So as I said, beyond the specificness of the learning mechanisms, behaviorists believe three main things. The first is that there's no innate knowledge. These three mechanisms are all an animal would need. The second is anti-mentalism. You would explain mental life without mental stuff like desires and goals, but just through principles that linked up stimulus and response. And the third is that these learning mechanisms will work for any stimulus and any response. And for any animal, the way humans learn is no different in principle from how rats learn, and how monkeys learn, and how pigeons learn. I think it's fair to say that nobody believes this anymore. Just about everybody agrees to at least some extent that these three claims are mistaken. First, there's a lot of evidence for unlearned knowledge. We'll see in subsequent lectures that was evidence for an eight contribution to capacities like learning language, perceiving a physical world, understanding number, certain aspects of sexual preference, and so on. Furthermore, even when it comes to learning, which is the focus of behaviorism, it turns out different animals have different learning mechanisms. The way birds learn birds song seems quite different from the way that humans learn language. The way some animals learn to navigate is just different from the way other animals learn to do this. The way some animals form attachments or develop cooperative behavior is often special to that sort of species of animal. And none of this is that surprising, natural selection will take different creatures on different paths. As for the sort of middle proposal that is unscientific to talk about observables. In some way, this is an understandable reaction to Freud. Who made all sorts of dramatic claims that were, as we said, unfalsifiable and difficult to pin down and horribly vague. But the truth is that all sciences talk about an observables. Biologist talked about genes even before we even knew what one looked like. Physicists will talk about string theory. And certainly for psychology, when explaining a complex and intelligent mechanism, it makes sense to appeal to internal representation. Suppose to had to ask how a complex chess machine worked. You would say, well, it has this sort of rule built into it, it has this sort of strategy, and so on. You would appeal to the internal workings in a machine and you'd be right to do so, because things like computers have internal workings that can only be explained in terms of things like models of the board, and rules, and strategies, and so on. Similarly, humans, there's something going on inside our head. One of the objections to behavior is that it's crazy to deny that we think even if we don't move. That we have dreams, that we have fantasies, that we deliberate over things, and a behaviorist might say that's fine, but that's not scientific. But it turns out that to explain even the most basic of human behaviors, you need to appeal to these internal representations, and true as well for animals. So it turns out that animals, for instance, get better and better at learning something, understanding something, even without reinforcement and punishment. And there is a rich literature developed by the psychologist Tolman, on what's called latent learning, which is learning without any sort of feedback. And this is all sort of common sense. You could imagine yourself looking over a map or making your way through a house, and coming to understand the spacial arrangement without any sort of reinforcement. The idea that reinforcement, at least in any sort of specific concrete sense, is not plausible. And here's a nice illustration of an experiment that shows how animals might really do in an interesting sense, have maps in their head. [MUSIC] >> This maze is designed to test whether rats can invent a new route in their heads. Most of the possible routes to the food are blocked off. [MUSIC] There is only one route to run down, and the rat appears to find its way to the food by trial and error, the same way Pippin learned to pick up the suitcase. [MUSIC] But it also appears to re-check the layout before using the route its learned. [MUSIC] Now comes the crucial part, the learned route is blocked off and new ones are opened. Can the rat use a mental map to think out a new route to the food? [MUSIC] Yes, and he does it the very first time. The maze demonstrates that laboratory rats are capable of more than trial and error learning. >> But let's step back to a broader question. Does behaviors in general provide good explanations for human action? And Skinner made the case for it in his classic and quite wonderful book, Verbal Behavior. But Verbal Behavior was subject to a book review by the linguist and public intellectual Noam Chomsky. And this book review became actually the most, I think, important book review in human intellectual history. I think it's fair to say that Chomsky's critiques in this book review shattered the foundations of behaviorism. Now in the course of Chomsky's review, it's not like he brought up scientific evidence and scientific arguments against Skinner's claim. Ultimately Chomsky championed a certain view on language development that was very nativist and very opposed to Skinner. But this isn't what he talked about in his review. In his review, he analyzed Skinner's claims in their own right. In particular, his claim that human behavior is shaped by reinforcement and punishment. What he ended up arguing is, that at least when it comes to people, let's put aside animals, his focus wasn't animals. At least when it comes to people, these notions are either flat wrong or so vague as to be unfalsifiable and uninteresting. So Chomsky points out for instance that Skinner raises questions like why do people talk to ourselves or why do we imitate things sometimes or create art? Give bad news to enemy or fantasize without pleasant situations. These are all good questions. And what Skinner says is for each of these, it's reinforcement. We do this because it's reinforcing. And Chomsky's point is this. It's either false or empty. It's not literally true that these things are reinforcing. In a sense that you get some sort of food, or pat on the head, or gift, or something that gives you pleasure as soon as you do it. When I talk to myself, I don't get any prize for it, I just like talking to myself sometimes. When I fantasize, the fantasy itself is its own reward, but there's no reinforcement. Now, what Skinner could say is that you do get reinforcement, it's just more abstract, it's rewarding, and that's the reinforcement. But Chomsky's point here is, when you make reinforcement so abstract, you lose any explanatory power. You just say we like it because we like it. So if Skinner is forced to say, as Chomsky argues he does, that creating art is reinforcing because we like creating art, that just comes down to the claim we create art because we like creating art, which is true, but psychologically uninteresting. And so it may well be that Skinner's theory of punishment and reinforcement, when made really explicit, in a way we did when we talked about training a pig or something, is a wonderful explanation of animal learning and the training of animals. But it seems to fall short when explaining why humans do what we do. And you can see this in other ways as well. So move beyond the Chomsky critique. It's not actually true in some sense that our behaviors are the product of law of effect. Because sometimes, we actually come to insights. If I was stuck in the puzzle box, I wouldn't flail away and do 100 things. Rather, I would think about it. I'd model it in my head, figure out what works and act upon it. So where does this leave us? Well the legacy of behaviorism seems to be this. It has given us a richer understanding of some very important learning mechanisms. There may be much more in our head, there may be innate knowledge, there may be insight, but I can't imagine anybody doubting that we have the capacity for habituation, classical conditioning and operant conditioning. And then this can explain certain really important aspects of human thought and human behavior. In addition to this, behaviorism and particularly the work of Skinner, has given as powerful tools for training and for teaching for people including children but also for non-verbal creatures, and has been an important contribution to the world in that. But where it goes wrong, I think, is that it underestimates the scope and power of human mental life. If the problem of Freud is that his theories were too all encompassing and too vague and too ungrounded in empirical effect to ultimately become a successful theory of the mind. The problem with Skinner is that when taken specifically, his theories just fail to explain the richness of human mental life. We'll talk about all sorts of things about the mind that the ideas of behaviorism are simply not sufficient to address. But I'll end by pointing out that, just like Freud, Skinner's ideas live on, and deservedly so. There's a lot to behaviorism that even now the most nativist, the most cognitivist psychologist still has to contend with.
