A standard and rapidly developing technology is the area of finite state technologies in lexicon building, exemplified by [Grefenstette, Schiller & Aït-Mokhtar (this volume)].
Computational implementations of lexica have often been optimised in terms of finite state automata in tree form (such as decision trees) or other graph forms, and psycholinguistic models have been developed on this basis ([Marslen-Wilson 1992], [Baayen, Schreuder & Sproat (this volume)]). The theoretical foundations of finite state technologies have been developed to a high level of sophistication in the past two decades. It is now regarded as an established fact, for example, that finite state models are quite sufficient for capturing not only phonological and morphological structures but also for modelling broad areas of lexically relevant phrasal syntax, particularly in spoken language.
An increasingly important application of finite state technologies is in `performance grammars' for efficient `shallow parsing', in which finite state approximations to context-free (or even more complex) grammars are developed. These developments are based on heuristic assumptions such as `finite state chunking' of lower level consituents with no commitment to deep tree structures, or a pseudo-context-free parse stack of finite depth.