本研究目的為探究使用臺灣手語聾人在閱讀字詞時的手語轉錄情形。研究對象為46位以臺灣手語為主要溝通工具之成年聾人，以及46位無手語能力的聽力正常成年人，共計92位。研究工具為依據手語音韻相似／不相似及語意相關／不相關四向度所自編之「語意判斷作業」，包括：「手語音韻相似／語意相關」(SR)，「手語音韻相似／語意不相關」(SU)，「手語音韻不相似／語意相關」(UR)及「手語音韻不相似／語意不相關」(UU) 120對字詞組，240個詞彙。實驗測試方式為請研究參與者判斷同時呈現之字詞組是否為語意相關。整體而言，本研究發現：(1)聽常者在四種語意判斷題型的反應時間未出現顯著差異。(2)聾人在四種題型的反應時間出現顯著差異，由快至慢依序為UR > SR > UU > SU。(3)無論是手語音韻相似或不相似題型，語意相關反應時間皆快於語意不相關。(4)無論是語意相關或語意不相關題型，手語音韻相似皆慢於手語音韻不相似。根據研究結果，本研究從手語干擾抑制效應，指出聾人閱讀時會將識讀之字詞轉譯（轉碼）至手語，啟動手語音韻解碼，出現手語轉錄歷程。此外，本研究亦發現聾人閱讀者會使用語意直接觸接來達成文字識讀，顯示其閱讀中文詞彙時除有手語轉錄現象，語意亦有一定程度之影響作用。

Background and Purpose

In the past few decades, research has consistently shown that deaf children do not achieve the same reading and writing skills as their hearing peers. Research has sought to explore the factors related to deaf children’s difficulty in literacy development. It is well-established and widely accepted that reading is the product of decoding skill and language comprehension (Gough & Tunmer, 1986). Accordingly, reading/written word recognition is viewed as the processes of establishing the relation between spoken language and its representation in print. Specifically, it is assumed that processing of written word recognition involves associating letters with sounds, blending the sounds together, and determining what word is represented by a string of letters. However, for deaf individuals using sign language, literacy acquisition presents a unique challenge due to their lack of direct access to spoken letters. It is suggested that although sign language provides a rich linguistic input, the transition to reading written text can be challenging. Previous research has used priming paradigm and developed pairs of written words that were not related in the written or phonological domain in spoken language, but shared sign units in a respective sign language (e.g. ASL, DGS) to understand the challenges and nature of the sign language recoding or representations that deaf signers use during reading. It was reported that when deaf signing participants had to detect semantic relatedness for pairs of written words, implicit phonological priming of the underlying signs speeds response time and, vice versa. Although these findings have shed light on the relationship between sign language representation and written word recognition, most of the evidence found in the research came from ASL or other European sign language users. Mandarin Chinese written language is a very different language system as compared to English in terms of its phonological system, word structure, which makes it a good candidate to explore this issue from a cross-linguistic perspective. However, little research has been done to study the underlying mechanism of written word recognition for deaf signers in Taiwan. Therefore, it is necessary to understand how Taiwanese deaf signers activate signs and their respective phonological forms when they are reading written words. The purpose of the current study is to understand whether Taiwanese adult deaf signers are able to make rapid use of ongoing sign language to facilitate recognition of written words and demonstrate recoding to sign language.

Method

Participants: Forty-six deaf adults (mean 46.28 years) for whom Taiwanese Sign Language (TSL) is considered their first language and Mandarin Chinese is the second language, and 46 hearing adults (mean 34.17 years) without any experience learning TSL participated in this study. All the deaf participants’ hearing threshold is greater than 90 dB and hold at least a high school diploma. Materials: The current study adapted the semantic relatedness judgement paradigm from Morford et al. (2011) to test whether adult deaf signers activate the Taiwanese Sign Language translations of Mandarin Chinese written words. One hundred and twenty Mandarin Chinese word pairs were developed. Sixty pairs were semantically related (e. g. tree-forest), and another 60 pairs were semantically unrelated (e. g. flower-fart). Among these same 120 word pairs, 60 pairs had phonological similarity in TSL (e.g., white-doubt), 60 pairs had phonological dissimilarity in TSL (e.g., butterfly-mirror). These test stimuli were further divided into four response conditions, namely: (1) phonologically similar in TSL/semantically related (SR), (2) phonologically similar in TSL/semantically unrelated (SU), (3) phonologically dissimilar in TSL/semantically related (UR), and (4) phonologically dissimilar in TSL/semantically unrelated (UU). Phonological similarity in TSL was defined as sharing a minimum of two formational parameters (i.e. handshape, movement, location, palm orientation, and facial expressions). Semantic similarity ratings and determination of phonological similarity in TSL were collected from 60 hearing Mandarin Chinese speaking adults and 6 deaf TSL teachers respectively. A semantic relatedness judgment task was adopted, in which participants were asked to judge word pairs (2 words simultaneously presented) shown on the computer screen to be “related in meaning” or “not related in meaning”. Response time (RT) was measured to the nearest millisecond from the onset of the second word pairs.

Results

The current study used a 2 (Semantics) × 2 (Phonology) repeated measures ANOVA across subjects (F1) and items (F2) and revealed effects of TSL similarity on semantic relatedness RTs. Overall, it was found that: (1) Word pairs sharing phonological similarity in TSL were judged more slowly than word pairs that were phonologically dissimilar in TSL. (2) Word pairs that were semantically related were judged more quickly than word pairs that were semantically unrelated. (3) Word pairs that were phonologically similar in TSL/semantically related were judged more slowly than word pairs that were phonologically dissimilar in TSL/semantically related. (4) A control group of hearing adults without any knowledge of ASL produced an entirely different pattern of results. Based on the results, it was concluded that phonological similarity in TSL tended to hinder deaf participants’ determination of the semantic relatedness of word pairs and caused slower reaction time.

Discussion

One of the most commonly asked questions in deaf education is whether and how signed language knowledge will affect the learning of reading. In line with previous studies, this study also adopted semantic relatedness judgement task to explore this question. However, in order to simulate real text reading context, the priming technique used in this study (i.e. presenting word pairs of two words simultaneously) is unlike previous studies where two words were presented sequentially. Nevertheless, the results found in the present study still provide evidence that Taiwanese adult deaf readers activate the TSL translations of written words or use sign language to decode the written words and support the results found in previous research. Specifically, the current study did not show sign language facilitating effect but an inhibition effect. That is, deaf participants were slower to reject semantically unrelated words with phonological similarity in TSL. In addition, semantic access to achieve word meaning comprehension also plays a crucial role during written word recognition. Based on the findings, the implication for the teaching of reading for deaf students is addressed hereby. It is suggested that education for deaf students should help them to learn the integration of visual and linguistic cues, given that sign language provides a rich linguistic input. Moreover, it is necessary that teachers should provide opportunities for deaf students to engage in discussions, storytelling, and other language-rich activities to support and increase their semantic or language skills. As for future research, it is suggested that using event related potentials (ERPs) or fMRI study to explore priming elicited effects of sign language activation during written word processing might shed light on this issue.