Learning about learning - it's a bit like being in a vortex of ever decreasing circles (or is that mixing metaphors?).
The behavourists had a nice simple idea about learning: cause and effect. Watson (no relation to he of Sherlock Holmes fame, as far as I know), said that all the fluffy, cognitive and social psychological research was meaningless, and that one should only study things that can be counted or measured scientifically. He was the pioneer of behavioural research, and famously created in a child called Little Albert a phobia of his white rat, by associating it with a loud noise - all in the name of psychological research. The ethics of this are undeniably dodgy, and in a less controversial but just as well known study, Pavlov conditioned dogs to salivate at the sound of a bell. He did this by creating an association between the bell and food, so that the dog eventually came to expect food when the bell was rung. In Django's case, the words "Are you hungry?" have the same effect!
These two studies concentrated on the learning or 'conditioning' of involuntary behaviour. BF Skinner went a step further and did studies on how to change voluntary behaviour. Using pigeons or rats in a box, he had a set up such that when they pecked a particular spot, or pressed a lever, they received a reward of food, thus reinforcing the behaviour. This type of learning is known as operant conditioning, because the association is between an action and a reward. Operant conditioning introduced the theory of positive and negative reinforcement, which has been very influential in behaviour modification efforts.
Positive reinforcement is simple - if the behaviour is correct, there is a reward, while negative reinforcement means that if the behaviour is correct, then something unpleasant (such a loud noise or pain) is stopped. Over time, the behaviour is reinforced and gradually developed - a process known as shaping. Skinner believed that these were the best ways to change people's behaviour, and was totally against punishment. He said that it doesn't work long term, and that it can lead to other unwanted behaviour.
In the above studies, the animals involved were reacting automatically to stimuli, known as 'stimulus-response' learning. They took no account of conscious thought being involved - unlike cognitive psychologists, the behaviourists did not consider that anything went on between the ears.
However, one researcher following on from Skinner did find that there may be more to animals' behaviour than simple stimulus-response. Tolman did experiments with different routes and mazes, which suggested that rats were using knowledge flexibly to make inferences about where the food was, rather than reacting automatically.
Behavourists work on the premise that research on animals can be extended to humans, and indeed some of their studies have proved very useful in human applications, particularly in classroom situations, treating phobias and aversion therapy.
However, it's hard to say that animal learning as a whole can be generalised to humans, partly because animals learn some things better than others - ie. where food is involved. Can you see a pigeon getting enthusiastic about pecking a button on the promise of a new pair of football boots...?