The Nature of Categorical Ambiguity and Its Implications for Language Processing: A Corpus-based Study of Mandarin Chinese
Chu‐Ren Huang, Chao-Jan Chen, Claude C.C. Shen · 2007
Ambiguity is a 'design characteristic' of the human cognitive system, even though scientific endeavors generally regard ambiguity as an evil that must be avoided at all cost in a representation system. Ambiguities are also costly to resolve in any psychological or computational model of language processing (e.g. Small et al.1988, Gorfein 1989, Schutz 1997, Ahrens 1998). However, the understanding of ambiguity may hold the key to the understanding of human cognition. In this paper, we take an empiricist approach towards a descriptive account of the nature of categorical ambiguity in Chinese. The two assumptions of this study are that i) categorical ambiguity is lexical by nature, i.e. a word is categorically ambiguous only when it is assigned multiple categories in the lexicon; and that ii) lexical representation of grammatical categories must be attested by actual use. Based on the above assumptions and given the fact that it is still impossible to enumerate contents of any individual mental lexicon (Huang 1995), we base our exploration on the information of an annotated corpus: the 5 million word segmented and tagged Sinica Corpus (Chen et al. 1996). A lexicon of over 14,000 entries is compiled based on the corpus. Any entries that are assigned