivations present in the word family. This validates the utility of high-frequency words.
(xii) Most of the correspondences between high-frequency English and common loanwords appear to be of high quality.
Thus there appears to exist in the Japanese lexicon a very useful "built-in lexicon" of useful cognates for high-frequency English vocabulary.
