Psychological Research Of Conditional Reasoning

The theme of this paper was the pragmatic research of conditional reasoning. Thelogic rules expressed two problems, one was the true or false of a sentence whichshould be interpreted by material implication, and the other was the justificationwhich should be interpreted by probability. The purpose of psychology research wasnot to confirm the logic rules to be wrong but to extend the explainable range ofinductive and reductive logic since Aristotle.Every progress of logic and linguistics could affect the research paradigm ofpsychology, logic research of conditional sentences had three aspects: hypothesis ofmaterial implication，hypothesis of possible World’s semantics and hypothesis ofsentence probability semantics. The theory of conditional psychology could beclassified in three orientations.The problem of conditional sentences initiated material implication paradox;hypothesis of possible world’s semantics could interpret subjunctive conditionalpsychology, while it did not have quantitative predictive power; although hypothesisof sentence probability semantics encountered the challenge of triviality result, itcould integrate the prior two hypotheses and it was the theme of reasoning research.This paper expounded the four main orientations of reasoning psychology:mental logic theory, mental model theory, conditional probability calculation model,and dual process theory. On the basis of content effects, this paper analyzed thecontents transformation from logical form to semantics and pragmatics.According to Speech act theory; this paper distinguished the pragmatic featuresof conditional sentence into seven styles: statement, suggestion, promise, threat, order,warning and causation. Statement corresponded to the indicative reasoning and thelast six styles corresponded to deontic reasoning.According to the analysis of conditional sentences and conditional reasoning, thispaper putted forward seven hypotheses:(1) Conditional reasoning research should surpass material implicationhypotheses and use sentence probability semantics as the research paradigm. (2) Naturally, conditional reasoning was symbol processing, it included threelevels of mental process: syntax, semantics, and pragmatics. It could be interpreted bysyntax reasoning, semantics reasoning, pragmatics reasoning and formal logicreasoning.(3) Wason’s selection task was the classical paradigm of conditional reasoning, itchanged deductive reasoning into inductive reasoning, changed the certaintyreasoning into uncertainty reasoning. It should be interpreted by theory of deductivereasoning.(4) Different probability model interpretations of Evans，Johnson-laird andOaksford were because the disunity between conditional probability and conditionalsentence probability. Reasoning should refer to conditional sentence probability.(5) The dichotomy of indicative reasoning and deontic reasoning could notinterpret the Wason’s selection task effectively. Every theory of deontic reasoning wasdomain particularity, Speech act theory and probability theory could integrate all thereasoning theories.(6) Under indicative sentence reasoning, there were interaction between factelements and probability elements.(7) Difficulties of reasoning problem, the reasoning level, verbal ability andverbal thinking tendency could affect conditional reasoning.According to the above hypotheses, this paper carried out four researches:As the theme of Wason’s selection task, Research one designed threeexperiments. Experiment one compared the differences of various reasoning effects.The results showed that the difficulty of Wason’s selection task was proved again,confirmation bias could not interpret Wason’s selection task, matching bias includedcontent matching and position matching and real contents could facilitate Wason’sselection task. Experiment two and three took the correct rate and reaction time as theindex, and showed that Wason’s selection task was the logic of discovery not the logicof verification. Subjects reasoned from the particular to the general not from thegenerality to the particularity.As the theme of reasoning content effects, Research two designed twoexperiments. Experiment four proposed four reasoning processes: syntax reasoning, semantics reasoning, pragmatics reasoning and formal logic reasoning. Firstlysubjects reasoned by syntax, then by semantics and pragmatics, only higher cognitionability subjects could reason by formal logic. The richness of the language couldfacilitate the solving of reasoning problems, while the complexity of language couldhinder the solving of reasoning problems. Experiment five discusses the feasibility ofthe dichotomy of indicative reasoning and deontic reasoning, it could not interpret thedifferent roles of violation and falsification for counterexample. The higher theprobability of conditional sentence was, the more subjects would rather accept theviolation for counterexample and more difficulty complete the reasoning task.As the theme of pragmatics, Research three designed three experiments.Experiment six distinguished deonic reasoning into six types, explored the interactionamong deontic reasoning, logic training and reasoning forms. Their results, logictraining could increase the reasoning acceptance of MP and MT; command and threatconditional sentences were the two kinds of opposite forms, the acceptance of otherconditional sentences was between the two. The acceptance of four reasoning formsunder threat conditional was higher than the acceptance of other conditional sentence.The acceptance of different reasoning forms was no significant difference in threatconditional sentence. The acceptance of reasoning forms was low in suggestion andthreat conditional sentence.Experiment seven explained the deontic reasoning by probability; the resultsdiffered from the conditional probability calculation model and P-Q mapping model.this paper found the higher relationship between the whole sentence probability andinferences. Experiment eight compared the effects of factual and probability factors inindicative conditional sentences. Their results, the various combinations of differentantecedents and consequents could affect the reasoning results, HH was higher thanother combinations in factual conditionals, while the same results were not found inno-factual and quasi-factual conditionals, especially for AC, the acceptance of HHwas lower than LL. In difference factual conditionals, MP was high and MT was low.The acceptance of AC was higher than other reasoning forms in quasi-factualconditionals. In no-factual conditionals, the acceptance of DA、AC and MT waslow,and the acceptance of all the probability combinations was lower than factual conditionals.As the theme of verbal factors, Research four designed two experiments,developed a verbal-image cognitive style analysis test based on the charactersmeaning and symbol constructions and explored the verbal effects in reasoning. Theirresults, the verbal effects were not significant in complicated reasoning. The scores ofverbal ability, Public Servant Exam and analogical reasoning in Public Servant Examhad no significant relation with language ability and Graphical reasoning ability.Cognitive style analysis test could differentiate the subjects effectively, and theacceptance of MT in Wason’s selection task and real contents conditionals reasoningrelated to verbal-image test scores significantly.Deduction and induction were the both ends of cognitive continuum theory ofBrunswik, conditional reasoning was between the two, and it also was the generaltype of hypothesis testing. Abstract logic conditional reasoning, indicative reasoningand deontic reasoning need different cognitive processes. When researchers lookedupon the conditional (belong to deductive reasoning) from the pragmatics andprobability, conditional reasoning was changed from the logic of verification into thelogic of discovery. The conditional research became domain special problems...