A broken bone is confirmed by an image and an infection by a culture. Depression, anxiety disorders and bipolar disorder are diagnosed from an interview, and the reason is structural rather than a gap waiting to be filled.
The categories were built from observation
Psychiatric classification developed by grouping symptoms that clinicians repeatedly saw occurring together, then naming the cluster.
The starting point was the pattern, not an underlying biological lesion, because the biology was not accessible when the categories were formed.
A diagnosis therefore names a recognised pattern of experience and behaviour. It does not assert a single identified cause.
Criteria are explicit and countable
Modern manuals specify which symptoms count, how many must be present, how long they must have lasted, and how much interference with daily functioning is required.
Duration and impairment thresholds are what separate a diagnosis from ordinary distress, since low mood after a setback shares its features.
Codifying the criteria this way was an attempt to make different clinicians reach the same conclusion about the same person, which earlier approaches did not achieve.
Biology does not respect the boundaries
Research has identified biological differences associated with psychiatric conditions at a group level, including patterns in brain structure, activity and genetics.
These findings overlap heavily between diagnoses, and they overlap with people who have no diagnosis at all. The distributions sit on top of one another.
A marker useful for diagnosis needs to separate individuals reliably, and group-level differences with wide overlap cannot do that.
One diagnosis probably covers several conditions
A symptom-based category will contain people whose underlying biology differs, if different mechanisms can produce a similar presentation.
Two people can meet criteria for the same disorder while sharing only a minority of the listed symptoms, given how the criteria are structured.
This heterogeneity is a leading explanation for why marker research has not produced a clean test, and why some researchers now study symptom dimensions instead of categories.
What a diagnosis is for
Its value is practical: it summarises a presentation, guides which treatments have evidence behind them, and allows communication between clinicians.
It is a working description that can be revised as the presentation changes. Any assessment belongs with a qualified clinician, since the criteria require judgement that a questionnaire cannot supply on its own.