Segments and Features in Word Production by Mandarin Speakers

• Main Authors: Yu-ru Peng, 彭郁如
• Other Authors: James Myers
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/40143670759634101906

碩士 === 國立中正大學 === 語言所 === 94 === In Levelt, Roelofs, and Meyer’s (1999) theory of lexical access, phonological encoding is the stage of taking morphemes as inputs to the phonological encoder which starts the process of metrical spell-out and segmental spell-out simultaneously. Roelofs (1999) ran four implicit priming experiments on Dutch to argue for the special status of phonological segments as planning units during phonological encoding. The results showed that facilitation was obtained only when response words shared initial segments, which means that sharing full segments is necessary for advanced planning in phonological encoding. However, Damian and Bowers (2003) found that the priming effect was disrupted by incongruent spelling in English, apart from any featural mismatch, because orthographic codes are mandatorily activated during the process of phonological encoding. In Roelofs’s experiments on Dutch, the featural overlap condition not only had different initial segments, but also had different spelling. Therefore, Damian and Bowers’s results suggested that the question of whether segments are the basic units remains unsolved, since the lack of feature priming effect found by Roelofs (1999) might be due to interference from letters. Different from English and Dutch, Mandarin uses a logographic writing system. In this study, four implicit priming experiments on native Mandarin speakers were carried out to investigate the influence of speaker groups (Dutch/English speakers vs. Mandarin Chinese speakers), languages (Dutch/English vs. Chinese), and orthography (phonographic vs. logographic) on the implicit priming effect. Experiments 1 and 2 used Chinese characters to examine whether different language and orthography would yield priming effect, just like English and Dutch. Chinese characters used in Experiment 1 and 2 contained no clue of pronunciation in order to avoid regularity effect. No priming effect was obtained even in the segmental-overlap condition in which response words in a set shared initial segments. Experiment 3 used zhuyin fuhao, a system of subsyllabic phonetic symbols to examine if this would encourage Chinese speakers to use subsyllabic processing during phonological encoding. The resulting mean production latency was shorter in the segmental overlap than in the featural overlap condition. However, the result was inconsistent with Roelofs’s finding in that shorter production latencies were obtained in the heterogeneous condition rather than in the homogeneous condition. Nevertheless, the results suggested that Chinese speakers indeed showed a subsyllabic effect in phonological encoding. Experiment 4 replicated Roelofs’s (2003) bilingual implicit priming paradigm to examine whether Chinese speakers would have subsyllabic processing for English words. The result was consistent with Roelofs’s (1999) finding that priming effect was only obtained in the segmental overlap condition.