Morphological Encoding in Chinese Word Production

• Main Authors: Train-Min Chen, 陳春敏
• Other Authors: Jenn-Yeu Chen
• Format: Others
• Language: en_US
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/94717112452368421826

博士 === 國立中正大學 === 心理學所 === 95 === The dissertation investigated the issue of whether the production of a Mandarin Chinese word involves the processing of its constituent morphemes. Experiments 1-6 employed the implicit priming task. In the task, the participants first learned a set of word pairs, the members of which were associatively related. Then, they were shown the first word of each pair and had to say out loud the second word as quickly and correctly as possible. When the second words to be produced in a block shared the initial portion, production latencies were typically shorter than when they did not share the initial portion. This has been referred to as the implicit priming effect. Experiment 1 demonstrated a reliable implicit priming effect when the disyllabic compound words to be spoken shared the first tonal syllable (i.e., syllable plus tone), but the effect did not vary whether the shared tonal syllables had the same orthography. Experiment 2 showed that the implicit priming effect was observable only when the disyllabic compound words to be spoken shared the first character, but not when they shared the second character. Experiment 3 showed that characters and their corresponding tonal syllables produced equal sizes of implicit priming effect. This was replicated in Experiment 4 using disyllabic monomorphemic words. Experiment 5 observed comparable implicit priming effects whether or not the shared characters corresponded to the same morpheme. Experiment 6 showed that the implicit priming effect of the shared morpheme was insensitive to the frequency of the morpheme. The combined evidence from the implicit priming task gave no support for the involvement of morpheme in Chinese word production. To examine whether the null results were restricted to the implicit priming task, the picture-word interference task was adopted in the next set of experiments. The task involved having participants name a picture while ignoring a distractor word that was morphologically related to the name of the picture (overlapped on morpheme, character, and tonal syllable), character related to it (overlapped on character, and tonal syllable), phonologically related to it (overlapped on tonal syllable), semantically related to it (overlapped on meaning), or unrelated to it. Under the conditions when the distractor word preceded the picture by 100 ms (i.e., SOA = -100 ms), there were reliable effects for morphologically and character related trials (both being facilitatory), but not for semantically or phonologically related trials. When the distractor word was directly superposed on the picture (i.e., SOA = 0 ms), there were reliable facilitatory effects for morphologically, character, and phonologically related trials. At both SOAs, morphological facilitation was equal to character facilitation, and character facilitation was greater than phonological facilitation. When the pictures were replaced by their names and the SOA was set at -100 ms, the results showed reliable morphological, character, and phonological facilitations, all of equal size. Finally, when the distractor word and the paired picture were separated by 7-10 unrelated trials (words or pictures), the result showed reliable morphological and character facilitations, and morphological facilitation was slightly greater (by 8 ms) than character facilitation. The evidence from the picture-word interference task suggested that constituent morphemes are processed in the course of word production in Chinese, but they contribute very little to the outcome of word production. The dissertation concludes with a discussion of why such might be the case.