Foucault vs. Herman: Does Diagnosis Discover or Produce?
Both sides here agree that diagnostic categories are made rather than found, and that they carry power. The disagreement is over what follows: whether the right response is to hold the category under permanent suspicion as an instrument of normalization, or to fight for a better one on behalf of the people the current ones fail.
The Positions
Suffering is not invented. What is made is the category, and the authority to apply it. Psychiatry did not discover the abnormal individual; it assembled the examination, the case record, and the norm against which a person is measured, and the abnormal individual appeared as the object those instruments could see. Power here is not something a psychiatrist holds and wields. It runs through the apparatus, and it is productive rather than merely repressive.
Diagnosis is political, and the history proves it. Hysteria appeared and vanished with the political fortunes of the women who had it; PTSD entered the manual in 1980 because a veterans' movement put it there. That is not an argument against naming. It is an argument for naming better. Survivors of prolonged captivity get handed borderline personality disorder, and the label puts the damage in the patient's character rather than in what was done to her. Refusing to name it frees no one. It leaves the worse name in place.
Clinical Implications
Foucault died in 1984 and never wrote about the DSM, so what transfers here is a method rather than a verdict on any particular manual. What it asks of a therapist is not the unanswerable "am I discovering something real or constructing it" but something narrower: what does this category let me do, to whom, and what kind of person does the client become once it is written down. The therapist who never questions diagnostic categories risks becoming an agent of normalization. The therapist who rejects all diagnosis risks being unable to communicate with other professionals or access resources for their clients, and, as Herman would add, leaves whatever worse name is already on the chart exactly where it is.
In Session
Consider a client who hears voices. A standard psychiatric frame diagnoses psychosis and considers medication. A Foucauldian frame does not ask who benefits from the word, since that assumes power is something somebody holds. It asks what the word makes possible: what gets recorded, who is now authorized to decide, and what the client becomes able and unable to say about themselves once they are a person with psychosis. A Hearing Voices Network approach treats the voices as meaningful experience to be worked with rather than symptoms to be eliminated. All three have clinical consequences: for the client's identity, treatment, and relationship to their own experience.
Toward Resolution
The instrumental answer is the working one: use diagnosis as a tool that opens doors (treatment, insurance, communication) while holding it as a historical construction rather than a natural kind. But Foucault has a good reply. Private skepticism is not an exit; the category still enters the record and still does its work, whatever the clinician privately believes. Herman's answer is the more demanding one, because it means contesting the category rather than using it with reservations. Complex PTSD is the test case, and the result is instructive: kept out of the DSM, admitted to the ICD-11 in 2018.