Scott Alexander, curated
← Back to curation

Symptom, Condition, Cause

Quality
72
Strong
Claude Shift
54
Moderate
RWI
2
of 10

Summary

A philosophy-of-psychiatry essay answering the recurring objection that psychiatric categories (autism, depression) aren't unitary. Scott: psychiatrists know this, and diagnoses needn't match root causes to be useful. Introduces a middle 'condition' level between symptom and cause via the pulmonary-edema analogy (one condition, many causes — heart/kidney failure, overdose; matched to a biological process = 'condition,' unmatched = 'syndrome') and stroke (ischaemic vs haemorrhagic — different causes/treatments, still a useful concept). Depression is likely a 'condition' (the SIGECAPS cluster) with many causes; his wild guess is that psychiatric disorders are 'computational conditions' (e.g. learning rate set too high). The subtype-identification program has been century-long and unproductive; better to understand what depression IS. 'All models are wrong, some are useful.'

Why this score

Quality 72 · Strong. Strong, upper. A clarifying conceptual contribution — the symptom/condition/cause trichotomy (with the pulmonary-edema and stroke analogies) usefully defends psychiatric nosology against the 'it's not unitary!' objection, and the computational-conditions guess is a sharp framing. Held at top-Strong as a focused conceptual essay rather than a major synthesis.

Claude’s paradigm shift 54 · Moderate. Moderate–Notable (54). The symptom/condition/cause framing + computational-conditions hypothesis was a fresh, clarifying contribution to psychiatric philosophy at publication.

Real-world impact 2 · Minor. Minor (2). A useful conceptual tool within psychiatric/rationalist discourse; no material reach.