- 作者: 曾進興; 黃國祐; 鄭?宜
- 作者服務機構: 國立高雄師範大學特殊教育系;國立中正大學心裡系
- 中文摘要: 雖然個別的字詞,在連續的語音信號當中,看不出是分離的獨立單位,但是聽話者卻可以清楚地辨認出一個個單獨的詞來。我們通常假定聽話者必須先將聲波切割為小的語音單位,然後在藉助這些語音單位從心面理詞典中蒐尋出對應的語詞。以印歐語言為材料進行的研究,顯示語音切割方式與個別語言的節奏性有很強的關聯。羅曼斯語系(如法語)使用者極可能先把語流切割作音節大小的單位,而日耳曼語系(如英語)使用者則可能依據輕重音的型態來決定語詞的界限。這一現象相當吻合語言心理學中一項普遍的前提,即語言的結構往往限定了人們處理語言訊息的方式。漢語(本文以國語為例)一般認為是「音節計時」的語言,因此在節律的特性上較接近羅曼斯語系。所以一個合理的推測是,漢語使用者在切音上可能表現得較像羅曼斯語者,而比較不像日耳曼語者。曾進興與曹峰銘(1994)藉著「音素偵測」的實驗作業檢驗了這個假設,實驗發現在真實音節中偵測音素,反應速度會比空缺(或不存在的)音節來得快。這種空缺劣勢效果顯示,以音節為切音策略的可能性極高。在本研究中,我們以四個新的實驗進一步探討這個假說的真確性,結果發現音節的特徵確實會影響音素偵測的速度,這表示在辨認連續的語流時,漢語使用者果然使用音節作為單位來進行語音的切割。
- 英文摘要: Studies on the Indo-European languages suggest that Romance language (e.g., French) users segmentthe speech stream into syllables whereas Germanic language (e.g., English) users make use of stress patternsfor locating word boundaries. Mandarin Chinese was hypothesized to have metrical properties that aremore similar to Romance than to Germanic languages, so that Mandarin speakers should also adopt thesyllabification strategy for speech segmentation. Tseng and Tsao (1994, Journal of National Chung ChengUniversity, 5:1, II, 181-204) tested this hypothesis with a phoneme monitoring task and found that phonemesin the initial position were identified faster when embedded in real syllables than in gaps (i.e., nonexistingsyllables). This gap inferiority effect was considered to be major evidence supporting the above hypothesis.In the four experiments of this study we present a new set of data along this line of inquiry. These ex-periments show that the properties of the syllable are the major variables determining phoneme identificationlatencies. It is our conclusion that Chinese speakers indeed make use of syllables as segmentation unitsduring speech processing.
- 中文關鍵字: syllable; speech segmentation; Mandarin Chinese
- 英文關鍵字: --