According to one more or less standard mythology, behaviorism, the ideology and methodology that reigned in experimental psychology for most of the century, has been overthrown by a new ideology and methodology: cognitivism. Behaviorists, one is told, didn't take the mind seriously. They ignored--or even denied the existence of--mental states such as beliefs and desires, and mental processes such as imagination and reasoning; behaviorists concentrated exclusively on external, publicly observable behavior, and the (external, publicly observable) conditions under which such behavior was elicited. Cognitivists, in contrast, take the mind seriously, and develop theories, models, explanations, that invoke, as real items, these internal, mental, goings-on. People (and at least some other animals) have minds after all--they are rational agents.
Like behaviorists, cognitivists believe that the purely physical brain controls all behavior, without any help from poltergeists or egos or souls, so what does this supposedly big difference come to? When you ask a behaviorist what the mind is, the behaviorist retorts: "What mind?"--and when you ask a cognitivist, the reply is; "The mind is the brain." Since both agree that it is the brain that does all the work, their disagreement looks at the outset to be merely terminological. When, if ever, is it right, or just perspicuous, to describe an animal's brain processes as thinking, deciding, remembering, imagining? This question suggests to some that the behaviorists may have been right about lower animals--perhaps about pigeons and rats, and certainly about frogs and snails; these simple brains are capable of nothing that should be dignified as properly "cognitive". Well then, where do we "draw the line" and why?
Do animals have beliefs? One of the problems with this question. which has provoked a lot of controversy among animal researchers and the ideologues of cognitive science, is that there is scant agreement on the meaning of the term "belief" as it appears in the question. "Belief" has come to have a special, non-ordinary, sense in the English of many (but not all) of these combatants: it is supposed by them to be the generic, least-marked term for a cognitive state. Thus if you look out the window and see that a cow is in the garden, you ipso facto have a belief that a cow is in the garden. If you are not ignorant of arithmetic, you believe the proposition that 2 + 2 = 4 (and an infinity of its kin). If you expect (on whatever grounds) that the door you are about to open will yield easily to your tug, then you have a belief to that effect, and so on. It would be more natural, surely, to say of such a person "He thinks the door's unlocked" or "He's under the impression that the door is open" or even less positively, "He doesn't know the door is locked." "Belief" is ordinarily reserved for more dignified contents, such as religious belief, political belief, or--sliding back to more quotidian issues--specific conjectures or hypotheses considered. But for Anglophone philosophers of mind in particular, and other theoreticians in cognitive science, the verb "believe" and the noun "belief" have been adopted to cover all such cases; whatever information guides an agent's actions is counted under the rubric of belief.
(The whole paper is now available in Daniel Dennett, Brainchildren, Essays on Designing Minds, MIT Press and Penguin, 1998.)