According to the radical behaviourist view of B. F. Skinner, all behaviour, animal and human, isexplained by history of reinforcement and punishment, but there are many examples ofbehaviour and learning that don't appear to be based on this reinforcement.Edward Thorndike studied learning in animals, usually cats. He devised a classic experiment inwhich he used a puzzle box to empirically test the laws of learning. Considering Thorndike'spuzzle box experiments, it really appears like we do learn from trial and error, or at least catsdo.Thorndike found out that the amount of time it took for the cats to escape from the puzzlebox, the amount of time it took for them to hit that lever and escape, decreased verygradually over time.Is that all there is to learning for animals and humans? Well, there's lots of evidence thatthere is more to it than that. For example, with insight learning, now we have experience withthis every day, being humans.We'll see a problem and we'll come up with some solution in our heads and perform somebehaviour for the first time that solves the problem. Now, that behaviour wasn't everreinforced. We came up with the idea in our minds.This work has been done in chimpanzees as well. In Wolfgang Köhler's work, chimpanzeeswere given problems and it was found that they often mentally explored a problem beforeexhibiting a specific response that actually solved it.Another example of apparently unreinforced learning comes from a study by Edward Tolman.In the study, rats were given access to a maze. The rats went into the start box, they turnedright or left several times and then ended up in the end box.FutureLearn 2There were three groups in this study. One had constant reinforcement. They went into thestart box and they had cheese at the end box every time. Group two only started receiving thecheese reinforcement after 10 days. Group three never received any reinforcement.What was found was quite interesting. They seemed to have evidence that there was latentlearning. We'll explain that now.Group one received constant reinforcement, so they were reinforced every time they wentinto the maze, and they learned quite quickly to make very few errors.Group three, the group that was never reinforced, made fewer errors at the end and thebeginning, but they didn't really improve that much.What's interesting is group two. They were the only group that only received reinforcementafter day 10.Interestingly, they behave pretty much like group three until day 10, as well they shouldbecause both groups were unreinforced up to that point.But when you start reinforcing group two, that is giving the cheese reward, they showinstantly that even though they weren't displaying learning up until day 10, they must havebeen learning the maze. Because they were instantly able to perform as well as group one, thegroup that was reinforced all along.It's latent learning. It's really latent because there was learning going on up until day 10. Theyjust weren't displaying it until there was a reason to display it. This is an example of learningwithout reinforcement.Another type of learning is observational learning, where learning occurs by observing thebehaviour of a model.Albert Bandura conducted a series of experiments on observational learning, collectivelyknown as the Bobo doll experiments. A Bobo doll is just this inflatable doll that you can hitand knock it over and it comes back up.Bandura found that children who were exposed to an aggressive model were more likely tobehave aggressively toward the Bobo dolls than those that had not seen an aggressive model.What this shows is that these children, even though they are not reinforced for showingaggression, will show aggression just in order to model the behaviour.© Monash University FutureLearn 3M P Lepper conducted a particularly interesting study with magic markers. In the study,children were allowed to play with magic markers. If you've ever seen children drawing andplaying with coloured markers, they love doing it. They don't need to be reinforced for it.In the study, some children were rewarded, but not all. Some children, would receive acertificate of participation and not others. Later, the children were monitored on how muchthey played with the magic markers.The findings from the experiment were very interesting. Children who had been rewardedactually played less with the magic markers later on, when there was no longer a reward. Theinterpretation here is that the extrinsic reward of the certificate undermined something thatwas actually intrinsically motivated, the desire to play with the magic markers.You can almost imagine that it's like turning a magic marker play into some kind of job. Youcan imagine to thinking, well, I was only rewarded for something because this is work. Perhapsthis is something I shouldn't want to do. You can imagine thinking like this would undermineintrinsic motivation of some activity.Together, all of these different kinds of studies, different bits of evidence show there's reallymore to learning than reinforcing and punishing