If children’s early words or word-like “phrasemes” have any meanings at all, then it should be possible to study and analyse their meanings. But how can early words and meanings be rigorously studied and analysed? In examining naturalistic production data from forty-seven subjects acquiring Mandarin during the first four years, this innovative study takes a radical, semantic approach to words and their meanings in child Mandarin i.e. the Natural Semantic Metalanguage (NSM) (e.g. Goddard and Wierzbicka 2002).

Amongst our findings, lexical exponents of sixty-one or so semantic “primes” posited in NSM are present in child Mandarin before the end of the fourth year. Many of these are among the earliest and the most frequent words that children produce. In addition, combinatorial properties of these lexical exponents also support hypotheses advanced about universal syntax within the NSM framework, despite challenges posed by a certain few exponents.

Early vocabulary comprises a great many semantically complex i.e. “non-prime”, words. Before an NSM prime acquires a lexical exponent, it may first be conceptually present as core semantic elements in the meanings of common non-prime words. This phenomenon is termed “latency” (following Tien 1999): a semantic prime is considered “latent” when it is first represented conceptually and expounded lexically only later in development.

On the whole, in adopting a representational system which is the NSM and which is commensurable with the adult system, this study has demonstrated that there is, in fact, developmental continuity between “the young child’s semantic system” and “the adult’s system” (Goddard 2001: 219).