When something serves a function, it is easy to overlook its origins. The tendency is to proceed directly to function and retroactively construct a story about origin based on the function it fills. In this article, I address this problem of origins as it appears in the sociology of knowledge, using a case study of the publication of the 3rd edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) in 1980. The manual revolutionized American psychiatry and the treatment of mental illness, because it served the function of classification that had become critical to the field of mental health by this time. But this function must be bracketed in order to reveal the "extra-functional" origins of the DSM-III. Using field theory, I argue that the manual was necessary for reasons other than the function it filled as a classification. Specifically, its origin lies in a series of conflicts among psychiatrists, psychoanalysts, and clinical psychologists within the field of mental health, which followed in the wake of the collapse of psychoanalysis as the dominant treatment type for mental illness. I reveal the generative formula behind the production of the DSM-III, capturing a variety of social processes that influenced the format of the manual and made it a useful classification, but which are not reducible to function. In this way, I reproduce its raison d'etre in a manner similar to how the DSM-III appeared for the people who produced it. This focus on generative formulas offers the sociology of knowledge a way to capture the epistemic importance of a range of different social processes. Most importantly, it avoids the functional fallacy of reducing origin to function, and ignoring the idea that innovations might appear necessary even without clear recognition of their functional consequences.