# Thinking and Deciding 4: Logic

### What is logic

Baron defines logic as “a normative model of inference, arrived at by reflection about arguments” (80). So logic is what we use as a standard to point out bad thinking. It’s interesting that Baron points out that it is reflection that leads us to our beliefs about logic. We come up with an argument style, and then try to generalize it. If we can think of counter examples, then it is not a good argument. This makes logic appear very inductive. Is there any better way to ground logic itself? I bet there is.

### Types of logic

Types of logic include propositional (if, or, not), categorical (all, some, not, no), and predicate. Logic is not a full theory of thinking because it is only concerned with inference, the drawing of conclusions based on evidence. Logic does not bear on how best to obtain evidence, just with the relation of evidence to conclusions. That’s a great way to describe the difference between rationality and logic. Logical thinking is often necessary for rationality, but not sufficient. You need good methods of gathering information in addition to good ways of adding that information up to get conclusions.

### Logical errors in hypothesis testing

Baron presents the four card problem here, but a thought occurred to me. The “if” statement is a bit ambiguous. If a card has a vowel on one side, then it has an even number on the other side. This could mean the barebones “if” or it could mean “if and only if.” This might explain some of the erroneous answers. The ambiguity of language is not helpful here.

Page 91 offers some evidence for rationalizations in the logic problems. Subjects made their decisions as to what seemed true, and then appeared to rationalize how they came to those conclusions. As mentioned in some other book (where was it? maybe Kurzban’s?), the content of the logic problem can make it easier to solve. Logic problems about social rules and obligations are easier than abstract logic problems.

### Extensions of logic

Baron claims that formal logic deals only with conclusive arguments, those involving absolute certainty. But I think this is totally false, since the premises may be true by different degrees of likelihood. So the uncertainty is a little hidden, but it does exist even in deductive logic.