The notion of literalness in linguistics is based on the following assumptions: Linguistic expressions are vehicle-meaning p airs (since literal meaning has to be the meaning of something). Linguistic expressions have to be cognised a nd used (especially uttered) in order for their meanings to be regarded as literally theirs. "Linguistie, vehicle-meaning relations are fixed and autonomous- - rather than having particular meanings in virtue of being used to express those meanings," linguistie' vehicle-meaning p airs are used to express certain meanings in virtue of having the meanings th at they have. This thesis criticises Chomsky's and Sperber and Wilson's attempts to establish the autonomy of "linguistie'vehicle-meaning pairs. I argue that " Both Chomsky and Sperber and Wilson fail to distinguish "linguistie' semantics from the "real" semantics of what "linguistic" vehicle-meaning pairs are used to express. " They persist in the idea that "linguistic" vehicles are specifically for being uttered (physically instantiated), thus defeating their own purpose of setting the linguistic absolutely apart from what it is used for. " Neither Chomsky's internalist conception of language nor Sperber and Wilson's relevance framework is able to account for the phenomenon of "language misuse", i.e. the use of a "linguistic" vehicle to express the "wrong" meaning. Burton-Roberts' representational conjecture is applied and developed in the presentation of an alternative non/ extra-linguistic account of "literal meaning" and "language use/ misuse". This account has it that neither "linguistic" vehicles nor "linguistic" vehicle-meaning relations are actually linguistic. It avoids the problems attending the notion of linguistic expressions as objects with sortally disjoint and arbitrarily conjoint properties (i.e. physically instantiable "vehicle" and mentally constituted "meaning"), and resolves the unease within Chomsky's Minimalist Program about the inclusion of phonology in I-language. Finally, by way of this resolution, I address some seemingly unrelated issues concerning vehicle-less "meanings" and the relations between l anguage,thought and consciousness.

Recent progress in psycholinguistic research on written Chinese allows us to develop a new approach to investigate the Chinese reading acquisition process. We hypothesized that Chinese children, much like children learning an alphabetic script, do not simply learn written words by rote. As they are taught words to be learned by rote, they develop an implicit understanding of the formal and functional characteristics of written Chinese. The formal characteristics refer to the graphic structure and the positioning of the stroke-patterns, and the functional characteristics refer to the semantic and phonological information conveyed in the stroke-patterns. The studies reported were designed to investigate the nature of children's learning of written Chinese. In two series of studies, a total of 236 children from Hong Kong, aged four to nine, created and decoded novel Chinese compound words. Results showed that young Chinese children attended to both the formal and functional constraints in reading and writing tasks. In the judging task, 4-year-olds were able to identify the type of orthographic elements - the stroke-patterns, but they could not place them in legitimate positions. The 6-years-olds were able to refer both to the position and the correct type of orthographic elements in differentiating pseudowords from nonwords. In the writing and reading tasks, four and five-year-olds were unable to utilize the semantic radicals to represent meaning, nor could they use the phonological components for pronunciation; six-year-olds could use the semantic radicals to represent meaning and only nine-year-olds could both use semantic radicals correctly and systematically referred to the phonological components for pronunciation. A significant age difference was found in all the experiments. The studies provide strong evidence that learning compound words in Chinese is not a simple matter of memorizing but involves the understanding of formal and functional constraints in the script. A possible application of these findings lies in the new direction offered for reading instruction where the non-generative, rote view of learning to read and write in Chinese can be safely abandoned.