A Psychological Study Of Conditional Reasoning Mechanism

The purpose of this study is to explore the conditional reasoning mechanism bymeansofpsychologicalmethodandthetheoryofinformation-processing.Based on logical and mathematical analysis, it is found that in present conditionalreasoning models, both the logical mechanism and non-logical mechanism areconfused with each other and logical truth is treated as factual truth. Resulting fromthe failure to differentiate logical reasoning mechanism from non-logical reasoningmechanism, present conditional reasoning models such as mental model and mentallogical model can neither explain the reasoning based on logical mechanism nor thereasoningbasedonnon-logicalmechanism.The study investigated mathematical implication of conditional probability andfound that the probabilities of antecedent and consequent of conditional propositionobtained from different probability space respectively did not meet the formula forconditional probability. Therefore, the conditional reasoning models based onconditional probability imply a supposition that there is a category to which both theantecedent andtheconsequent belongandreasoners canuseit as commonprobabilityspace for measuring probabilities of antecedent and consequent. However, there is noanyevidencetoprovethesupposition.To verify the conditional reasoning models based on conditional probability, threeexperiments on parameter correlation, probability consistency and common categorysupposition were performed. Experimental results show that subject's reasoning doesnot meets the requirements of parameter correlation, probability consistency andcommoncategorysupposition.On the basis of inspection on information process of reasoning, the study putsforward an assumption that there is multiple-mechanism for conditional reasoning,which means reasoner may start more than one mechanism to deal with conditionalreasoning task. The various type of mechanism can fall into two main categories:logical mechanism, non-logical mechanism. For each type of reasoningmechanism, adefinition was given. An information processing diagram of conditional reasoningmechanismwasconstructed.Anew conditional reasoningmodel: P–Q castinginterrelation model, is presentedin the light of investigation on non-logical conditional reasoning mechanism. For theconditional reasoning related to LH and HLtypes of proposition, the prediction made by P- Q casting interrelation model is the same as that made by probabilisticcomputationalmodel.ButfortheconditionalreasoningrelatedtoLLandHHtypesofproposition, the prediction made by P- Q casting interrelation model is opposite towhat made by the probabilistic computational model. An experiment was designedand performed to find the fitter model between the two models. Experimental resultsshow that subject's reasoning performance is identical with what the P- Q castinginterrelationmodelpredicted.A theoretical and experimental study was carried out on the reasoning based onlogical mechanism. As a result, a conditional reasoning model based on logicalmechanism was presented, which can explain the reasoning process led tonon-probabilistic conclusion. Experimental results show that propositional attitudeplays an important part in determining which type of reasoning mechanism will beswitched on and the subjects who never learn logic can also perform the reasoningbasedonlogicalmechanism.The study maintains that the appearance of logical paradox such as implicationparadoxandthepresentationofG?del'sincompletenesstheoremshowthattheformalaxiomatic method in logic has come to the limit, and logic need to find new ways todevelop. Psychological research pattern based on analysis of information-processingisjustthewayneededfortheadvancementoflogic.