It is well known that language is used in different ways in different interpersonal and physical contexts. We do not usually speak to the bank m anager in the m anner and tone that we do to our chil­ dren. Dialogue in a courtroom is different from that round a din­ ner table. Aside from these specific examples, there are general claims that can be made about the form and content of com­ m unication as a function of context. I wish to examine one of these claims that refers to the sociolinguistic validity of work in the referential communication tradition. The claim, forcefully put by Erickson (1981), is that the results of research into speaker and lis­ tener skills are underm ined bv the artificiality of the paradigm used. Because interpersonal communication is an enterprise in which m eaning is socially negotiated, any context which deliber­ ately ignores that fact is of very limited value.