Introduction

Chinese is well known to be a language with widespread zero anaphora (Chao 1968, Lü 1980, 1986[1942], Wang 1985[1943], Gao 1986[1948]). Following Ross (1982), C.-T. J. Huang (1984) suggests that languages be classified in analogy with McLuhan's (1964) ‘hot-cool’ division of the media. Accordingly, there are three types of language in relation to the explicitness with which they encode anaphoric relations: ‘hot’, ‘medium’ and ‘cool’. Chinese is characterised as a very ‘cool’ language in that it allows the massive use of zero anaphora and the interpretation of anaphoric expressions in Chinese relies heavily on inference, context and world knowledge. A similar typological classification can also be found in Gundel (1980), in which Chinese is considered to be a language that allows great freedom for the occurrence of zero anaphora. Furthermore, zero anaphors frequently take precedence over pronouns at both sentence and discourse levels in Chinese, which leads Li & Thompson (1979, 1981) and C. N. Li (1992: 260) to the observation that zero anaphora ‘should be considered the norm in Chinese discourse’.

Within GB theory – a parametric theory of UG – zero anaphors are treated as empty categories (ECs), entities that are assumed to be phonologically null but syntactically present.