- 期刊
WORD SUPERIORITY EFFECT IN THE VISUAL PROCESSING OF CHINESE
摘要
There are many contradictory reports in Chinese word recognition research. One of the reasons may lie in the inconsistency between Chinese characters and words as unit of perception during reading. Characters are orthographic units, which are separated from other writing units by spaces at both sides. Words, on the other hand, are linguistic units, which are listed as an entry in the lexicon with its syntactic, semantic, and phonological information. It can not be interrupted by exterior elements, and must be moved as a unit. In most languages orthographic units and linguistic units are isomorphic; such is the case in classical Chinese. However, in modern Chinese, because of phonological changes and expansion of vocabulary, these two units are no longer congruent. A word may be composed of one, two, three, or even more characters. For example, 天＂sky＂, 天花＂smallpox＂, and天花板＂ceiling''. Although the vast majority of morphemes consist of one characters, there are exceptions. For example, 葡萄＂grape＂ and 玻璃＂glass＂. The purpose of this research is to clarify what is the unit of Chinese word recognition, character or word? Because words are composed of morphemes, we like to know if the type of morpheme construction and number of morphemes will influence Chinese word recognition, and if they do, which is more influential? Most of the previous researches on this topic employed lexical decision experimental paradigm. Unfortunately, due to the confusing criterion of Chinese lexicon, those results are not consistent. The present study adopted Reicher (1969) and Wheeler (1970) word superiority effect paradigm in Chinese to investigate the saliency of characters and words in Chinese word recognition. Three experiments with different morpheme types and morpheme number conditions were carried out. The results showed that words were more salient than characters in Chinese word recogntion, and within words of the same number of characters, the morpheme number influences word recogntion more than the morpheme type. Words consisting of one morpheme were recognized more accurately than those consisting of two morphemes. In other words, mono-morphemic words were accessed faster than bi-morphemic words. Such a finding supports the hypothesis that word recognition is mediated by morpheme construction.
關鍵字
無資料並列摘要
漢語文字辨識的討論一直有許多爭議，其中一個原因是漢語字詞的不一致。簡單的說，「字」是書寫單位（orthographic unit），每個字的兩邊都會有空白把它和其它字隔開。「詞」則是語言處理的單位（linguistic unit），每個詞都有其在語意、句法、音韻等的獨立地位。詞不可以隨意在其中插入其它的成份，也不可以隨意被拆開搬移。在大多數的語言中書寫單位和語言單位是一體的，古漢語亦是如此。但是由於語言的變遷，現代漢語這兩種單位卻漸漸分家了。首先，一個詞可能由一、二、三、或更多個字組成，如「天」、「天花」、「天花板」等。同時雖然大多數詞素由一個字組成，仍有少數的例外，如「葡萄」、「玻璃」等。本文的重點在於探討漢語字詞辨識中字和詞何者重要。同時由於詞由詞素組成，一個詞中各種詞素的組合情形及詞素數目是否影響辨識則是另一個探討的重點。過去探討漢語文字辨識多用詞彙判斷作業。鑑於各個研究間的結果並不一致，本文採用另一實驗模式：詞優效應。詞優效應越大，在辨識上越顯著；辨識率越高，則詞彙觸接越快。實驗依刺激項的字數分為三組實驗，各為一字詞、二字詞及三字詞。目的是要在同一的字數的詞中控制詞素組合及詞素數目。實驗結果發現在辨識文字時，詞比字顯著，有較強的詞優效應。另外在同一字數的詞中，詞素數目對辨識的影響比詞的種類大。由一個詞素構成的詞其辨識率較由兩個詞素構成者佳。此一發現支持詞彙觸接過成中須經歷詞素建構之過成的假設。