This research aims to critically examine Japan’s language policy regarding Japanese Sign Language (JSL) and the Deaf community. By analyzing the historical evolution of shuwa, the study evaluates the persistent tension between oralist and manualist educational approaches while highlighting the lack of legal recognition for JSL as an indigenous language. Key objectives include investigating the systemic challenges in implementing bilingual education, such as teacher qualifications and instructional methods. Furthermore, the research seeks to clarify the pedagogical and policymaking confusion caused by the ambiguous use of the term "shuwa" and assesses the contemporary impact of regional dialects and cochlear implants on linguistic rights and accessibility.

Outcome：

2025
The Act for the Promotion of Measures Related to Sign Language (Shuwa Shisaku Suishin Hō) was enacted on June 18, 2025, and promulgated and came into force on June 25, 2025. The Act is significant in that it positions sign language as "a language and other important means of mutual communication" and calls on the national and local governments to promote measures such as supporting the acquisition of sign language, developing environments for its use, raising public awareness, and training personnel, thereby systematizing sign language policy at the national level. However, the Act raises two concerns. First, it remains unclear which type of sign language the term shuwa refers to. Second, the expression "a language and other important means of mutual communication" is itself interpretively ambiguous — it is uncertain whether the Act recognizes sign language as a "language" in the full legal sense or merely positions it as "one of several important means of communication." As a result, the extent to which sign language is legally guaranteed as a "language" remains unclear. Research into the significance of the Shuwa Shisaku Suishin Hō in the history of sign language policy in Japan is ongoing.

2024
Based on the survey conducted in FY2023, a paper (in press) was written that comprehensively addresses the linguistic and cultural dimensions of sign language in Japan. The paper first clarifies the linguistic features of Japanese Sign Language (JSL) and the distinctions among the three manual communication systems used in Japan — JSL, Manually Coded Japanese (MCJ), and Japanese Hybrid Signed Language (JHSL) — and then examines the confusion arising in policy, educational, and interpreting contexts from the term shuwa functioning as a hypernym in Japanese that encompasses all of these systems. It then provides a diachronic analysis of the historical tension between oralism and manualism and the transitions in language policy, demonstrating that JSL has yet to receive full legal recognition as an indigenous language in Japan. The paper further identifies the lack of JSL proficiency among teachers at schools for the deaf, problems within the teaching certification system, and structural challenges in interpreter training programs. Additionally, it considers the crisis facing regional sign languages threatened with extinction and the impact of the growing prevalence of cochlear implants on linguistic rights and accessibility. These outcomes represent the culmination of a diachronic and multifaceted investigation into the linguistic and cultural background of sign language in Japan, directly addressing the central aim of this research: a critical examination of language policy.

2023
I conducted a literature review on language policy regarding sign language and the Deaf community in Japan.

Japanese Sign Language (JSL) has units equivalent to syllables, each formed by the simultaneous combination of syllabic components such as movement, handshape, location, palm orientation, and metacarpal direction. However, even among combinations of individually well-formed syllabic components, there exist many combinations that are judged ill-formed (ill-formed syllables). This fact indicates that JSL is subject to constraints equivalent to the "phonotactic constraints" found in spoken languages. This research draws on a database of well-formed syllables and a database of ill-formed syllables — both of which decompose syllables into their component elements, encode them symbolically, and register them in Excel — in an attempt to identify and extract the factors of ill-formedness contained in ill-formed syllables. In doing so, multiple machine learning algorithms are actively employed.

Outcome：

2025
Research was conducted with the aim of elucidating the rules and constraints governing syllable formation in Japanese Sign Language and the underlying principles behind them. Analysis using the well-formed and ill-formed syllable databases yielded new findings regarding the phonotactic constraints on Type III syllables from two perspectives: handshape and location. With respect to handshape, it was demonstrated that the A handshape appearing on the non-dominant hand always co-occurs with a specific posture and carries the meaning "male/person," and therefore does not qualify as an "unmarked handshape" in Battison's sense but is instead to be analyzed as an instance of "phoneme morphologization." With respect to location, a new constraint was proposed stating that the articulation of Type III syllables is restricted to within the "A-zone (the anterior-to-the-body zone)," and it was shown that this framework can consistently account for the well-formedness of syllables including apparent exceptions. In the area of machine learning, an AUC optimization method based on Positive-Unlabeled (PU) learning was introduced to account for the continuous variation in syllabic components, and improvements in representation learning for handshapes were confirmed. These outcomes represent multifaceted advances toward a full elucidation of the phonotactic constraints of JSL and the automatic identification and extraction of ill-formedness factors through machine learning.

2024
Research was conducted with the aim of elucidating the rules and constraints governing syllable formation in Japanese Sign Language (JSL) and the underlying principles behind them, yielding the following three outcomes. (1) It was found that the phonemic handshape /O/, previously considered an unmarked handshape, cannot be specified for the non-dominant hand in Type 3 syllables (in which the two hands take different handshapes) — only the allophone [O] is permitted under the Redundancy Rule — and that it cannot occur in syllables whose nucleus consists solely of a handshape change. Questions also arose regarding the unmarkedness of /O/ from the perspectives of its low frequency of occurrence, narrow distribution, and greater structural complexity compared to other unmarked handshapes. (2) Data were obtained suggesting that the relationship between syllables and morphemes in signed languages may be structured differently from that found in spoken languages. (3) In the area of machine learning and deep learning, Adaptive Occlusion Sensitivity Analysis was introduced to verify the basis for predictions made by the syllabic component classification model and to assess the validity of its learning process. Analysis of the model's internal attention regions visually confirmed that the model trained with transfer learning classifies syllables by focusing on the dominant hand.

2023
Debugging of existing coding data was carried out, along with revisions and modifications to coding methods reflecting the latest research findings. The database was subsequently updated and refined on that basis. Work was also conducted to finalize the inventory of phonemic handshapes.