Abstract
The paper analyses an interview describing how K came to be defined by her friends as mentally ill. The method of analysis assumes that the structure of the conceptual scheme `mental illness' which the reader uses in recognizing `mental illness' is isomorphic with that organizing the text and hence is discoverable `in' it. The full text of the interview is presented as the data. The analysis explicates the interpretation of the text as a method of reading. The text is found to provide instructions for its interpretation and for the authorization of its facticity. K's mental illness is to be located in the collection of instances of K's behaviour which the interview records. How is behaviour to be described as `mentally ill type' behaviour? It is suggested that the interview as a whole organizes a `cutting-out' procedure whereby K's behaviour is presented as making sense neither to her friends nor to the reader of the text. The procedure involves showing for each instance of her behaviour as well as for the collection as a whole that K's behaviour is not properly provided for by relevant social rules or definitions of the situation. To be recognizable as `mentally ill type' behaviour examples of K's actions must be constituted as anomalies rather than as deviations from a norm or rule.