English Words of Japanese Origin: A Study of Dictionaries and Corpora

This thesis investigates English words of Japanese origin through different dictionaries and corpora. The purpose of the thesis is to explore which Japanese borrowed words are listed in some British and American dictionaries, what kind of changes they undergo when they are listed in dictionaries, and which Japanese-based loanwords are used most frequently in the UK and the US. In addition, a comparison in the use of Japanese loans between British and American English is made owing to a lack of previous research in the area. This is a quantitative corpus study, and frequently used Japanese-based loanwords collected chiefly from the third edition of OED and also six other dictionaries are examined through British and American corpora. The corpora employed for this study are BNC1994, Spoken BNC2014, COCA, and COHA. First, Japanese loans in the OED are sorted by frequency and the list of the top 100 Japanese loanwords is created. Then, each term is examined carefully by checking their linguistic labels and removing some rare words from the list. After that, other British and American dictionaries are used to supplement the list for my corpus study with some newer Japanese-based loanwords. All the loans listed in more than five dictionaries are candidates for my corpus study, and eventually my list of Japanese loans for my study contains 113 words in all, as shown in Table 8. This study also looks into the use of linguistic labels in several British and American dictionaries, and the results suggest that not many Japanese loanwords are marked: only four different labels, domain, time, slang, and regional labels are detected among Japanese borrowed words analysed for this study. Interestingly, however, some loans to which labels are attached, such as skosh marked [American slang], and soya marked [chiefly British], are examined through British and American corpora and the findings are in accord with the labels. Furthermore, loanword transformations introduced by Daulton are discussed in greater detail in this thesis. The main finding is that numerous loans undergo an orthographic change and receive a penultimate accent when they are listed in dictionaries. A unique finding is napa (cabbage), which undergoes semantic restriction. The results of my corpus study suggest that the numbers of tokens are rather small in the two British Spoken corpora, BNC1994DS and the core set of Spoken BNC2014. Nevertheless, some changes in the use of Japanese loans in British spoken English over 20 years are observed, such as the very high frequency of the Japanese martial art karate in BNC1994DS and highly ranked food-related loans, such as sushi, soya, tofu, and ramen, in the core set of Spoken BNC2014. The search for Japanese loans through the American corpora COHA and COCA yields much more occurrences than do British corpora and it is worth comparing the results from these two corpora diachronically. Interestingly, in both corpora, soy is the most frequently used Japanese loan, and as many as 22 of the top 30 Japanese loanwords overlap, despite differences in their rankings. It is worth noting that one of the very unfamiliar loanwords for native Japanese, tycoon, is found to have been used constantly in American English for many years, when looking into the frequency distribution through COHA and COCA. As regards the comparison of British and American corpus results, it is interesting to observe that BNC1994 shows a higher frequency of Japanese borrowed words related to economy and business, such as Nikkei, zaibatsu and keiretsu, as well as Japanese traditional sports, karate and judo. On the other hand, American corpora, especially COCA, yield more instances of food-related loans in addition to tsunami, Zen, anime, and manga. Combining dictionary study with corpus study leads to careful and empirical observations, and it is fascinating to find out that the results from the two studies achieve consistency