The Probability Heuristics Model of Syllogistic Reasoning

The probabilistic approach to human reasoning

“…This is consistent with the probabilistic approach where the possibility has already been raised that some people might use System-2 processes to test conclusions generated by System-1 (Ref. 18). The critical question is the balance of System-1 versus System-2 processes in human reasoning.…”

“…Consequently, PHM uniquely makes predictions for the 144 syllogisms that are produced when Most and Few are combined with the standard logical quantifiers. We showed previously that these heuristics pick out the p-valid conclusions for these new syllogisms, and reported experiments that confirm the predictions of PHM when Most and Few are used in syllogistic arguments 18 .…”

“…This approach was first applied to Wason's selection task 5 , where many models have been proposed [6][7][8][9][10][11] and investigated [12][13][14][15][16][17] . More recently the probabilistic approach has been extended to the other core areas of the psychology of reasoning: syllogisms [18][19][20] and conditional inference [21][22][23][24] , where the need for a probabilistic approach has been most apparent [25][26][27][28][29] .…”

Natural Logic for Natural Language

“…In an empirical set-up of [7], the inference from 'No B C, All B A' to 'NotAll A C' (the fesapo pattern) is only recognized as valid in 8 percent of the cases, while in a staggering 61 percent of the cases, subjects think, erroneously, that the conclusion No A C follows from the premises. The only cases where the scores are still lower for endorsement of a valid conclusion are cases where the conclusion follows by existential import from a universal negative conclusion that is also valid, and that is recognized in a majority of cases as being valid.…”

The New Paradigm and Mental Models

“…Another main plank in JKG's argument is that MMT provides a better account of syllogistic reasoning than the probability heuristic model (PHM) [6]. However, the metaanalysis [7] they report comparing PHM with MMT used accuracy as a measure but did not allow PHM to predict no valid conclusion (NVC) responses.…”