The use of categorical person reference terms such as ‘woman’, ‘gentleman’, ‘lady’, etc. (sometimes referred to as ‘membership categorisation devices’) has seemed to offer a
solution to the problem of when gender is relevant in talk, since it is widely taken for granted that a speaker who refers to herself (or another) as – for example – a ‘woman’ is showing herself to be oriented to gender, thereby warranting the analyst’s treatment of her as such. Based on conversation analysis of a single recorded interaction, this
paper shows that ‘woman’ is not necessarily relevantly gendered for participants, and that – even when it is – it is not only gender, and in fact not most saliently gender,
that is always achieved through its use. It suggests that an exclusive preoccupation with the production of the category term ‘woman’ and its associated attributes as
the main focus of analysis obscures the actions in which participants are also, or otherwise, engaged.