Verb Bias in Mandarin Relative Clause Processing

Abstract

Recent research has indicated that a verb`s preferential subcategorization frame plays a fundamental role in guiding the choice of the type of complement that follows a verb. For example, sentences with a sentential complement-biased verb, such as "admitted", will cause less processing difficulties when it is followed by a sentential complement than when it is followed by a direct object. Such information that verbs carry, a.k.a. verb bias effect, has been shown to influence the processing of languages in head-initial languages such as English. Up to date, no studies have been found to investigate verb bias effect in head-final Mandarin relative clauses. The present paper thus aims to investigate the influence of verb bias during online Mandarin relative clause processing. In addition, the present study also aims to further examine two language processing models, constraint-based model and garden-path model, to see which may be more correct in predicting the processing results. Findings of the present study show that, similar to English speakers, Mandarin speakers are also capable of using the information embedded in the verb to disambiguate and predict sentence structures, thus supporting the predictions of the constraint-based model.

近來研究指出動詞內的結構意涵對語句處理有重要影響。舉例來說，英文中的「admitted」後面比較常接子句，而非直接受詞。在語句處理時，若讀者發現「admitted」後面果真接著子句時，讀者所產生的處理困難會因此減少。目前為止，這些研究針對的大部分都是英文。然而此一現象是否可以推論至中文卻不得而知，因為就關係子句來說，英文和中文的結構相距甚遠！在英文裡，關係子句之「頭」出現在關係子句之首。而在中文裡，它出現在關係子句之尾。此特殊的結構剛好提供學者得以操弄並進一步確認語言學上不同派別的理論與假設。本文旨在研究在中文關係子句之「頭」在子句尾出現的狀況下，以中文為母語的使用人士是否仍得以借用到動詞本身的結構意涵來預測並處理語句。研究發現儘管中文的結構和英文截然不同，中文使用者仍然可以快速的接收動詞內的結構意涵，並用此處理語句。此外，本文結果也支持語言使用者在閱讀中文句子的時候會同時接收不同訊息來處理句子。