In this article, we contribute to understanding the interactional aspects of making clinical diagnosis in mental health care. We observe that therapists, during the “problem presentation” sequence in clinical encounters, often use a specific form of diagnostic formulations to elicit more diagnostically relevant information. By doing so, they often substitute one type of verb with another, following a diagnostic hypothesis. Specifically, in interviews that arrive at a diagnosis of neurosis, therapists formulate with behavioral verbal processes; in interviews that arrive at a diagnosis of psychosis, they do so with material ones. Such formulations often prove useful to define clinical diagnoses. They can, however, also be dangerous in that they may favor the therapist’s agenda over the patient’s. Our analysis helps therapists not only better understand the diagnostic process but also reflect upon their own use of diagnostic formulations and become aware of the clinical effects of their interactional performance.