- Convenors:
-
Sachiko Kiyama
(Tohoku University)
Matthew Zisk (Tohoku University)
Send message to Convenors
- Chair:
-
Sachiko Kiyama
(Tohoku University)
- Format:
- Panel
- Section:
- Language and Linguistics
Short Abstract
This panel examines Japanese writing from historical and psycholinguistic perspectives. Through studies on ateji, early romanization, Chinese–Japanese homographs, and kana recognition, it shows how a hybrid writing system emerged through language contact and is efficiently processed by readers.
Long Abstract
The Japanese writing system is among the most complex in the world, employing a combination of morphographic sinograms and three phonographic scripts (hiragana, katakana, and Roman alphabet). Moreover, each sinogram may have multiple readings depending on context. This hybrid system was established over a long historical period as an adaptation of morphographic sinograms to the structural demands of Japanese. In particular, Japanese required a written means for representing grammatical function words such as particles and verbal suffixes, reflecting its agglutinative nature, as well as an efficient way to disambiguate a lexicon rich in homophones , a consequence of its relatively simple phonological structure.
This panel sheds light on crucial aspects of how the Japanese writing system was shaped through contact with foreign languages, by presenting newly examined materials from historical linguistics and psycholinguistic experimental findings on script processing. The first two presentations focus on the history of Japanese writing. The first examines the use of Japanese ateji from a cross-linguistic perspective, comparing the methods used in ateji with similar practices found throughout the Sinographic Cosmopolis, and showing that such methods are not unique to Japan. The second looks at the earliest Roman alphabet transcriptions of the Japanese language, found in the letters of Francis Xavier in the late 16th century, and explains how these conventions helped shape the transcription practices of the Jesuits during the Christian century in Japan.
The final two presentations demonstrate how proficient readers of contemporary Japanese effectively use this hybrid writing system. The first examines the reception of shared sinographic compounds in Chinese and Japanese, through a psycholinguistic experiment with Chinese–Japanese bilinguals, focusing on how orthographic similarity mediates the activation of mental representations across the two languages. The second investigates the distinctive nature of Japanese kana through experimental studies of kana recognition by native speakers of Japanese, highlighting how kana scripts are processed as functionally specialized components within the writing system. This panel argues that the Japanese hybrid writing system is not irrational, but rather an efficient system for its users and a crystallization of Japan’s historical reception of writing and its contemporary cognitive perception.
| Abstract in Japanese (if needed) |
Accepted papers
Paper short abstract
This paper divides Japanese ateji into phonetic, semantic, and phono-semantic rebuses. It argues that such practices are not unique to Japan, but have clear parallels across the Sinographic Cosmopolis, from the liùshū principles to regional reading practices in Sinitic and non-Sinitic languages.
Paper long abstract
Japanese ateji can be broadly divided into phonetic (e.g., takusaN 沢山 ‘many’), semantic (e.g., biiru 麦酒 ‘beer’) and phono-semantic (e.g., kurabu 俱楽部 ‘club’) rebuses. While ateji are commonly viewed as a phenomenon unique to Japanese, we can find numerous parallels throughout the Sinographic Cosmopolis (the region of East Asia in which sinograms were traditionally used as a written medium, including China, Taiwan, Vietnam, the Koreas and Japan), from the principles underlying sinogram formation to regional and vernacular reading practices in both Sinitic and non-Sinitic languages.
Phonetic rebuses are the most widely attested form of rebus writing. They form an integral part of Xǔ Shèn’s theory of liùshū 六書 ‘the six principles of writing’ in Shuōwén jiězì 説文解字, in which they are given the title jiǎjiè 假借 ‘phonetic loans’. Such phonetic rebuses are abundant in Literary Sinitic texts and serve as an indispensable strategy for representing abstract morphemes in Sinitic languages and transcribing non-Sinitic languages. In Japan, this method was employed in the form of man’yōgana, and later kana, as a means for transcribing Japanese phonographically.
While less common than their phonetic counterparts, semantic rebuses are also widely attested outside of Japan. As far back as the 1920s, Nakahara Yomokurō argued that the liùshū principle of zhuǎnzhù 轉注 ‘semantic transfers’ was essentially meant to convey semantic rebuses, a claim later supported by Kōno Rokurō. While the exact meaning of zhuǎnzhù remains contested, the use of semantic rebuses has been reported both historically, in the form of tóngyì huàndú 同義換讀 ‘synonymous interchangeable readings’, and contemporarily, in the form of xùndú 訓讀 ‘regional readings’, both methods which employ sinograms to represent non-homophonous synonyms.
Phono-semantic rebuses find a parallel in the liùshū principle of xíngshēng 形聲 ‘phono-semantic compounds’, in particular, the subcategory yìshēng 亦聲 ‘[meaning] with sound’, in which the choice of phonetic is not entirely arbitrary. Early parallels are found in the sinification of place names (e.g., the Nara period kōji-rei 好字令 ‘auspicious character edict’), while modern parallels are abundant in Sinitic transcriptions of loanwords (e.g., Mandarin kěkǒu kělè 可口可樂 ‘Coca Cola’ or the use of mèng 夢 ‘dream’ in duōlā-a-mèng 哆啦A夢 ‘Doraemon’).
Paper short abstract
This paper analyzes Japanese words found in Francis Xavier’s letters to reconstruct early missionary transcriptions. Focusing on sources before 1591, it examines their spelling conventions and significance for understanding Late Middle Japanese phonology and the history of Japanese romanization.
Paper long abstract
This research focuses on the romanized Japanese lexical items found in the letters attributed to Francis Xavier, written during his time as head of the Jesuit mission in Japan in the mid-16th century. These letters constitute the earliest surviving written examples of Japanese transcribed using the Latin alphabet and represent a crucial, yet underexamined, source for Late Middle Japanese phonology and early missionary transcription practices. Although they have long been studied from historical and religious perspectives, their linguistic value has not been sufficiently explored.
Previous research on missionary linguistics focuses mainly on later printed works, such as Sanctos no Gosagyo, or early manuscripts such as Manoel Barreto’s miscellany, both dating to around 1591. By contrast, Xavier’s letters predate the establishment of the Jesuit mission press in Japan and offer insight into transcription systems in their early stage. The letters contain Japanese lexical items following mainly Portuguese and Spanish spelling conventions, reflecting how early missionaries perceived and represented Japanese sounds, prior to the adoption of Jesuit mission press in 1591.
By examining orthographic patterns, phonetic representations, and internal inconsistencies, the research identifies which transcriptional features persisted in subsequent systems and which were later modified or abandoned. This approach also clarifies the influence of European linguistic frameworks, particularly Old Portuguese and Early Modern Spanish, on early Japanese romanization practices.
The corpus is based on critical editions of Xavier’s letters, with most of these being copies written by members of the Jesuit order, dating several decades before the first printed edition of Sanctos no Gosagyo, resulting in a diverse dataset with a wide range of textual differences. These differences are treated not as simple errors but as evidence of the missionaries’ developing Japanese linguistic competence and the influence of European languages.
By focusing on Xavier’s letters, this research demonstrates their value for reconstructing early and otherwise undocumented transcription systems such as that of Juan Fernandez, contributing to the fields of Japanese historical linguistics, missionary linguistics, and broader discussions of early modern language contact.
Paper short abstract
Research on alphabetic scripts shows that visual information from graphs quickly activates sound information during reading, but other phonographic scripts remain underexplored. This study examines the time of such activation in Japanese kana and explores how this process varies across individuals.
Paper long abstract
When people read words, visual information of written graphs rapidly activates information about their sounds. Research on visual word recognition has shown that readers of phonogram-based writing systems such as English and French automatically engage both orthographic (visual-form) and phonological (sound-based) information. Yet the time course of these processes remains underexplored in writing systems that employ syllabic scripts. Japanese, an agglutinative language, adopted sinograms from Chinese; however, the need to represent affixes and other grammatical elements in writing contributed to the development of kana, a syllabic script with a highly transparent correspondence between graphs and sounds. This historical development gave rise to a distinctive mixed writing system that combines different script types (sinograms for lexical roots and kana for grammatical and phonological specification). The present study explores how contemporary readers process kana, and whether the timing of orthographic and phonological activation varies across individuals, utilizing a psycholinguistic experiment where fifty native Japanese-speaking students underwent a masked-priming lexical decision task in which briefly presented strings of graphs (primes) precede target words. The experiment manipulates Prime Type (Orthographic–Phonological prime vs. Phonological prime) and Prime Duration (33 ms vs. 50 ms) to examine the relative timing of orthographic and phonological activation during kana word recognition. In addition, we collect measures of cognitive and individual characteristics to explore how individual differences relate to reading. Because kana has shallow orthographic depth, we expect a smaller temporal lag between orthographic and phonological activation than previous studies have reported for alphabetic scripts. By situating an experimental investigation of effective kana processing within the historical development of Japanese writing, this study aims to advance a culturally grounded and more inclusive account of visual word recognition, while providing a foundation for barrier-free approaches to literacy that accommodate individual variability in multimodal information processing.
Paper short abstract
This study examines the reception of shared sinographic compounds in Chinese and Japanese using a masked translation priming experiment with Chinese–Japanese bilinguals. Results show recognition of these words varies systematically with the degree of orthographic overlap and the processing stage.
Paper long abstract
Japanese writing has developed through long-term contact with sinograms, resulting in a vast shared lexicon of sinographic compounds in Mandarin Chinese (hereafter, Chinese) and Japanese. While these items typically share semantic and orthographic properties but differ phonologically, existing research has largely focused on descriptive contrastive linguistics rather than real-time cognitive processing. From a psycholinguistic perspective, these cross-language translation equivalents are often treated as cognates, defined by functional form–meaning similarity rather than etymological origin. Following this approach, this study examines the Chinese-Japanese shared sinographic compounds composed of two sinograms, focusing on the role of orthographic overlap in bilingual lexical access. While previous studies have reported that cognates are more easily processed than noncognates, most evidence stems from alphabetic writing systems, where orthography and phonology closely interact. Chinese–Japanese two-sinogram cognates therefore provide a useful case for examining orthographic effects with reduced phonological correspondence. The study reports a masked translation priming experiment in which Chinese (first language: L1)–Japanese (second language: L2) bilinguals performed a semantic categorization task on Japanese target words. To delineate the time course of lexical activation, prime duration was manipulated at two levels (33 ms vs. 50 ms) across four conditions: orthographically identical cognates (e.g., 学生, “student”), orthographically nonidentical cognates (e.g., 价值 vs. 価値, “value”), noncognates (e.g., 行李 vs. 荷物, “luggage”), and unrelated pairs (肥皂 vs. 息子, “soap vs. son”). At the 33 ms prime duration, both identical and nonidentical cognates were processed faster than unrelated pairs, whereas non-cognates were not, indicating that early priming arose solely from orthography. At the 50 ms prime duration, identical cognates maintained a robust processing advantage compared to both nonidentical cognates and noncognates, whereas the priming effect from nonidentical cognates was comparable to that from noncognates. These results demonstrate that the cognitive reception of shared sinographic compounds is highly sensitive to the degree of orthographic overlap and processing stage. While initial lexical activation is non-selective and driven by orthographic similarity, language-specific formal constraints emerge to modulate this activation as processing unfolds, offering new insights into the dynamics of the bilingual mental lexicon.