Abstract: Phonological processing concerns access to and use of mental phonological representation, which is essential to language comprehension and production. However, most findings and theories on phonological processing are based on spoken or written languages, with relatively little evidence from sign languages. As natural human languages, sign languages can also be analyzed at a phonological level. But they are articulated in a visual-gestural modality and have some unique features different from spoken languages. For example, 1) the manual and largely simultaneous articulation of phonological units or parameters (e.g., location, handshape, and movement), and 2) a tight relation between phonology and semantics due to the prevalence of iconicity, i.e., sign forms visually resemble their meanings. These features may result in some phonological processing mechanisms different from those found in spoken languages, hence posing challenges to current language processing theories. To reveal what is fundamental and what is modality-specific about language processing, this article attempts to review recent works on phonological processing in sign languages, with a focus on its empirical evidence and neurological mechanisms.
On the one hand, studies from different paradigms and techniques show that parameter-based phonological information is activated and used for sign recognition. First, signs with higher phonological neighborhood density are recognized slower than those with low density, indicating that phonological competitors are activated and compete for identification. Priming studies have found that prime-target sign pairs with one- or two-parameter overlap are responded to at a different speed or elicit different amplitude N400s from those unrelated pairs. In addition, eye-tracking studies using the visual world paradigm show that, compared to unrelated pictures, participants spend more time looking at competing pictures whose corresponding signs share one or two parameters with the target signs. These findings indicate that phonological processing in sign languages is psychologically real, and that the processing units may involve both individual parameters and two-parameter combinations. However, results are mixed regarding the actual effects of the units, with some showing facilitative effects of location overlap, some showing inhibitory effects, while others showing no effects of location but facilitative effects of handshape and location-handshape overlap.
On the other hand, brain image studies using PET, fMRI and TMS have showed that sign languages share similar neurological mechanisms of phonological processing with spoken languages. In sign perception and comprehension tasks, signers tend to activate the superior temporal cortex bilaterally when viewing signs or sign-phonetic and syllabic units. In explicit phonological judgement tasks, a left-lateralised network is engaged, including the left inferior frontal cortex, superior marginal gyrus, and the superior parietal lobule. These processing regions have also been found in speech sound perception and word rhyme judgement, indicating that there are some similar phonological processing mechanisms in the two language modalities. Differences also exist, though, at the whole brain level, significant activation is found in the left occipital lobe in handshape judgement but not in the frontal-parietal network and the cerebellum bilaterally in speech rhyme judgement. Yet it is unclear whether these differences are related to modality-specific processing mechanisms or the sensory properties of phonological parameters.
To sum up, research available has provided some evidence for the psychological reality and neural mechanisms of phonological representation and processing in sign languages, but there is still a lack of consistency about the processing units and their roles as well as evidence on the modality-specific processing mechanisms. Thus some suggestions are given for future studies: (1)Future investigation may look at the roles of phonological parameters; (2) Further exploration of the neural mechanisms of sign phonology processing, especially on the brain activities in specific parameter perception and comparison of mechanisms may invoke in different tasks; (3)Enrichment of studies on various kinds of sign languages, as different languages may have their own unique phonological, grammatical or syntactical properties, which may result in cross-linguistic differences in processing mechanisms.

Key words: sign languages, phonological processing, phonological parameters, neural mechanisms

摘要： 语音加工在言语理解与产出中具有重要的作用。手语是基于视觉-手势模态的自然语言,也具有语音结构,但与口语的语音结构之间有很大的差异。本文总结和分析了近年来有关手语语音加工的研究,发现手语加工中存在基于音韵参数的语音信息激活和使用,且手语的语音加工神经机制与口语具有相似之处,尤其在双侧颞叶和左侧额下回等部分脑区存在重合,同时在左侧枕叶和缘上回等脑区具有一定的特异性。最后,文章从语音加工单元和手语语种等方面展望了未来的研究方向。

关键词: 手语, 语音加工, 音韵参数, 神经机制