The Tohoku University 21st Century COE Program in Humanities:
Open Lecture Series No.2

Psycholinguistic Studies of Kanji and Kana Processing

Tadao Miyamoto, Ph.D.
Department of Linguistics
University of Victoria
British Columbia, Canada
V8W 3P4

The Mental Lexicon is a well studied topic, comprising three main subtopics: how individual lexical items are organized in our brains; how they are accessed; and how they are acquired. The field of Japanese psycholinguistics has a long history of academic excellence. Japanese psycholinguists are especially noted for their studies on the processing of written words, with particular focus on how written words are accessed. This interest stems from the uniqueness of the Japanese writing system, which includes both logographic (kanji) and phonographic (kana) symbols. Investigations of kanji/kana processing have primarily aimed to determine whether the processing of written Japanese is unique or whether it is similar to lexical processing in other languages. This talk addresses several issues relating to kanji/kana processing, drawing on reviews of previous studies on eye-movements, morpheme-stripping, kinetic memory, laterality and aphasia.

Eye Movement Studies:
Eye movement studies are grounded in the assumption that the speed of eye movements correlates with the speed of comprehension of written materials. Eye movement studies are frequently used to infer the nature of the processing of written words. The eye movements consist of fixations (of 200-250 msec each) and saccades (which occur approximately every six characters). Fixations on characters involve foveal vision, while saccades, which are used to help locate the next appropriate fixation point, employ parafoveal vision; our eyes move from one fixation point to the next. These eye movement mechanisms are employed in the lexical processing of written material in all languages. It is a universal fact that while content words serve as sites for fixations, function words are consistently skipped over. In the processing of English texts, 15-20% of eye movements are regressive, with the eye returning to the beginning of sentences following an initial failure to decode their meaning. In contrast, regressive eye movements are found to constitute less than 10% of all eye movements in the processing of kanji/kana mixed texts. These findings imply that there may be non-universal aspects involved in Japanese lexical processing, suggesting that Japanese texts are ideal for reading due to the visually rich mixture of logographic and phonographic symbols.

Morpheme Stripping Studies:
One trend in Japanese psycholinguistics studies of the 1970s was the examination of the "dual route" hypothesis. According to this hypothesis, the processing of Japanese texts was unique in employing two separate routes. The first route was dedicated to the processing of kana words, involving an initial decoding of the graphemic features of the kana, followed by the processing of phonological and semantic properties. The second route was employed in the processing of kanji, in which semantic properties could be directly inferred from graphemic features due to the logographic nature of kanji characters. Studies on the dual route hypothesis were inconclusive on the question of the universality of processing mechanisms. Over time, this experimental paradigm has been replaced by studies on "morpheme stripping," a significant research topic in the West, where numerous studies had demonstrated that written words are internally analyzed in the course of lexical processing. For instance, in the processing of both prefixed words (e.g., revive) and pseudo-prefixed words (e.g., relish), affixes are stripped from roots; individual words are never processed as unified "chunks." This experimental paradigm was applied to the study of kanji processing. Kanji are of two types, simple or complex, with most of the latter consisting of two radicals. With respect to the morpheme stripping paradigm, the issue was whether in the processing of such complex kanji characters, the reader actually strips individual radicals or whether s/he processes whole complex kanji characters as chunks. This paradigm was further extended to compound kanji words, in an effort to determine if individual kanji characters in compound words are stripped or not. A general conclusion of these studies is that morpheme stripping takes place with Japanese kanji words, lending support to the notion of universal mechanisms in orthographic processing.

Kinetic Memory Studies:
Across languages, written words are retrieved through a decoding of their accompanying graphemic, phonological and semantic representations. However, studies on kanji indicate that these words can also be retrieved by releasing the kinetic memory associated with individual kanji characters. This evidence stems from studies on the traditional practice of kuusho, the finger-writing strategy for recalling a kanji character. When Japanese writers are uncertain of a kanji character, they may begin tracing its outline on the palm of their hand to initiate the motor sequence that unlocks the remainder of the kanji configuration. Previous pedagogical and experimental studies indicate that kinetic memory is an integral part of the lexical properties of individual kanji characters. Unlike non-logographic languages, it seems that for Japanese, the graphic representations stored in the brain for the motor output of kanji incorporate kinetic information and form part of the mental representation of a given kanji character. In this sense, the mechanisms used in the processing of kanji may not be wholly universal.

Laterality Studies:
Hemispheric laterality is an extensively investigated issue in both the West and Japan. Results from early studies in clinical and experimental psychology gave rise to the impression that orthographic processing in Japanese is unique. These studies found that phonemically encoded kana are processed exclusively in the left linguistic hemisphere, while visuospatially oriented kanji are processed in the right hemisphere. This dichotomous view of kanji/kana processing is also utilized in Nihonzin-ron to support the claim that the Japanese brain --- and, for extreme adherents, the Japanese race --- is unique. A large number of studies have since explored the dimensions of kanji/kana processing from different perspectives. Findings from these studies call for a rearticulation of the "dichotomous view": hemispheric preference may not be dependent on orthography, but rather on the type of task involved. In this account, graphemic tasks call for right hemisphere superiority, while both phonological and semantic tasks call for left hemisphere superiority, lending support to a universalist account of the processing of written words.

Aphasiology:
The processing of kanji/kana words is a dominant theme not only among Japanese experimental psychologists, but also within the medical field of aphasiology; the majority of medical papers written by Japanese aphasiologists published in the West concern this issue. Of the variety of non-localized aphasic syndromes, two are relevant to the processing of kanji/kana: alexia, or disturbance in the comprehension of written materials; and agraphia, or the loss or impairment of the ability to produce written language. Impairments in the processing of alphabetic languages may be associated with the combination of [+/- Alexia] and [+/- Agraphia], yielding four possible syndrome types. In the case of Japanese, there are eight possible syndrome types: the four types delineated above, multiplied by the two script types of kanji/kana. The challenge for aphasiologists is to map these eight aphasic syndromes to very precise locations in the brain, based on the locations of lesions associated with each syndrome type. Despite the advance of diagnostic tools, this task has proven frustrating to neurologists. No clear claims on the universality of written word processing have emerged from this line of inquiry.

Conclusion:
From the viewpoint of universality, kanji/kana processing is an extremely fertile area of inquiry in the field of Japanese psycholinguistics, and will remain so for the foreseeable future in both Japan and the West.