Terry Joyce: Japanese writing system

Japanese writing system (JWS)

[Updated: 20200422]

Within present revision of website, this page is still under construction.
The present page consists of two topics; namely, morphographic writing system which outlines my arguements that this is the more precise terms to refer to how kanji function as a form of graphematic representation, and orthographic representation and variation, which outlines evidence from Joyce, Hodošček & Nishina (2012) for the fungible nature of the JWS.
When revision complete, this page will also briefly introduce the notion of intentionality, espoused by Joyce & Masuda (2019) as framework for the motivational factors that underlie both the conventional orthographic and non-conventional, or variant, graphematic representations of the JWS.

Amongst the writing systems of the world, the Japanese writing system (JWS) undeniably warrants special attention!

This claim may appear quite bold and provocative, but it is definitely not without justification. Indeed, reflecting on both the frequency and sheer variety of superlatives that have been used in describing the JWS, its validity is pretty difficult to refute.

As recorded in more detail in Joyce (2002, 2011), writing system scholars have described the JWS as being the most complex (Sproat, 2000; Kess, 2005; Gnanadesikan, 2009), the most complicated (Coulmas, 1989; Fischer, 2001; Robinson, 1995), the most elegant (Kess, 2005), the most intricate (Coulmas, 1989; Kess, 2005), the most onerous (Unger, 2004), as being without inferiors (Sansom, 1928), or simply describing it as “one of the worst overall systems of writing ever created” (DeFrancis, 1989, p. 138).

While this page certainly aspires to some day becoming pretty comprehensive about all aspects of the JWS, given the considerable degree of attention that that will eventually entail, the present incarnation is admittedly rather more selective in mainly singling out what I, personally, regard as being one of the central aspects or issues for understanding the JWS (at least, it has shaped and continues to drive much of my own research); namely, to appreciate more fully how kanji function primarily as a morphographic writing system (Joyce, 2002, 2008, 2011).

However, in the spirit of striving towards its grander aspirations, this page also introduces two other, more recent, research projects that address different but complementary aspects of the JWS. The first is concerned with the notions of orthographic representation and variation within the JWS (Joyce, Hodošček & Nishina, 2012). [Please note that the BCCWJ-corpus-based word lists created and reported on in that paper are being made publicly available within a separate Corpus word lists page, from where the lists can be downloaded.] The second, ongoing, research project focuses more on the official jōyō kanji list (Joyce, Masuda, & Ogawa, accepted).

Links to main sections of this page, to facilitate navigation.

Morphographic principle

Orthographic representation and variation

References

Morphographic principle

As already indicated, essentially, much of my research has been pursuing a psycholinguistic approach to investigate the simple question of how do kanji function as a writing system? And, stemming from firm belief that a large part of understanding how a writing system functions depends on identifying the principles that govern the combination of graphic units, much of my research has focused on the two-kanji compound word as its principle frame of reference.

Given that the nature of a writing system is reflected in the structure of the literate mental lexicon, I believe that the underlying transition from writing system to mental lexicon to be simply a shift in perspective -- from external form to internal representation. But, in the interests of helping to disseminate the message as quickly as possible, because the whole issue seems already too obscured by misconceptions and 'myths', allow me to state without further delay the answer I believe to be supported by my research; namely, that kanji within the Japanese writing system can be most appropriately characterized as being a morphographic writing system and that this is reflected in the organization of the Japanese mental lexicon.

It is sincerely hoped that visitors will carefully assess this claim about the morphographic nature of Japanese kanji within a complete picture that connections with the dominant principles of Japanese word formation and with priming results obtained from constituent-morpheme priming experiments conducted in proposing the Japanese lemma-unit model of the mental lexicon, as sketched out with a few more words within the Japanese lemma-unit model page (JLUM).

At the head of this page, it was noted that the JWS has something of a negative reputation for its complexity. This is undoubtedly due, in part, to its multi-script nature, using a mixture of kanji, two sets of kana, and increasingly the alphabet, as elements of an integrated system, as well as the dual-reading system of on-readings (Sino-Japanese pronunciations) and kun-readings (native Japanese pronunciations) for kanji. However, I have argued that the dominant image of the JWS is primarily a reflection of the controversy that has surrounded the appropriate linguistic classification of kanji as a writing system.

Within my Ph.D. thesis (Joyce, 2002) and, as summarized in Joyce (2011), I have reviewed a number of important linguistic classifications of writing systems, such as that presented in Gelb's (1952) seminal book on grammatological research, Sampson's (1985) highly controversial classification that prompted passionate replies from DeFrancis (1989), as well as more recent classifications by Sproat (2000) and Rogers (2005) that, breaking with the common depiction as tree-figures, classify systems according to the two dimensions of type of phonography and degree of logography/ morphography (respectively) involved within a system. And, based on such reviews, I have discussed the theoretical perspectives concerning the relationship between language and writing which are at the heart of the often heated debate over terms, such as word-syllabic (Gelb, 1952), ideographic (Diringer, 1962), morpho-syllabic (DeFrancis, 1989), logosyllabic (Daniels, 1990, 1996, 2001), logographic (Faber, 1992), morhosyllabic (Coulmas, 1992), as well as core syllabic + high level of logography (Sproat, 2000) and moraic + high level of morphography (Rogers, 2005), which have, at various times, been applied to kanji.

Without recounting all the background again here, as already noted, I believe that kanji within the Japanese writing system may be most accurately characterized as being a morphographic writing system. Now, given the wider definition that many currently accord to the commonly-encountered label logographic, my preference for the term morphographic might appear rather pedantic to some scholars. However, I believe that the typological labels we apply should be as informative as possible about how different writing systems function in representing language. This means that we must seek to identify the dominant principle underlying the use of the orthographic units and classify them accordingly. Seen in this light, continuing the practice of using the label logographic, while, at the same time, acknowledging that morphographic is “more precise”, as some scholars do, surely does very little to deepen our understanding of writing systems. The shift in focus that the revision in terminology requires is of fundamental significance for our understanding of kanji in the Japanese writing system, in particular, and of writing systems, in general, as well as for psycholinguistic research into the organization of the literate mental lexicon.

Orthographic representation and variation

Having advanced this appeal for wider acknowledgement of the morphographic nature of kanji, in a spirit of expansion, this page now turns to share a few corpus-based observations concerning orthographic representation and variation within the JWS, obtained within Joyce, Hodošček and Nishina (2012). Striving to explore the complex relationships that can exist between units of language and units of writing, Joyce et al (2012) specifically addressed some difficult issues surrounding the notions of orthographic representation and variation within the JWS from the context of creating word lists from the recently completed Balanced Corpus of Contemporary Written Japanese (BCCWJ) Project (Maekawa, 2007; NINJAL, 2011).

In addition to providing download links (for two data archives), the Corpus word lists page includes a table of file names for the created corpus word lists. In total, 31 files were created according to the two kinds of word unit utilized within the BCCWJ and 14 different word class categories. The two word units are short-unit words (SUWs), which essentially equate to morpheme, and long-unit words (LUWs), which are closer to phrases.

The simplest piece of evidence that the relationship between lemma types and orthographic representations is far from straightforward in the case of the JWS can be gained from a summary of the generated word lists, presented as Table 3, p. 262, in Joyce, et al (2012) and recreated within the Corpus word lists page. The totals row for SUWs indicates that amongst the 104,701,996 SUW tokens, there are 175,708 lemma types which are associated with 277,792 orthographic types. Accordingly, the mean lemma-orthographic type ratio (LOTR) for SUWs of 0.63 indicates that, on average, SUW lemmas have approximately two orthographic representations.

Instead of attempting to thoroughly investigate the notions of orthographic representation and variation, which touch on a range of sociolinguistic issues, such as language policy, Joyce et al (2012) primarily sought to identify some issues as deserving of future examination and to provide some demonstration of how the BCCWJ could contribute greatly to such research. The first issue was concerned with the treatment of lemmas within UniDic, where the specifications for an efficient morphological parsing dictionary may entail sacrificing some traditional distinctions of nuance between kanji. Most directly related to the extraction of the corpus word lists, the second issue was basically about defining units of language for a highly agglutinative language like Japanese. While the utilization of both SUWs and LUWs arguably yields the greatest flexibility overall, it is also fair to comment that neither unit corresponds neatly with a general notion of the word. Accordingly, both linguistic units must be used in order to extract the widest range of word types from the BCCWJ.

The final concern or objective of the Joyce et al (2012) paper was to attempt to quantify the extent of orthographic variation within the JWS as represented by the corpus word lists extracted from the BCCWJ. To that aim, the study calculated LOTRs and plotted their distribution. The resultant distributions indicate considerable levels of orthographic variation with the JWS, with up to 34 orthographic variants for one SUW noun lemma and mean variations of 8.44 for SUWs and 5.80 for LUWs (based on the most frequent 100 lemmas). With highly-similar distributions of orthographic variation across the four main word classes of nouns, verbs, i-adjectives and adverbs, the quantitative data for the corpus word lists clearly demonstrate that orthographic variation is a major characteristic of the JWS, at least, in terms of the orthographic representation of the most common Japanese words.

References

Coulmas, Florian. (1989). The writing systems of the world. Oxford: Basil Blackwell.

Coulmas, Florian. (1992). Writing systems. In William Bright, (Ed.). International encyclopedia of linguistics, vol. 4. (pp. 253-257). New York, Oxford: Oxford University Press.

Daniels, Peter T. (1990). Fundamentals of grammatology. Journal of the American Oriental Society, 110, 727-731.

Daniels, Peter T. (1996). The study of writing systems. In Peter T. Daniels & William Bright, (Eds.). The world's writing systems. (pp. 3-17). New York, NY: Oxford University Press.

Daniels, Peter T. (2001). Writing systems. In Mark Aronoff, & Jamie Rees-Miller, (Eds.). The handbook of linguistics. (pp. 43-80). Oxford, England: Blackwell.

Diringer, David. (1962). Writing. London: Thames and Hudson.

DeFrancis, John. (1989). Visible speech: The diverse oneness of writing systems. Honolulu, HI: University of Hawai'i Press.

Faber, Alice. (1992). Phonemic segmentation as epiphenomenon: Evidence from the history of alphabetic writing. In Pamela Downing, Susan D. Lima, & Michael Noonan, (Eds.). The linguistics of literacy. (pp. 111-1134). Amsterdam: Benjamins.

Fischer, Steven Roger. (2001). A history of writing. London: Reaktion Books.

Gelb, Ignace J. (1952). A study of writing: The foundations of grammatology. Chicago: University of Chicago Press. [Revised edition, 1963, Chicago & London].

Gnanadesikan, Amalia E. (2009). The writing revolution: Cuneiform to the Internet. Oxford: Wiley-Blackwell.

Joyce, Terry. (2002). The Japanese mental Lexicon: The lexical retrieval and representation of two-kanji compound words from a morphological perspective. Unpublished doctoral thesis. University of Tsukuba, Japan.

Joyce, Terry. (2008). The significance of the morphographic principle for the classification of writing systems. Typology of Writing Systems: Sixth International Workshop on Writing Systems, 18-19 September, 2008. Braunschweig, Germany.

Joyce, Terry (2011). The significance of the morphographic principle for the classification of writing-systems. (Typology of writing systems special issue edited by Susanne R. Borgwaldt and Terry Joyce). Written Language and Literacy, 14, 1, 58-81. DOI:10.1075/wll.14.1.04joy

Joyce, Terry, Hodošček, Bor, & Nishina, Kikuko. (2012). Orthographic representation and variation within the Japanese writing system: Some corpus-based observations. (Special issue: Units of language - units of writing, edited by Terry Joyce and David Roberts), Written Language and Literacy, 15(2), 254-278. DOI: 10.1075/wll.15.2.07joy

Joyce, Terry, Masuda, Hisashi, & Ogawa, Taeko. (accepted). Jōyō kanji: Recent revision, characteristics, and role as core component of the Japanese writing system. Presentation to be given at ‘The Architecture of Writing Systems’ 8th International Workshop on Writing Systems and Literacy, Institut für Germanistik, Oldenburg, Germany, 4 - 5 October 2012.

Kess, Joseph F. (2005). On the history, use, and structure of Japanese kanji. In Katsuo Tamaoka (Ed.), Corpus studies on Japanese kanji, (Glottometrics 10) (pp. 1-15). Hituzi Syobo: Tokyo, Japan and RAM-Verlag: Lüdenschied, Germany.

Maekawa, Kikuo. (2007). Design of a balanced corpus of contemporary written Japanese, Proceedings of the Symposium on Large-Scale Knowledge Resources (LKR2007) (1-3 March, Tokyo Institute of Technology, Tokyo, Japan), 55-58.

National Institute for Japanese Language and Literature (NINJAL) [Kokuritsu Kokugo Kenkyūjo]. (2011a). Tokuteiryōiki kenkyū nihongo kōpasu kenkyū seika hōkoku [Priority-Area Research “Japanese Corpus”: Research Report] [DVD format of data and research reports]. Tokyo: General Headquarters, Priority-Area Research “Japanese Corpus”.

Robinson, Andrew. (1995). The story of writing: Alphabets, hieroglyphs & pictograms. London: Thames & Hudson.

Rogers, Henry. (2005). Writing systems: A linguistic approach, (Blackwell textbooks in linguistics, 18). Malden, MA & Oxford, UK: Blackwell Publishing.

Sampson, Geoffrey. (1985). Writing systems: A linguistic introduction. London: Hutchinson & Stanford: Stanford University Press.

Sansom, George B. (1928). An historical grammar of Japanese. Oxford: Clarendon Press.

Smith, Janet S. (Shibamoto). (1996). Japanese writing. In Peter T. Daniels, & William Bright, (Eds.). The world's writing systems. (pp. 209-217). New York & Oxford: Oxford University Press.

Sproat, Richard. (2000). A computational theory of writing systems. (Studies in natural language processing). New York & Oxford: Oxford University Press.

Unger, J. Marshall. (2004). Ideogram: Chinese characters and the mth of disembodied meaning. Honolulu, H: University of Hawai'i Press.

Yamada, Toshio. (1967). Japan - The writing system: Historical research and modern development. In Thomas A. Sebeok, (Ed.). Current trends in linguistics, Vol. 2Current trends in linguistics, Vol. 2. (pp. 693-734). The Hague, Mouton.