Rationality, cognitive bias, and artificial intelligence: a structural perspective on quantum cognitive science

Human beings are not completely rational; there is some irrationality, as well as bounded rationality, involved in the nature of human thinking. It has been shown through recent advances in quantum cognitive science that certain aspects of human irrationality, such as cognitive biases in the Kahneman-Tversky tradition, can be explained via mathematical models borrowed from quantum physics. It has also been shown in quantum cognitive science that human rationality exhibits a special sort of non-classical phenomenon as observed in quantum physics as well, namely the phenomenon of contextuality, which extends the notion of non-locality, what Einstein called “spooky action at a distance”. In the present paper we elucidate and articulate the nature of human rationality and irrationality as observed in cognitive bias experiments and cognitive contextuality experiments. And we address the question whether non-human agents, such as animals and robots, can exhibit the same sort of cognitive biases and cognitive contextuality. Technically, we shed new light on these (quantum) cognitive experiments from the viewpoint of logic and category theory. We argue, inter alia, that the logic of cognition is substructural or monoidal, rather than Cartesian (which encompasses classical, intuitionistic, etc.), just as the logic of quantum mechanics and information is substructural or monoidal. The logic of reality is thus intertwined with the logic of cognition; the logical link between physical reality and the conscious mind would possibly allow us to go beyond the Cartesian dualism separating matter and mind as intrinsically different entities.