The first influential linguistic research on natural sign languages of the deaf is contained in a monograph by Stokoe (1960).I would like to thank Harry van der Hulst for useful comments on this review. That work demonstrated that sign language (American Sign Language in this case) could be shown to have a level of structure that corresponds to the phonological level, in that it consists of a finite list of meaningless units that combine to form all the lexical items of the language. By substituting just one of these units with another, minimal pairs could be identified. Stokoe further categorised these units into three broad types: handshape, location and movement, a categorisation which has persisted in most subsequent work. The fact that this seminal work addressed phonology rather than any other level of linguistic structure is a significant one, precisely because the physical modality of transmission is so different from that of spoken language. Despite this fact, and despite the obvious iconicity of much of the lexicon, sign languages do in fact have an abstract submorphemic level of structure.

Since Stokoe's discovery, linguists have been trying to describe that structure, and to compare it with that of spoken languages, developing approaches which are more and more informed by theoretical work on phonology of spoken languages. Two inherent problems have challenged researchers in this field: (i) to pinpoint the similarities and differences between the phonology of spoken languages and that of sign languages; and (ii) to understand the mutual relevance of sign language phonology and general phonological theory. Brentari's book about the phonology of American Sign Language (ASL) is the most recent attempt to rise to these challenges.