Scott Alexander, curated
← Back to curation

Can It Be Wrong To Crystallize Patterns?

Quality
73
Strong
Claude Shift
55
Moderate
RWI
2
of 10

Summary

A sharp conceptual essay on naming as cognitive technology. The setup: rationalist concepts (e.g. 'mysterious answers') feel obvious-in-retrospect, unlike factual knowledge (the capital of Cuba), because they aren't new information but the detection-and-naming of a pattern you'd half-seen. Crystallizing a pattern with a name ('trivial inconveniences,' 'Schelling points,' 'political correctness,' 'witch hunt') lets you hold it in mind, formulate a meta-level policy toward it, and dismiss instances 'in a single mental motion' -- which matters because debate is bottlenecked by mental-workspace and steps. He notes a leftist's shrewd request that he stop using 'political correctness': strike the term and a whole class of thoughts gets harder to think (unlike 'President,' which has easy substitutes). Then the title question: can crystallizing a pattern be WRONG? Cases like 'ley lines' (the concept tempts belief) and 'international Jewry' / 'mainstream media' / 'patriarchy' show that a name can smuggle an agency-fiction connotation, 'singling it out as A Unique And Interesting Problem' -- a HUGE practical problem -- but those are connotation failures, and renaming just reloads the connotation in ten minutes. Hunting for a clean case (the C.S. Lewis / Christian-clarity example) he can't find one he both recognizes as a crystallized pattern and judges net-harmful (Lewis's trilemma he actually likes). Provisional verdict: crystallization mainly misleads the already-at-risk via connotation, but for someone who can sort connotations it's usually good.

Why this score

Quality 73 · Strong. Strong (73): a genuinely original and useful conceptual contribution -- naming-as-cognitive-technology, the 'single mental motion' point, and the connotation-smuggling danger -- developed at real length and connected to live discourse (political correctness, witch hunts). Upper-Strong rather than Excellent because it ends provisionally/unresolved and is a focused meditation rather than a definitive framework.

Claude’s paradigm shift 55 · Moderate. Notable (55): the crystallizing-patterns frame and the analysis of how names smuggle agency-fiction connotations are a fresh, non-obvious synthesis, building on his Categories/concepts work and the LW concept-handles tradition rather than inventing the whole frame.

Real-world impact 2 · Minor. A genuinely original conceptual contribution — naming as cognitive technology (crystallizing a pattern lets you hold it, form a meta-policy, and dismiss instances 'in a single mental motion'), with the connotation-smuggling danger. Conceptual influence within rationalist discourse, ends provisionally, no material change — low RWI.