Edward L. Thorndike

American psychologist (1874–1949), a founder of educational psychology and comparative (animal) learning research, whose puzzle-box experiments produced the law of effect — behavior followed by a satisfying consequence is "stamped in" — and whose 1920 paper "A Constant Error in Psychological Ratings" coined the term halo ("the constant error of the 'halo,' as we may call it"); the fixed compound halo effect is a later crystallization, not his wording. He matters to this vault as the origin, a century upstream, of a mechanism the vault's modern source-evaluation cluster rediscovered without naming: the bleed of a global impression into ratings meant to be independent.

Thorndike's law of effect is also a recognized ancestor of reinforcement learning's credit-assignment problem, tying him to the vault's AI-history threads as well as its epistemics ones.

References

• claim-thorndike-1920-halo-effect-ratings-too-high-and-too-even
• observation-halo-effect-names-the-2025-reliability-credibility-leak
• entity-halo-effect
• claim-thorndike-1904-attributes-jingle-fallacy-to-aikins-not-himself — 2026-09-17: correction, not a new topic. This page's entity-jingle-fallacy connection has read as "Thorndike coined the jingle fallacy (1904)" since first_seen; a direct read of his own 1904 primary shows he is citing Herbert Austin Aikins's term, not naming his own. Thorndike's role in this vault's jingle-fallacy thread is the citer, not the coiner.
• Capture: 10-inbox/raw/2026-08-03-hop-thorndike-halo-effect.md
• 2026-09-17 (propagation-repair): was "whose 1920 paper … coined the term halo effect" -> now the 1920 paper coined halo (p.28), with halo effect named as the later crystallization, not Thorndike's wording. Brings this page in line with the 2026-08-09 ruling on claim-thorndike-1920-halo-effect-ratings-too-high-and-too-even. The 2026-08-10 propagation check there looked only at the 1917/Scott misattribution.