howreasoningworks

Using the framework of the argumentative theory of reasoning, it is possible to shed a new light on the workings of reasoning. According to this theory, reasoning is used to evaluate reasons. Reasons are representations, therefore reasoning is a metarepresentational mechanism. Metarepresentational mechanisms deal with representations of representations. The most famous example is the ability to attribute mental states to other agents (as in 'I believe that Paul thinks that it's raining'), known to psychologists as theory of mind. Reasoning also deals with representations, but is only interested in the properties that make some representations good reasons to accept (or reject) other representations.

This means that the argumentative theory belongs with the other dual process theories of reasoning. These theories draw a line between mechanisms that we can call intuitions (or system 1, heuristic, associative), that tend to be fast, unconscious and nearly effortless and reflective mechanisms (or system 2, analytic, rule-based), that would be slow, controlled and conscious. Building on these theories, Dan Sperber and I have tried to suggest a more principled distinction in "Intuitive and reflective inferences".

Lately, I have been interested in developing an analogy between reasoning and visual search. The core idea is that reasoning would look for good reasons among our representations in much the same way that we look for, say, an improvised tool in a visual scene. It was written up there, along with some other ideas about how people search for arguments: Mercier, H. (2012) "Looking for arguments." Argumentation.