The Weak Completion Semantics Can Model Inferences of Individual Human Reasoners

Christian Breu, Axel Ind, Julia Mertesdorf, Marco Ragni

Publikation: Kapitel i bog/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

Abstrakt

The weak completion semantics (WCS) based on three-valued Łukasiewicz logic has been demonstrated to be an adequate model for general human reasoning in a variety of different domains. Among the many experimental paradigms in cognitive psychology, the Wason Selection Task (WST) is a core problem with more than 200 publications demonstrating key factors of the systematic deviation of human reasoning from classical logic. Previous attempts were able to model general response patterns, but not the individual responses of participants. This paper provides a novel generalization of the weak completion semantics by using two additional principles, abduction and contraposition: This extension can model the four canonical cases of the WST for the Abstract, Everyday, and Deontic problem domain. Finally, a quantitative comparison between the WCS predictions of the extended model and the individual participants’ responses in the three problem domains is performed. It demonstrates the power of the WCS to adequately model human reasoning on an individual human reasoner level.

OriginalsprogEngelsk
TitelLogics in Artificial Intelligence - 16th European Conference, JELIA 2019, Proceedings
RedaktørerFrancesco Calimeri, Nicola Leone, Marco Manna
ForlagSpringer
Publikationsdato9. maj 2019
Sider498-508
ISBN (Trykt)978-3-030-19569-4
ISBN (Elektronisk)978-3-030-19570-0
DOI
StatusUdgivet - 9. maj 2019
Udgivet eksterntJa
Begivenhed16th European Conference on European Conference on Logics in Artificial Intelligence - Rende, Italien
Varighed: 7. maj 201911. jul. 2019

Konference

Konference16th European Conference on European Conference on Logics in Artificial Intelligence
LandItalien
ByRende
Periode07/05/201911/07/2019
NavnLecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science)
Vol/bind11468
ISSN0302-9743

Fingeraftryk

Dyk ned i forskningsemnerne om 'The Weak Completion Semantics Can Model Inferences of Individual Human Reasoners'. Sammen danner de et unikt fingeraftryk.

Citationsformater