Orthographic representation and variation within the Japanese writing system Some corpus-based observations

Given its multi-scriptal nature, the Japanese writing system can potentially yield some important insights into the complex relationships that can exist between units of language and units of writing. This paper discusses some of the difficult issues surrounding the notions of orthographic representation and variation within the Japanese writing system, as seen from the perspective of creating word lists based on the Kokuritsu Kokugo Kenkyujo's 'Balanced Corpus of Contemporary Written Japanese' (BCCWJ) Project. More specifically, the paper (i) reflects on the treatment of lemmas within UniDic, the morphological analyzer dictionary developed for the project, (ii) notes some concerns for extracting word lists that stem from the project's approach towards defining orthographic words which draws on its conceptualization of short and long unit words, and (iii) attempts to quantify the extent of orthographic variation within the Japanese writing system as represented by the BCCWJ.