Humans and Large Language Models Infer a Multidimensional Structure of Moral Motivation
Identification
- Clé publique
- radar:r-a0a44f4481a268d59ce29d980036e573
- Auteurs
- Paul C. McKee · Andrew James · Breanna K. Nguyen · Madison Monroe-Mohajerin · Scott A. Huettel · Walter P. Sinnott-Armstrong
- Année déclarée
- 2026
- Langue déclarée
- en
- Type source
- preprint
- Première détection
- 2026-11-14T02:29:22.697169+00:00
- Dernière observation des métadonnées
- 2026-11-14T02:29:22.697169+00:00
- Dernière vérification HTTP
- Non effectuée
Institutions déclarées par la source
Duke University
Pays d’affiliation déclarés
US
Résumé documentaire — langue originale
Moral evaluation depends not only on what people do, but on why observers believe they acted as they did. Although motive inference has long been recognized as central to moral judgment, moral motivation itself has rarely been conceptualized as a multidimensional psychological representation. We investigated whether observers infer four distinct moral motivational dimensions (Approach Good, Avoid Bad, Approach Bad, and Avoid Good) from natural-language narratives and whether this representational structure is recovered by frontier large language models (LLMs). Human participants (N = 1,031) and 23 frontier LLMs evaluated 344 real-world moral narratives using identical procedures. All four motivational dimensions were inferred frequently in both populations and exhibited structured patterns of co-occurrence. Human observers clustered antisocial motivations (Approach Bad and Avoid Good) more strongly than prosocial motivations, whereas LLMs exhibited the opposite pattern. Across all evaluative outcomes, inferred motivations systematically predicted judgments of rightness, wrongness, praiseworthiness, and blameworthiness. Supervised machine learning analyses further demonstrated that the complete four-dimensional representation consistently outperformed every reduced two-dimensional alternative, indicating that each motivational dimension contributed unique predictive information. Although LLMs broadly recovered the multidimensional structure observed in humans, they inferred antisocial motivations less frequently, concentrated those attributions within fewer narratives, and produced systematically compressed response distributions characterized by lower variance and entropy. Together, these findings suggest that moral motivation is best understood as a multidimensional representational system and that representational richness provides a useful framework for evaluating both human moral cognition and its approximation by artificial systems.
Identifiants
- OPENALEX
- https://openalex.org/W7223214499
- DOI
- 10.31234/osf.io/u4r3q_v1
Repères documentaires
- Famille documentaire
- preprint
- Types déclarés par les sources
- preprint
Les mentions d’entités ne constituent pas une preuve de reconnaissance juridique. La juridiction étudiée est distincte du pays d’affiliation des auteurs.
Sources
OpenAlex · Document 1
Citer cette référence
Paul C. McKee; Andrew James; Breanna K. Nguyen; Madison Monroe-Mohajerin; Scott A. Huettel; Walter P. Sinnott-Armstrong. Humans and Large Language Models Infer a Multidimensional Structure of Moral Motivation. 2026. https://doi.org/10.31234/osf.io/u4r3q_v1
JSON · CSL-JSON · BibTeX · RIS · CSV · JSON-LD ·
APA, Chicago, ISO 690 : formats bibliographiques fondés sur les champs disponibles. OSCOLA / Bluebook : données structurées uniquement ; aucune conformité juridique complète n’est revendiquée sans vérification des éléments nécessaires.
- APA
- Paul C. McKee; Andrew James; Breanna K. Nguyen; Madison Monroe-Mohajerin; Scott A. Huettel; Walter P. Sinnott-Armstrong. (2026). Humans and Large Language Models Infer a Multidimensional Structure of Moral Motivation. https://doi.org/10.31234/osf.io/u4r3q_v1
- Chicago
- Paul C. McKee; Andrew James; Breanna K. Nguyen; Madison Monroe-Mohajerin; Scott A. Huettel; Walter P. Sinnott-Armstrong. Humans and Large Language Models Infer a Multidimensional Structure of Moral Motivation. 2026. https://doi.org/10.31234/osf.io/u4r3q_v1
- ISO 690
- Paul C. McKee; Andrew James; Breanna K. Nguyen; Madison Monroe-Mohajerin; Scott A. Huettel; Walter P. Sinnott-Armstrong. Humans and Large Language Models Infer a Multidimensional Structure of Moral Motivation. 2026. https://doi.org/10.31234/osf.io/u4r3q_v1
Warning: Undefined array key "last_link_verified" in /home/clients/e27d81e12f04d2bfd747aa264540ac56/sites/_UDTR_APP/legal_library_reference.php on line 31
UDTR — Bibliothèque juridique mondiale des droits de la Nature. Humans and Large Language Models Infer a Multidimensional Structure of Moral Motivation. Notice documentaire: https://www.declarationuniverselledesdroitsdelarbre.org/global-rights-of-nature-radar/reference/r-a0a44f4481a268d59ce29d980036e573/ · Dernière vérification HTTP : non effectuée
Données de cette version figée
Dossiers et relations documentaires
Rôles des sources
Source bibliographique / publication d’origine · https://doi.org/10.31234/osf.io/u4r3q_v1
Un lien vers une autorité ne prouve pas à lui seul une affirmation juridique. Les résumés et les textes intégraux restent soumis aux conditions de leur source.
Statut documentaire
Pertinence à examiner
Bibliothèque
Cette identité n’est pas encore intégrée à la Bibliothèque juridique.