All it takes to be a Behaviorist
All it takes to be a Behaviorist is to believe that your questions about psychology can be answered --- Really and Truly answered. When someone asks a question like "Why did he decided to climb the mountain?", all you have to do is believe that the question references no magical or mystical entities. You don't have to be certain you know what "decide" means, you just have to believe that, whatever it means to "decide", the person asking the question observed it, or at least that it was possible for them to have observed it. That is, she is asking why he "decided", because "deciding" is a non-magical thing that people do in the world, under certain circumstances. As such, the decision can be explained through reference to prior events on multiple time scales, and at multiple levels of analysis, just like any other observable happening.
To be a touch more specific: All it takes to be a Behaviorist is to believe that questions about psychology are properly understood as questions about behavior. Questions about psychology are questions about why one or more people did particular things in a particular set of circumstances. All questions about psychology.
Oh, I know what you are thinking, "But what about conditioning (Pavlov, Skinner), or logical derivations of psychological laws (Hull), or dynamic systems (modern Embodied Cognition), etc., etc., etc.?!?"
Those things are great. Lots of discoveries were made by people who happen to also be behaviorists, including lots of general principles about the development and organization of behavior. But you can fully believe that operant conditioning works, for example, without believing that the behaviorist approach to psychology is correct. Ditto almost any of the other propositions advanced by individual psychologists in the behaviorist tradition.
The behaviorist position is the logical culmination of the rejection of magical entities and dualisms. Charles Sanders Peirce asserted that you have not had a complete thought until you have considered all the consequences of that thought. Behaviorism is, I would assert, where you end up if you think clearly when rejecting a non-physical soul as a force that causes behavior. This way of thinking does not deny that people do mental things, but it denies any attempt to assert that mental entities are separate things, that have an existence outside the context of actions happening in a world. There are, of course, psychologists who deny the existence of some (or all) mental things, but that is a position held separately from their being a behaviorist. There are also many ways in which individuals can understand the mapping of questions about psychology onto questions about behavior. Skinner, for example, attempted to map many (but not all) questions about psychology onto questions about the developmental history of the organism.
In addition to trying to clarify and promote Behaviorism, I am a proponent of a particular approach to behaviorism that has its origins in American Philosophy. There is an intellectual trajectory that I believe is behind the early popularity of Behaviorism in the United States, which flows from Charles Sanders Peirce to William James and to the cohort of psychologists whose education was steeped in James's work. Following James's death, those working to advance this trajectory were on the verge of bringing forward something great. Then Watson came along and sidetracked everything. Reading most current histories, it would be easy to think behaviorism arose from Watson like Athena from the head of Zeus, but there is ample evidence to the contrary. While it is easy now to read into William James anything you might want to see, the connection between James's work and the emerging behaviorism is made clear in the works of Edwin Bissell Holt, who was James's student, colleague, and friend. The influence of this trajectory can be seen in the work of Holt's students, including the "Purposive Behaviorism" of E. C. Tolman and the "Ecological Psychology" James J. Gibson and others. Building off of James's work, Holt tells us that he is pursuing:
In this way of thinking no mental events are denied: Mental events are understood as things people do, not things people have. Mental verbs are embraced, but mental nouns are viewed with suspicion. It makes sense to talk about a person as being conscious of particular aspects of his or her surroundings, but it makes less sense to talk about a person as "having consciousness" or of having certain objects "inside" consciousness. Similarly, it does not make sense to talk about a person having "a memory", stored somewhere, like a series of 1s and 0s on a disk. Rather people sometimes have the ability to behave as a reliable function of events in the past. Performing such behaviors is re-membering.“a way of thinking which aims to escape, both in philosophy and in psychology, from the absurdities of subjectivism and any form of psych-physical parallelism… ‘consciousness,’ the metaphysical entity, does not exist; that it is merely the last lingering echo of the primitive ghost-soul. Conscious phenomena of course exist, [James] said, and the problem of cognition exists, but not mental substance…”
An intellectual great-grandchild of Holt is Nicholas Thompson. Nick has spent the past 50 years or more developing what he calls the "Natural Design" perspective. Without going into to much detail, this approach claims that mental terms refer to certain patterns that can exist between arrays of behaviors and arrays of circumstance. I don't know if Nick's system can work for all psychological questions, but do believe it can greatly clarify some psychological questions, and it can help to clarify the relationship between biology and psychology. Given that my original training was in animal behavior, this is a major interest of mine.
In conclusion, much of my current efforts are aimed at 1) elucidating what it means to be A Behaviorist, 2) exploring the trajectory from pragmatism to radical empiricism to radical behaviorism to a variety of mid-20th century movements to modern embodied cognition, and 3) the potential of Thompson's system to clarify issues in the field.
"Mental events are understood as things people do, not things people have. Mental verbs are embraced, but mental nouns are viewed with suspicion."
ReplyDeleteOK, but granting the premise that mental events are actions (which no cognitivist actually disagrees with) does not delegitimize the use of "mental nouns" to descriptively capture particular properties of those events. We use static short-hand descriptions of action all the time. If you described to me a road trip where you drove (a verb) from DC to Tacoma, should I be similarly suspicious about any references you made to your "route" (a noun)?
Good question!
ReplyDeleteThe classic examples are about consciousness and memory.
The first seems very antiquated, but I think it is still relevant: The traditional assertion that the experienced world is "in" your consciousness, is alive and well in the modern assertion that the experienced world is "in" your brain. The Late-William-James / E. B. Holt position is that the experienced world is the world. There are surely additional details to be worked out, but that is the assertion. Thus claims like "Brain part X is, in humans, typically plays an important role in conscious processes" is a bit vacuous, but might be acceptable. In contrast, claims like "Brain part X is where consciousness is" are flat out a priori.
The problem with memory is more obvious in the messy animal literature. The quick version, which you know well, is: Just because an animal's current behavior has altered as a result of something in the past does not mean the animal has "a memory" of the past event. I don't deny that sometimes people re-member things and I'm not completely sure how best to describe such situations from this perspective (though I like Tonneau's solution at the moment). I am, however, sure that I sit in on talks all the time in which people talk about rat's memories, as if such things were tucked into the rats brain, with the implication that these memories are rexperienced, revised, reconsolidated, etc., where I think they little evidence to support that view.
I'm not sure how exactly to map these examples to your analogy of the trip. In English we let verbs change to nouns and vice versa pretty often without trouble, but for some reason it seems to cause a lot of trouble in psychology. When you returned from a trip I could say, for example, "did you have a nice drive?" and you might say "yes" or "no". If I followed up with "where did you put it, and can I have it next?" you would roll your eyes and say "har, har", because we all know full well that though I refereed to "your drive" as a noun, I didn't really mean that it was "thing" in any traditional sense.
On the other hand, if someone told me she had just interacted with a man who "had a narcissistic personality", and I asked her where he kept it, there could well be an elaborate conversation indicating that he kept in his brain. We might agree on many methods we could use to try to find it there. At a conference I could put up a colorful slide with a region of interest, say "we have long wondered where narcissistic personality was in the brain... and I have found it" with applause following thereafter. Not only is this bad thinking, but it leads to a weirdly narrow path of investigation. If I was convinced that the researches and consumers of the research were treating this, as some form of short hand, then I don't think it would bother me so much. All evidence suggests that many researchers, and certainly most consumers of psychological research, really believe such things when they are said.
Recall Skinner's critique of hypothetical constructs in psychology. They are perfectly good things to use, he said, except that the next generation of psychologists always seems to forget they are hypothetical.
"Recall Skinner's critique of hypothetical constructs in psychology. They are perfectly good things to use, he said, except that the next generation of psychologists always seems to forget they are hypothetical."
DeleteI think Gilbert Ryle also extended on this idea in "The Concept of Mind", where he basically argues that it's accurate to reject the 'mental nouns' as things in themselves but it makes sense to retain the terminology as shorthand because it makes it easier for us.
I have only stumbled upon your blog recently, but as a grad student who is still learning the ecological and embodied perspectives, I find it a great resource.
ReplyDeleteIn reply to your comment, I would ask how you would interpret the findings of a study Andy Clark describes in "Supersizing the Mind" where individuals must program a VCR. In the study, individuals either have free access to a panel of reference information, or they have to press a button to access it. In this study, people seem to use different strategies to solve the VCR problem--either referring directly to the information panel, or storing the information from the panel and then going back to the programming task. What is happening when the information is being stored? At some level, couldn't we say that some pattern of brain activity represents that stored informatoin?
Uhg... I just wrote an elaborate reply that was deleted somehow. I will try to recreate it quickly, but apologies for any lack of eloquence.
ReplyDelete--------------
Alex, many thanks for the question. Another tough one!
I don't have a firm answer, because are many options represented by those I generally agree with. Andrew and Sabrina would claim you should not talk about representations at all. A bit on faith, they would claim that if you had a proper handle on the task, you would find that you did not need such talk at all, i.e., such talk is indicative of poor experimental control. Tony Chemero argued pretty convincingly that while you often can use representation-talk in psychological contexts, such talk is much less informative than a good model of what is happening, i.e., the terms just don't deliver in the ways people think they do. Dan Hutto argued pretty convincingly that "basic minds" do not need representations at, but I would bet he would say this situation does involve representations, and would have no problems so long as you used the term carefully.
So, with that out of the way, I probably still owe you a bit of hand waving as to what I think.
First, I think we could agree on the following: There is a mystery about how people can respond at one point in time based on something they read at a prior point in time. A complete description of how that happens will involve a lot of details about changes in the brain.
I will add to that a suspicion that we probably do not share: I suspect that if we had a complete description of what was going on in the brain, we would agree that it does not look much like anything is being stored.
From that, I further conclude: If nothing is being stored, it cannot possibly be the case that "representations" are being stored.
I know that is probably not a totally satisfying answer, but hopefully it at least clarifies the basic tenants being asserted.
Hi Eric -
DeleteThanks for this post. I have trouble deciding from web discussions what positions are considered to be "Behaviorist" and what aren't - this helps. I like your rather inclusive definition since it lets us avoid that question.
There is a mystery about how people can respond at one point in time based on something they read at a prior point in time. A complete description of how that happens will involve a lot of details about changes in the brain.
I don't know why this would be any more mysterious than other types of memory, if you accept a couple of reasonable (IMO) assumptions:
1. The objective of an imperative sentence is either to create in the hearer/reader a new behavioral disposition or to activate an existing one.
2. The essence of a memory is having a learned disposition to respond to members of a family of sensory stimuli in a certain way.
The text of an instruction manual can be viewed as a sequence of imperatives. Per assumption 1, each individual instruction is intended to create a disposition to execute an action in a certain context. For example, imagine that you're reading the entry in an instruction manual that describes turning on a simple appliance. There will often be a drawing of a control panel and a description of the action to be taken, say, "Press the red button labeled ON shown in the drawing below". The text of the instruction will be recognized as specifying an action to be executed once you are viewing the control panel and have located the button. When viewing the control panel, you experience a neural activity pattern that should adequately match the one experienced and learned when viewing the drawing. The imperative "push a button" is a familiar action that your brain has been trained to execute, so the motor neuron excitations necessary for that execution can become associated with that learned visual neural activity pattern. Per assumption 2, the combination of the "stored" neural activity pattern and the "stored" motor neuron excitation commands constitute a memory module. (See note below.)
There are indeed many details beneath this top level description, but the basic architecture seems plausible, and it allows us to address the question of representations a bit more directly. In principle, the neuronal activity pattern to which the memory module responds could be converted back into a rough sketch of the drawing or of the panel, so one could call it a representation of either or both of them. But since operation of the module can be completely described without that interpretation, I don't see what it would buy us. OTOH, the possibility (likelihood?) of that interpretation's causing unnecessary confusion seems clear.
I suspect that if we had a complete description of what was going on in the brain, we would agree that it does not look much like anything is being stored.
I'm not sure how to interpret "much [of] anything" and "stored" here. Surely memory involves some neuronal structures (whether or not anything like the modules hypothesized above), so in that sense something is "stored". But if you mean "much of anything that could profitably be called a 'representation'", then I agree.
Note: In "How to Create a Mind", Kurzweil describes in considerable depth modules that he calls "pattern recognizers" the architecture of which, as best I recall, is somewhat along these lines.
Thanks Charles!
Delete"In principle, the neuronal activity pattern to which the memory module responds could be converted back into a rough sketch of the drawing or of the panel, so one could call it a representation of either or both of them. But since operation of the module can be completely described without that interpretation, I don't see what it would buy us. OTOH, the possibility (likelihood?) of that interpretation's causing unnecessary confusion seems clear."
If I am understanding you correctly, this is the position Anthony Chemero advocates for in Radical Embodied Cognitive Science.
About the storage thing... If agree that the brain isn't storing representations, or agree that isn't a very useful way to talk, that is at least 75% of the issue out of the way.
As for the rest... I think we need a clear model for what "storage" means. If I store vegetables for the winter, they are in a jar, in my cupboard, and I can go to the jar, open it up, and get them right back out. Assuming I prepared them well for storage, they will be fairly identical upon retrieval, even over a long period of time. The memory-storage talk came into psychology, so far as I can tell, with the dualistic assertion that there was a VERY close analogy between that situation and the way memories existed in the other-worldly realm of the mind. When psychologists started looking for memory traces, they were trying to do a quick substitution between "other-worldly realm of the mind" and "structure of the brain". Of course, I can't rule out that one day we suddenly get that analogy to work, but so far 100 years of looking hasn't done so.
Neuronal structures are amongst the many things that change in the body as a result of experience. I'm not sure neuronal changes should be inherently privileged over the others, but maybe that is a distraction. Sticking to the brain: One characteristic of the ultra-complex brain, is that it can maintain stable behavioral responses despite dramatic changes in underlying structure. Thus, it seems unlikely to me that, were we ever to get a proper handle on what the brain is doing, we would find the "storage" analogy useful.
To hazard a metaphor I'm not quite sure works: When I flip a switch to turn on a life, the structure of the electrical system in my house has changed. However, it would be weird to point at the switch and say that the switch's current position "stored" my having moved it to that position. It would be even weirder to call the switch a "memory module."
I think we are really in agreement re "storing". I was concerned that in saying something like "nothing is stored" you might be suggesting that there was no brain change involved. Now that I see you weren't, I'll happily agree that expressions like "a memory is stored in the brain" are harmful, suggesting as they do that a "memory" refers to some entity that exists independently of brain structure and can be put into the brain container in the same sense that peas are stored in a jar..
DeleteGreat post, although I fear that phrasing the behaviorist position like this, "questions about psychology are properly understood as questions about behavior" will inevitably lead to the common (mistaken) objection of how psychological phenomena aren't external behaviors.
ReplyDeleteObviously 'behavior' here includes cognitive processes, but have you run into people trying to refute behaviorism on the basis of the "Super Actor" argument before and, if so, what do you think is the best way to respond? Would it help if we stopped referring to it as 'behavior' and used a different term?
Thanks Mike!
ReplyDelete"Obviously 'behavior' here includes cognitive processes"
I'm not sure about this, assuming that by "cognitive processes" you mean something like "brain processes". This is one of the big tensions in the James/Holt/Thompson trajectory. I suspect James would say that the question is not very interesting or important; I'm suspect that Holt would allow just-brain occurrences, if you phrased things carefully; Thompson would probably say that brain processes on their own don't count (for explanatory purposes), because that move entails slipping between levels of analysis. Of course, there are many other positions, but even within this single intellectual trajectory, I think we have covered the full range of options pretty well... So I guess I would like to stick with the quote you pulled out, and insist that the minimal criterion for being a behaviorist is neutral with regards to those different options.
Also... due in no small part to reading too much Orwell, I believe in fighting for words. Thus, at this point, I'm not inclined to give up "behavior" or "behaviorist", though I suppose someone might convinced me in the future that it is to do so in the future.
"I'm not sure about this, assuming that by "cognitive processes" you mean something like "brain processes"."
DeleteNot really, I'm referring more to what Skinner would call "covert behavior", which are things we would typically distinguish from overt, external behaviors. So when we talk of behaviors like attending and remembering, etc, we are talking about a set of behaviors which are distinct from pressing a button or waving your arm around. This distinction may not have a fundamental basis in that they are probably controlled by the same processes, have the same effects, and so on, but I think the distinction is useful on a communicative level - in that people don't tend to think of 'remembering' as a behavior.
These processes are "mental" or "cognitive" in that they describe abstract extended patterns of behaviors or responses, rather than physical instances of a behavior. At least, that's how I view it, if that makes sense?
"Thompson would probably say that brain processes on their own don't count (for explanatory purposes), because that move entails slipping between levels of analysis."
This is a great way of putting it and one that I hadn't actually twigged on to before. I would include the qualifier that this would only be true for behaviors which don't have a direct biological basis, like innate responses. So when researchers talk of finding Brain Pattern X that correlates to Behavior Y, Thompson's argument would be entirely correct, as we aren't saying anything new since that brain pattern is just a proxy for the behavior we observe.
But when we get to innate behaviors where a response really does come about due to a fixed structure in their brain, it obviously doesn't make sense to talk of reinforcement histories and choice theories, etc, and the only explanatory variable is the brain. But otherwise I agree with that description.
"Also... due in no small part to reading too much Orwell, I believe in fighting for words. Thus, at this point, I'm not inclined to give up "behavior" or "behaviorist", though I suppose someone might convinced me in the future that it is to do so in the future."
Haha, that's a good approach to have. I'll be honest in that my approach has been tempered over the years because I think people tend to respond better when I correct some misunderstanding of behaviorism. That's why I now tend to describe psychology as: "The study of behavioral and cognitive processes" even though I think 'cognitive processes' are just a subset of behavior, or I discuss radical behaviorism as including the role of cognition in the development of behavior, even though that 'cognition' is still just behavior (just not the external behavior that is imagined by the logical/analytical behaviorists).
Ah yes, covert behavior. Skinner's discussion is pretty close to early Holt (which Skinner was influenced by). Holt tries to talk of it as "muscle tonus", by which he means changes in the muscle system not strong enough to move the body (or countered by other changes in the muscle system so that the body does not move). So for Holt, there needs to be some change that is, in principle measurable, even if we wouldn't normally notice just by looking at someone. Though I think there are many details to work out, I like the covert behavior idea.
DeleteThompson's position is where most of my efforts are going at the moment. He tries very hard to separate the descriptive and explanatory content of terms (in biology and psychology). It leads to some pretty cool places.
Emailed to me from Lee Rudolph:
ReplyDeleteYour blog doesn't want to let me comment today (with the "Name/URL" option; I don't subscribe
to any of the other options), so here's what I would have said there:
==begin==
This exchange has prompted me to realize that one way to describe a large part of my current
project/program as how and why to do ontology without nouns. (Eric, you'll be in
receipt of the latest fruits of that project Real Soon Now. I hope.)
==end==
This is a nice article..It is very easy to understand..POS 355 Entire Course
ReplyDeleteApologizes for commenting very late but I found this piece very interesting. I had some questions or perhaps a comment though.
ReplyDeleteYou wrote:
>>>All it takes to be a Behaviorist is to believe that your questions about psychology can be answered --- Really and Truely answered. When someone asks "Why did he decided to climb the mountain?", all you have to do is believe that the question references no magical or mystical entities. You have to believe that, whatever it means to "decide" the person asking the question observed it, or at least that they could have observed it. As such, the decision can be explained through reference to prior events on multiple time scales, and at multiple levels of analysis, just like any other observable happening.<<<
This is an interesting definition of behaviorism but I'm not sure if many cognitivists would area with it because it places them under "behaviorism". There is also the unexplained idea of what it means to "really and truly answer" a question that is asked about psychology.
To me, the difference in cognitivist and behaviorist position is not one that rejects magical/mystical entities but rather what epistemological criterion is used to “really” answer the questions. Both are appropriate based on what each enterprise sets out to do. So in a way, both cognitivists and behaviorists
The best overview of these issues I've seen was by Hayes in "Mentalism, behavior-behavior relations, and a behavior-analytic view of the purposes of science", where he talks about the differences in terms of what counts as an explanation in behavior analysis (prediction and control, inductive abstractive theories) compared to cognitivism (prediction, mentalism, hypothetico-deductive). Now granted he was talking about a specific branch of behaviorism but I think that the real difference between behaviorism and cognitivist lies in either studying behavior directly as the subject matter or using behavior as indicators of mentalistic theories. However, I agree that these two branches of psychology have a common heritage in the scientific method and non-magical approaches but it's still useful to distinguish them as cognitivists and behaviorists do see themselves as different groups.
Imad,
ReplyDeleteThanks for the comment! Please do not hesitate to comment on any of my posts that might interest you.
You know that there are approaches to psychology (I'm thinking humanist, existential, transcendental, etc.) that claim the cognitivists and behaviorists are very similar from their point of view. More to your point, however, I think that most behaviorist would argue that cognitivists are still wed to magical entities. That is, they want the brain to be full of things other than cells and fluids. They have taken the 19th century "faculties of the soul" and turned them into "modules in the brain" without realizing the baggage that comes along for the ride.
Rifting off of your summary of Hayes, I think you make the point when you say:
>>>I think that the real difference between behaviorism and cognitivist lies in either studying behavior directly as the subject matter or using behavior as indicators of mentalistic theories. <<<
So, the cognitivists want the behavior to serve as a proxy for something other than the behavior, and they claim to be REALLY interested in that other thing. The behaviorist (as I define it here) would ask what those other things might be, and be very suspicious of any answer they get. For example, it is fine to say that observing people buying pies in a store is a good proxy for estimating how many pies the store manager should order in the next shipment. However, I am not sure what you would mean if you said that pie buying behavior is a good proxy for estimating which pie people "wanted". Would you say that there is a "want" in the person's soul, or in the person's head, which CAUSES them to behave in the way observed? If so, then most behaviorists would say you are engaging in magical thinking. You have slipped levels of analysis, they would claim, and have inaccurately anthropomorphized the brain.
Of course, the behaviorist position might ultimately be found wanting, but that is the position as I understand it.