Sunday, 19 January 2014

Has radical behaviorism reached retirement age?

Simon Baron-Cohen

The is an online magazine of sorts which has 'conversations' instead of 'articles' that are written by some of the leading minds and academics of our time. To kick off the New Year they decided to gather a range of responses from scientists and science educators on the topic of: "What scientific idea is ready for retirement?". This is, in my opinion, quite an interesting topic to discuss and there were some thought-provoking answers. There were also some questionable ones, like Sam Harris' suggestion that we need to expand the definition of science which can be summed up by him saying he doesn't understand why his ideas on morality and free will are wrong - but that's an issue for another time.

The response that I'm interested in is the one written by Simon Baron-Cohen where he answers: Radical Behaviorism. I'll be honest, when I was first linked to this article I thought it was a very risky approach from him. His other parodies and caricatures relied on the audience having very little in the way of formal education and I thought Borat was a much funnier character than a mad scientist who attempts to criticise an idea he doesn't understand.

But then I realised I had confused Sacha with his cousin Simon. What I thought was an elaborate yet ultimately unconvincing portrayal of a bad scientist turned out to be an actual psychologist making silly claims. But why are the claims silly? Let's review:

"Every student of psychology is taught that Radical Behaviorism was displaced by the cognitive revolution, because it was deeply flawed scientifically."

Damn. First sentence out of the starting gate and immediately we run into a hurdle. I'll ignore for now the claim that the cognitive "revolution" took place and displaced radical behaviorism and focus instead on the idea that radical behaviorism is "deeply flawed scientifically". How is it possible for a philosophy of science to be scientifically flawed? By its very nature it is not empirical or scientifically testable.

But maybe he'll present good reasons below.

"Yet it is still practiced in animal behavior modification, and even in some areas of contemporary human clinical psychology."

Radical behaviorism underpins the experimental approach of psychology. Even if we grant him the idea that it has been displaced by cognitivism, we are still forced to accept that cognitivism did not replace the behaviorist methodological paradigm - it simply built upon it. So all of psychology, to some degree, "practices" radical behaviorism.

"The central idea of Radical Behaviorism—that all behavior can be explained as the result of learned associations between a stimulus and a response, reinforced or extinguished through reward and/or punishment—stems from the early 20th century psychologists B.F. Skinner (at Harvard) and John B. Watson (at John Hopkins)."

...That's not what radical behaviorism claimed at all. And suggesting that radical behaviorism "stemmed" from Skinner and Watson seems to imply they both were radical behaviorists or had equal hands in its development, when of course we know that John Watson was a methodological behaviorist. Radical behaviorism stemmed from methodological behaviorism only in the sense that it was a reaction to it.

Baron-Cohen even later seems to have a moment of self-awareness when he comments on the fact that the sentence bolded above is a strawman, by saying: "Skinner was painted as if he believed the newborn human mind was no more than a blank slate, although this was something of a straw man, since in at least one interview Skinner clearly acknowledged the role of genetics". So how is this possible? How can a person simultaneously believe that all behavior is the result of learned associations and accept that behavior can also be a product of genetics?

The answer is: you can't. If you accept the latter then, by definition, you reject the former. And that's what Skinner did (I've commented on some of these issues in more detail in my article "Misunderstanding behaviorism" so I won't rehash the arguments here).

So what is his actual reason for wanting to retire radical behaviorism? He states here:

"[It] is scientifically uninformative. Behavior by definition is the surface level, so it follows that the same piece of behavior could be the result of different underlying cognitive strategies, different underlying neural systems, and even different underlying causal pathways. Two individuals can show the same behavior but can have arrived at it through very different underlying causal routes".

Well, of course. But no behaviorist would deny this. Back when I was in Behavioral Psychology 101 they taught us this very concept by detailing this situation: suppose you see two men running down the street, what is the cause of their behavior? We don't know because topographically similar behaviors can have distinctly different causes - one man might be running because he's late for work, the other because he is on fire.

Baron-Cohen goes further and explains that the problem is that radical behaviorism ignores the underlying cognitive patterns and neurobiology. Oh, okay then. But wait a minute, the whole point of radical behaviorism is that it accepts the causal role of cognition. And the importance of neurobiology to understanding behavior is discussed, and researched, by a number of behaviorists today and was even commented on by Skinner himself!

It's almost as if he's attacking the strawman he set out to warn us about earlier...

"Given these scientific arguments, you'd have thought Radical Behaviorism would have been retired long ago, and yet it continues to be the basis of 'behavior modification' programs, in which a trainer aims to shape another person's or an animal's behavior, rewarding them for producing surface behavior whilst ignoring their underlying evolved neurocognitive make-up".

This is probably one of the biggest clues as to where he's going wrong: "surface behavior". He seems to be arguing that behavioral modification techniques (we're no longer talking about radical behaviorism it seems) only mask problem behaviors rather than actually fixing them. His evidence for this? Non-existent. His entire argument seems to be predicated on the idea that since 'behaviorism' has 'behavior' in its name, then it must only be concerned with external behaviors. This obviously ignores the major contribution that radical behaviorism brought to psychology: the notions of covert and private behaviors.

When combined with the myth that radical behaviorism ignores cognition we could easily understand why Baron-Cohen is so confused that radical behaviorism has lasted so long. The reason why, of course, is that it doesn't claim the things he says it does.

"The orca dutifully produced the behaviors to get the rewards (food) but, over the years in captivity, he was involved in 3 deaths of people. It has never been documented that orcas have killed a human in the wild, so this may have been a reaction to the Radical Behaviorists who were training this orca to show new behaviors, whilst ignoring millions of years of evolved social and emotional neurocognitive circuitry in the animal's brain, circuitry that does not just vanish in captivity".

The actual argument here seems to be quite difficult to discern so I'll break it into two basic components: 1) things associated with captivity caused the deaths of humans, and 2) new behaviors are incompatible with [insert neurobabble].

To the first point, this has to be a misunderstanding on my part. Orca typically live in areas that aren't frequented by human swimmers whereas those in captivity are exposed to human swimmers on a daily basis for hours at a time. Of course there are going to be more attacks in captivity than in the wild, in the same way that there are going to be more deaths in Japanese made cars than classic hot rods - there are simply more of them on the road.

The second point is simply nonsensical. Why would novel behaviors produce aggressive behaviors on the basis that they are "new"? If an organism had a breakdown every time a "new" behavior "ignored millions of years of evolved social and emotional neurocognitive circuitry in the animal's brain" then we'd have serious problems on our hands - in captivity and the wild.

More importantly, the fact that novel behaviors can be incompatible with instinctive behaviors is an idea formally proposed by behaviorists! In their paper "The Misbehavior of Organisms", Breland and Breland discuss some of the problems associated with what they termed 'instinctive drift'. The long and short of it is that when they were training pigs to collect novelty sized coins to deposit in a piggy bank for a bank commercial they noted that after a while the pigs would stop depositing the money. Initially the coins had acted as a token reward whereupon depositing the coin resulted in food reinforcement but soon they started treated the coins as they would actual food, burying them in the mud to return to later.

"Kidnapping one individual orca and placing him or her in captivity not only isolates the animal from their social community, but it reduces their life expectancy, and causes signs of ill-health, such as the frequent collapse of the dorsal fin. The use of Radical Behaviourism towards such animals in captivity is doubly unethical, because of its lack of respect for the animal's real nature. The focus on shaping surface behavior ignores who or what the animal really is".

This clearly has nothing to do with radical behaviorism. With or without the philosophy of science underpinning psychology this behavior would undoubtedly continue because there is a demand for it and somebody can profit from it. Orca aren't being put in Sea World and taught to do tricks because of radical behaviorism, people have been setting up circuses and training animals to behave in specific ways for much longer than radical behaviorism has been around.

This is like criticising evolution on the basis that it was unethical for Hitler to round up entire sections of society to kill them off in the hopes that the new generation will be more aesthetically pleasing.

The summary:

"There may be ethical lessons here when we think about the still widespread use of behavior modification of humans in contemporary clinical settings: the need to respect how a person thinks and feels, respecting their real nature, rather than simply focusing on whether they can be trained to change their surface behavior".

I agree. An ethical form of clinical psychology would respect how a person thinks and feels, respect their "real nature", and focusing more on just "surface behavior" to mask problem behavior. But, of course, this is exactly what psychology informed by radical behaviorism does.


  1. I won't consider Watson's behaviorism as methodological. Perhaps Hull but not Watson.

    Otherwise I basically agree with the article.

    1. Thanks for the comment, Jorge, but I'm interested in why you disagree with the characterisation of Watson as a methodological behaviorist. How are you defining it and why doesn't Watson meet the criteria?

      Methodological behaviorism is simply the idea that since we have no way of accessing inner states, we should just operate as if they don't exist. Hence the "methodological" qualifier, which means that there is not a rejection of the existence of something but rather there is simply a pragmatic understanding that it's best to work as if it didn't exist (in the same way that it's used in "methodological naturalism" to refer to the assumption of naturalism rather than the assertion that naturalism is true).

      Watson, as far as I can see, meets such a definition and is the archetypal methodological behaviorist. Hull on the other hand was a neobehaviorist as he rejected Watson's argument that we should ignore inner states and instead posited that unobservable processes can be inferred (as they are in other sciences) with the help of operational definitions.

      But, with all that said, behaviorism does have a bit of a PR problem in the sense that many different authors have mixed and redefined terminology so I wouldn't be surprised if you'd read resources that define things differently to how I'm using them.

  2. Thanks Mike. I'm a behavior analyst and agree with your analysis including that of Watson acting methodologically. This field is beginning to realize when comments like these are made that a response is important to prevent misbelief by omission. However, many clinicians lack this self-promotion skill. We are getting better, but thanks for the help. --Martin Ivancic

    1. Thanks for the support, Martin. I agree that articles like that require rebuttal, otherwise we risk repeating the mistakes of the past (specifically thinking of the delay in responding to Chomsky's review of "Verbal Behavior").

      I'd love to say that my attempt at a rebuttal did our position justice but I read another response here that I think does a far better job than I did!

  3. Thanks for the rebuttal. Wish The would reprint it, but that is not likely to happen. I do have queries about two of your statements: "Radical behaviorism underpins the experimental approach of psychology." "...the whole point of radical behaviorism is that it accepts the causal role of cognition." Regarding the first point, the scientific method derived from radical behaviorism has tended to be single-case experimental designs, whereas experimental psychology has a history of between-group designs. Granted early work in sensation and perception used uncontrolled single-case methods,but most experimental psychologists love those control groups. Regarding the second point, while radical behaviorists recognize cognitions as covert behaviors, I do not see where they say that cognitions are causal of other (external) behaviors. This issue of the role of cognitions is one of the fundamental differences between radical behaviorism and cognitive psychology..

    1. Hi Maurice, thanks for the comment.

      On the first point, what I'm talking about there is more than just the actual methodology and I suppose I'm meaning something more like a paradigm. The behaviorist ideas on how to approach the subject matter of psychology has been integral to our success as an experimental science; for example, treating verbal reports and introspection as behaviors that are subject to their own contingencies besides the underpinning mental states that they are supposed to represent. I think this is true of a number of contributions behaviorism made to the field, like the questioning of anthropomorphic and mentalistic explanations, and the use of explanatory fictions.

      On the second point, Skinner often discussed cognition as being part of the causal chain involved in the development of specific behaviors. His argument, which has often been [mis]represented as an argument against the existence or utility of cognition, was that cognitions are not causes in themselves. That is, they are not self-causes. Whilst Skinner accepted the possibility of studying cognitions as explanations for behavior, he just argued that it makes more sense (logically and pragmatically) to take one step back in the causal chain to analyse and manipulate the environment.

      What cognitivism did then, at least in my opinion, was to extend upon the ideas of radical behaviorism, rather than reject it. In other words, they argued that (at least in some cases) it makes more sense logically and pragmatically to analyse the cognitive step of the causal chain to understand how a behavior came about. But as evidenced by the development of CBT and the fact that a number of cognitive psychologists are behaviorists, there is no inconsistency or incompatibility of being a behaviorist and accepting the causal role of cognition.

      There's an interesting discussion on the history of private events in radical behaviorism by William Baum here, although he does go on to reject the distinction under his form of molar behaviorism.

  4. Turned to google after reading the edge piece and was thankful to find your rebuttal. After a number of years in education, I began working in ABA and am just beginning my BCBA coursework. A long-time believer, however in determinism and the question of human will, I was absolutely floored when I began to read Skinner's About Behaviorism. I had had the usual cursory experience with him in undergraduate psychology courses. Yet in my readings far and wide over the years trying to better get at social justice through understanding child and human development, I don't think I'd heard once about how profound the implications of radical behaviorism are, not to mention how hard the science is behind them.

    Anyhow, my background isn't psychology, and so as I study ABA I keep coming up against the "cognitive schism", if you will. Yet I'm continually struck by the seeming misunderstandings people have, and some of the almost vitriolic hostility expressed towards what I find is a beautiful and rich - and as clear a framework for understanding human behavior as I've come across. The edge article was a perfect example of the... weirdness people have in understanding behaviorism. When he was going on about Orcas, I mean... I make progress with the kids I work with on a daily basis that could bring one to tears it is so profound and life-altering for them and their family, and it is a DIRECT result of verbal behavior! And this guy is at Cambridge?!!!

    1. Hi Vidoqo, thanks for your comment!

      I agree, the misunderstanding of behaviorism here was so absurd to the greatest degree. If you haven't seen it already, I have another article on a similar topic here: The Law of Feynman.

      Your experience with psychology undergrad is unfortunately common. I didn't really hear much about Skinner or behavioral psychology until one day I was sitting in an evolutionary psychology tutorial group and my tutor was making jokes about how ridiculous behaviorism was as a philosophy of psychology. It sounded so horrifically absurd that I had to look it up but, as you'd expect, the reality of the situation didn't match up his description.

    2. An example of what is still taught

    3. An example of what is still taught

  5. Mike, well-written responses to Baron-Cohen's critique.

  6. This comment has been removed by the author.