Japanese language education is an area of research that deals with themes that span language, education, and society. This can be confirmed by the history of Japanese language education over the past 40 years. I will consider what kind of facts the frequency analysis model of quantitative linguistics reveals about various research issues in Japanese language education, based on specific cases. In particular, the following three facts will be revealed: 1) research based on the survey and analysis of a large corpus of native speaker data has the potential to resolve issues in Japanese language education across language and education, 2) research based on the survey and analysis of a learner's corpus has the potential to address issues in Japanese language education, especially second language acquisition, 3) research based on the survey analysis of small-scale, self-made corpus is effective in solving the problems of Japanese language education in society.

View full abstract

In this paper we identified problems in the connective expression use of Japanese learners based on an analysis of connective expression usage patterns and proposed an appropriate corrective teaching method. Using a list of connective expressions sourced from large-scale corpora, we first classified them into base or derived forms and identified their demonstrative status, presence in dictionaries, and semantic function. We then compared tendencies in connective expression use within 5 different corpora, including learners’ bachelor’s theses and other compositions. Compared to other native corpora, learners were observed to use connective expressions at the highest rate. Furthermore, the most frequently used expressions were sourced from a limited repertoire of semantic functions. Finally, as compositions written by Japanese learners of similar proficiency were observed to employ different patterns of use across registers we suggest incorporation of multiple registers as an essential element for any research utilizing learner compositions.

View full abstract

This study, based on the analysis of the I-JAS data, investigates the influence of learning environments and classroom instructions on the accuracy and complexity of the Japanese sentences used by JSL learners. We analyzed the speech data of 60 learners in three groups; JFL in China, JSL-Classroom learners in Japan and JSL-Natural learners in Japan. We measured the accuracy of their oral data, examining error-free T-units and also measured the complexity of sentences, counting the numbers of different clause types in T-units. The findings are as follows: First, JSL-N learners speak Japanese more correctly than those studying in other environments. Second, JSL-N learners use a greater number of complex sentences with subordinate clauses. Third, in JSL-C, similar to JFL, subordinate clauses are not as numerous as JSL-N, but compound sentences are seen as many as in JSL-N. The results indicate the influence of the learning environments on the accuracy and complexity of learner's Japanese, and they may also show that JSL-C learner's Japanese represents a transitional stage prior to the acquisition of a grammar of greater complexity.

View full abstract

This paper examines the use of Principal Component Analysis (PCA) for measuring and clustering the stylistic information of adverbs. The target items were basic adverbs for Japanese learners and semantically synonymous adverbs that serving different stylistic functions. There were 164 items, in total. The standardized frequency of each adverb in 14 genres of BCCWJ was calculated, and then a Principal Component Analysis (PCA) are conducted. Using the two primary principal component scores, the adverbs were divided into groups. An axis of “Bidirectional versus Unidirectional” was obtained as PC1 and an axis of “Self-Expression versus Public Announcement” was obtained as PC2. This result matches that of stylistic study of English (Biber 1988). Adverbs were divided into four groups with the two axes: Group A “Adverbs for self-expression”, Group B “Adverbs for Public Expression”, Group C1 “Adverbs for Sophisticated Communication”, and Group C2 “Basic Adverbs for Daily Communication (Basic).” Thus, this analysis can divide synonymous adverbs into stylistically different groups.

View full abstract

Unlike balanced corpora, numerous learner corpora are not satisfactorily large-scale or representative. However, because previous analyses were conducted fundamentally by using the same method as balanced corpora, the present study examined the statistical validity of the existing method. According to the findings, linguistic units (e.g., misuse expressions, sentences, and morphemes) are consistently used by each learner. For this reason, there is no satisfactory assumption of independence for statistical analyses, and there are cases in which the outliers distort the results. Thus, the method of collecting each learner’s frequency and analyzing learners as observation units is a valid approach to replace the existing method.

View full abstract

This paper explains the questionnaire survey service and questionnaire creation tool provided by Internet-related companies except for the Google form and CGI explained in this series (8) and (9). This report is focusing on self-type internet research (DIY Research) in consideration of reducing cost and the possibility of using in academic research fields. Self-type Internet research means that researchers themselves design and create online questionnaire, collect, aggregate, and analyze the answers using the tools provided by the internet survey company. Questant, SurveyMonkey, MOMONGA Survey and 100nin Enquete are covered. The detail of each services and tools are introduced. Their chrematistics are summarized and using for academic researches are discussed.

View full abstract