Mental Probability Logic

The project continues my work on the development of a new theory of human reasoning, called "Mental Probability Logic". Mental Probability Logic will be a psychological competence theory of human reasoning that accounts for uncertain and defeasible inferences. The theory assumes that the "if-then" constructions of everyday arguments are interpreted by humans as high conditional probabilities. Humans draw inferences by applying defeasible rules, like the Probabilistic Modus Ponens. The rationality of human inferences will be evaluated by "coherence", the basic concept of subjective probability theory. The theoretical development will be complemented by conducting ten experiments to investigate the predictions of the Mental Probability Logic. Especially probabilistic versions of well known argument forms, and a basic and central system of defeasible reasoning will be investigated empirically.

The two central questions of this project were: (1) How do people draw conclusions from incomplete and uncertain knowledge? (2) What makes inferences under uncertainty rational? Both questions were investigated in the framework of probability logic. Probability logic investigates how conclusions can be drawn from uncertain and incomplete premises. The probabilities of the conclusions are usually between some lower and upper probability bounds, but sometimes one cannot infer anything about the conclusion based on the available premises: then, any probability value in the unit interval is coherent. From the probabilities that people respond to reasoning tasks, one can infer how they interpret the task material and especially conditionals (Ifthen constructions). It is subject of long debates in psychology andindependentlyof debates in philosophy how conditionals are interpreted. The project contributed to both debates and elaborated strong empirical and theoretical evidence that uncertain conditionals are best understood as conditional probabilities. People endorse basic rationality principles that govern the revision of conclusions in the light of new evidence. Moreover, project work revealed a novel finding concerning shifts of interpretations: While about half of the participants responded by conditional probabilities at the beginning of the experiments, almost all participants responded by conditional probabilities at the end of the experiments. Pioneering experimental work in this project on well-known paradoxes of the material conditional explains why people do not endorse the paradoxes, which provides evidence in favour of the conditional probability interpretation. Moreover, the project investigated how people negate conditionals in the context of the recent experimental philosophy movement. The rationality of classical fallacies of argumentation under uncertainty was explained in probabilistic terms and a new formal measure of argument strength was elaborated in the project. All this work constitutes a new competence theory of human uncertain reasoning, which is called Mental probability logic (MPL). MPL makes important contributions at the forefront of the new paradigm psychology of reasoning. Apart from psychology, the project made strong interdisciplinary connections to philosophy (formal and naturalized epistemology), mathematics (probability theory), and cognitive science (modelling uncertain inference). Project work culminated in 14 papers, 37 contributions to scientific meetings (among which 19 were invited), and the completion of one PhD thesis as well as three masters theses. The main project results belong to the domain of basic research. However, understanding human uncertain reasoning helps to design reasoning trainings for people in various fields including medicine, finance, education, law, science, and politics.