The debate about knowledge-production in sociology has pitted “internalist” accounts, which pay close attention to the places, practices, and tools of knowledge, against “externalist” accounts of institutions and fields. Using psychoanalysis as a case, this paper develops an approach that integrates these traditions by comparing the differentiation of places, tools and practices of knowledge production. The paper shows that, in a context in which other areas of practice increasingly differentiate research, diagnosis and treatment in spaces, tools, and professional roles, psychoanalysis invokes that differentiation rhetorically but refuses to differentiate its practice. Psychoanalysts insist on a specific setting – the couch and the psychoanalytic relationship – as central to all aspects of their knowledge-production but they do not adapt this space to pursue any of these purposes in their own right. This analysis explains some of the problems psychoanalysis has with its environment and the specific form divisions take within psychoanalysis. As an unusual case of non-differentiation, psychoanalysis highlights the role differentiation plays in other areas of knowledge-production.

Standard clinical diagnostic procedures are often inappropriate and frequently not feasible to apply in population-based studies, yet ascertaining accurate disease status is essential. We conducted a systematic review to identify algorithms, criteria, and tools used to ascertain 17 chronic diseases, and assessed the feasibility of developing algorithms for the CLSA. Of the 29,616 citations screened, 668 papers met all inclusion criteria. We determined that the information included in a disease algorithm will differ by condition type. The diagnosis of some symptomatic conditions, such as osteoarthritis and arthritis, will require substantiation by clinical criteria (e.g., x-rays, bone density measurement) while other conditions, such as depression, will rely solely on self-report. Asymptomatic conditions, such as hypertension, are more difficult to ascertain by self-report and will require additional physiologic measures (e.g., blood pressure) as well as laboratory measures (e.g., glucose). This pilot study identified the tools necessary to develop disease ascertainment algorithms.