In recent discussions there has been a tendency to refer to 'gesture' and 'sign' as if these are distinct categories, sometimes even as if they are in opposition to one another. Here I trace the historical origins of this distinction. I suggest that it is a product of the application to the analysis of sign languages of a formalist model of language derived from structural linguistics, on the one hand, and, on the other, of a cognitive-psychological view of 'gesture' that emerged in the latter half of the twentieth century. I suggest that this division between 'gesture' and 'sign' tends to exaggerate differences and obscure areas of overlap. It should be replaced by a comparative semiotics of the utterance uses of visible bodily action. This will be better able to articulate the similarities and differences between how kinesics is used, according to whether and how it is employed in relation to other communicative modalities such as speech.