A Limited Defense of Mentalism
Our science illuminates behavior-environment relations. It does not appeal to mental processes to explain the delicate dance between behavior and environment. The philosophy of radical behaviorism that grounds our science doesn’t deny the existence of private events but rather views them as dependent variables (behavior) to be explained. Therefore mental states (like peace), feelings (love), and cognitive processes (understanding) are not considered to be causes of other things that individuals do.
Behavior analysts like this approach because it supports a precise, objective science of behavior and thus provides the basis for effective technological applications. By contrast, mentalistic accounts, we’ve been told by Skinner and many others, undermine good science and practice.
I doubt I’m relating anything you didn’t hear in your first introduction to behavior analysis, and I’m not here to propose that for more than a century behaviorists have embraced some catastrophic error in the construction of their science. Still, just for fun, and with an eye to thoughtful musing about mentalistic explanations, read the following account of how a bird, a Cooper’s hawk (Astur cooperii), incorporated human technology into its hunting.
One Brilliant Bird
By way of background: The Cooper’s hawk is a North American raptor frequenting wooded areas and feeding mainly on birds and small mammals. It’s speedy on the wing, but prey are speedy too, so Cooper’s hawks hunt by ambush, typically bursting forth at the last minute from behind cover – a stand of trees, for instance — to surprise its victims.
In The Cooper’s Hawk Incident we’re discussing, the cover was not some natural feature but a line of automobiles backed up at a pedestrian crossing signal. Here’s how Smithsonian magazine summarized The Incident:
Whenever a human pushed the pedestrian crossing signal, a long line of cars would back up down the street. The savvy creature then used the vehicles as cover to launch a sneak attack on a group of unsuspecting birds gathered in a nearby home’s front yard…. Someone would push the pedestrian crossing signal, which would activate a loud, rhythmic clicking sound designed to let individuals with visual impairments know it was safe to cross the street. The noise also seemed to serve as a cue for the hawk, which would fly into a nearby tree and perch on a low branch…. As cars began stopping behind the crosswalk, the hawk seemed to bide its time. Once the queue got long enough—usually about ten cars—the bird sprang into action, swooping down and flying just a few feet above the sidewalk along the line of vehicles. Then, once it reached a certain house, it would make a hard, 90-degree turn and veer into the yard across the street, where prey birds were feeding on the ground.
That’s the WHAT of The Cooper’s Hawk Incident. As for the WHY, the author of the original observations, Vladimir Dinets, wrote:
The hawk understood the connection between the [pedestrian crossing] sound and the eventual car queue length. The bird also had to have a good mental map of the place, because when the car queue reached its tree, the raptor could no longer see the place where its prey was and had to get there by memory. (Dinets, 2025, “Street smarts: how a hawk learned to use traffic signals to hunt more successfully“)
To hunt as observed, the hawk had to develop a precise mental map of the street (as the target flock was invisible to the hawk until the last stage of attack), notice the connection between the sound signals and the length of the car queue, and figure out that only longer queues provided cover for the entire approach…. The behavior described here is an impressive feat of intelligence, going a long way to explain the species’ ability to successfully colonize such unusual and dangerous environment as urban landscape. This intelligence was likely pre-existing rather than evolved in the novel environment…. One can only imagine the level of knowledge and understanding of the environment possessed by hawks living in more natural landscapes. (Dinets, 2025, “Street smarts: a remarkable adaptation in a city-wintering raptor”)
A Chicken-Little Philosophy of Science
If you’re a good behaviorist you probably dislike how this explanation relies on understanding, knowledge, memory, intelligence, mental maps, and so forth.
But ask yourself: Why is that a problem for you?
A thoughtful paper by José Burgos and Peter Killeen (“Suing for peace in the war on mentalism”) provides some perspective on what we’ve been taught. According to Burgos and Killeen, standard behaviorist objections to mentalism tend to entangle two thrusts. Thrust 1, that behaviorism supports good science and practice, Burgos and Killeen take as a given. About Thrust 2, the contention that mentalism is broken and harmful, they are more circumspect.
Upon careful inspection, Burgos and Killeen argue, many of the specific criticisms that behaviorists lob against mentalism are, logically speaking, either weak or downright nonsensical (I’ll let you check out the specifics for yourself). If that’s the case, why do we embrace them? Because, say Burgos and Killeen, these criticisms serve a function by promoting a sort of Chicken Little catastrophism. In their view, our standard objections to mentalism are less about the features and functions of mentalism itself and more an attempt to “pump up motivation by fighting against hypothetical enemies.” (see Postscript 1).
Hence we exclaim: Mentalistic explanations subvert science and savage our society! Only behaviorism can save the day!
Such a rally-the-troops approach to the behaviorist “war on mentalism” employs the same rhetorical basic strategy as American conservative politicians’ histrionic allegations that liberal policies have caused children to identify as animals rather than as humans, forcing uber-woke schools to accommodate them by providing litter boxes instead of traditional human bathrooms. It’s the same tactic employed by governments that vilify immigrants or ethnic groups or a neighboring nation to evoke citizens’ fears and mobilize public support for their policies.
The irony, say Burgos and Killeen, is that our dread-enemy narrative is unnecessary, because there are plenty of compelling reasons to pursue a behavioral agenda on its own merits:
What behavior analysts fight for—their positive accomplishments in basic research and applications—provides a stronger, more constructive and fruitful motivation. Behavior analysts are excellent experimentalists, and their research is progressing well.
In other words: Mentalism doesn’t have to be bad in order for behaviorism to be good.
Mentalism as Metaphor
But is mentalism bad? Is it automatically, irreconcilably, irredeemably detrimental to science? Burgos and Killeen don’t weigh in on that directly, but check out an interesting series of papers (e.g., here and here) by Tom Zentall, who has published quite a bit of basic behavioral research. Zentall agrees with the behavioral aversion to treating mental “stuff” as actually responsible for our behavior (e.g., “To say that a stimulus is represented in the brain and that it elicits a response adds an unnecessary internal state that serves little purpose;” Zentall, 2012, p. 322). But it’s equally a mistake, he says, to think of mentalistic constructs as proposals about Real Things:
The goal of behavioral research is to understand behavior, and it can be argued that the careful use of cognitive concepts can help us understand behavior… The argument that cognitive terminology interferes with an understanding of behavior is correct only if one accepts the cognitive term as an explanation of the behavior. If instead one asks how one would demonstrate evidence for such a process in the absence of a simpler account, one can either learn more about the underlying process… or reject the need for cognitive explanation (p. 322).
Here Zentall is asking about the heuristic value of mentalism: Can it, under certain circumstances at least, help scientists be more effective? Does it suggest experiments that might not otherwise have been conducted, and thereby reveal behavior-environment relations that might not otherwise have been discovered?
Edward Tolman considered himself a behaviorist… [but nevertheless] asked if animals were capable of forming a cognitive map (a representation of the environment) by training rats to take one path to a goal and then blocking that path. Tolman reasoned that if the rats took a direct novel path to the goal, it would argue for an emergent property of learning that would be difficult to explain in simple S-R terms…. (p. 322)
Tolman proposed that we ask what behavior would look like if it were guided by mental maps, and design experiments to test for that kind of behavior. The result would not verify the existence of a mental map, but it might reveal interesting behavior-environment relations, including some not readily suggested by other conceptual frameworks (see Postscript 2).
As was Tolman, Zentall is concerned not with whether such constructs as mental maps are real but rather with how the behavior of scientists is affected by proceeding as if they were. He concludes that:
Is there value in positing the existence of a cognitive map? On the one hand, it can lead to the idea that an animal has somewhere in its head an image of its environment that it can consult whenever needed. On the other hand, it can provide a shorthand way to describe the results of a novel path experiment. More important, it can lead us to conduct further research into the conditions under which animals are capable of taking a novel path to arrive at a goal. (p. 322)
So, Is This Sky Falling?
I want to apply a similar lens to those descriptions of The Cooper’s Hawk Incident. How might a reader be affected by those mentalistic accounts mentioned above? To answer that question, we first need to be clear about which organisms whose behavior is of interest. I’m assuming that one key goal here is dissemination impact, i.e., the sharing of science with a general, nonspecialist audience (in this case, people who aren’t experts in behavior science generally or raptor ethology specifically). I don’t have readership data for the Smithsonian article, but Dinet’s report in Frontiers absolutely blew up. By only 11 days post-publication, it had 15,484 views and almost 1000 downloads. Its Altmetric Attention Score was already in the top 5% of all articles ever published and it had been mentioned in at least 71 news reports. Also, according to Altmetric’s sentiment analysis, these mentions were overwhelmingly positive.
Most behavior analysts would gladly impale themselves on a hawk’s pointy talons if doing so would bring that kind of attention to their work.
To be sure, a little interpretive caution is advised here, because if the internet has taught us anything it’s that for a lot of people animals are prime entertainment, and yeah, the Dinet article featured a precocious beastie. But we know that animals per se are not sufficient to arouse public fascination, because scientific studies of animal behavior are published all the time without a whisper of nonspecialist attention (see Postscript 3).
I suggest that, regardless of topic, when general readers are drawn to science information, animal focused or not, it’s because they can understand what’s being asked in the research and what the answer is (and if they can get this far, they might actually learn something). This is where accounts of the Cooper’s Hawk Incident bear close inspection.
Linguistic Soft Skills
You see, it’s not just the facts you share that matter; it’s how you relate them. To wit: Take an important innovation like successful autism treatment, bury it in behavioral bloviation, and you might as well be saying something like, “Low occasion setter salience results in learning conditional stimulus partial reinforcement instead of occasion setting” (title from from a May ’25 JEAB article; Postscript 3). Presentation matters. Just ask Catherine Maurice, now one of ABA’s most effective advocates, who initially reacted very negatively to what she was told about ABA.
The key to recruiting interest: What you say about behavior has to make contact with audience verbal repertoires, and I propose that it’s the very mentalism in tellings of The Cooper’s Hawk Incident that has helps to fuel its popularity.
Look, as Skinner often reminded us, all people grow up in a verbal community that is saturated with mentalistic language. Perhaps this language is an impediment to good science, as behaviorists like to claim, but impediments cut both ways. Clear, specialized scientific language is (at best) hard for nonspecialists to understand. For instance, below are some statements from Dinet and what they may imply from a behavioral standpoint. Your typical reader-on-the-street may think the left column is pretty cool but see the right column is somewhere between uninteresting and unintelligible.
| what dinet said | what a behavior analyst hears |
|---|---|
| “The hawk understood the connection between the [pedestrian crossing] sound and the eventual car queue length.” | Stimulus control resulting from exposure to three-term contingencies |
| “The bird also had to have a good mental map of the place, because when the car queue reached its tree, the raptor could no longer see the place where its prey was and had to get there by memory.” | Chaining |
| “The behavior described here is an impressive feat of intelligence, going a long way to explain the species’ ability to successfully colonize such unusual and dangerous environment as urban landscape.” | Shaping by complex environmental contingencies that produces complicated sequential repertoires |
| “One can only imagine the level of knowledge and understanding of the environment possessed by hawks living in more natural landscapes.” | Generality of principle |
Explaining the Cooper’s Hawk Incident mentalistically meets human organisms where they are. And when you make contact with their verbal repertoires you create interest, and with that a chance that the audience may learn something. Minus that interest, nothing gets taught.
This is something all good clinicians understand. They sometimes take liberties with our scientific language in service of promoting a beneficial client outcome. So, for instance, they may ask a parent not to “reward” tantrums with attention (rather than explaining the technical distinction between rewards and reinforcers). They may tell novice therapists to give clients “reasons to make good decisions” (rather than unpacking the quantitative specifications of the matching law and its implications for the relative frequency of desirable and undesirable behaviors). And so forth. These are not precise ways of speaking, but they may be understandable enough to get the listener engaged productively with a problem in behavior. Look carefully at those accounts of The Cooper’s Hawk Incident, and you’ll find much the same vibe.
I’m not claiming that the authors of those hawk articles mentioned above had precise behavioral accounts in mind, and strategically chose to use looser, mentalistic language as part of some systematic audience shaping process. But the effect is the pretty much the same regardless: Because readers became engaged, a lot of them got to see how incredibly malleable behavior is — in this case, predation behavior that might normally be assumed to be biologically hard wired and unchangeable (“instinctive”).
A lot of the objectionable mentalistic language in those accounts of hawk predation isn’t too hard translate into behavioral terms, and in a very general sense it conveys something absolutely true: that when the conditions of learning are right organisms can acquire some pretty amazing (and “unnatural”) abilities. In a world that constantly underestimates the power of learning (in part because WE fail to sell this point effectively), is that such a terrible thing?
Postscript 1: Enter the Dragon
If behaviorists have been obsessed with being Not Mentalists, as Burgos and Killeen suggest, why has this been the case? It’s said that every intellectual movement has a dragon, that is, a predecessor movement that it claims to have vanquished to make the world a better place. Two such movements helped to set the stage for early behaviorism.
First, in human psychology, Introspectionism was an attempt to catalogue the contents of human consciousness. It colorfully failed to produce reproducable results or principles that could be commonly agreed upon. Second, in animal psychology there was a long tradition of anthropomorphism, that is, the tendency to attribute human-like psychological functioning to nonhumans. This reached a crescendo with the (in)famous Clever Hans episode, in which a horse was widely believed to be capable of counting… only to be eventually revealed, in 1907, as merely responding to subtle cues from its owner. Anthropomorphism is mentalistic to the extent that the observer interprets human psychology mentalistically.
It can be argued that what brought down both Introspectionism and early 20th Century anthropomorphism was not mentalistic tendencies per se but rather a reliance on bad scientific methods. It doesn’t matter what you’re interested in… create crappy experiments and you will mold a sketchy science. In demanding better methods, behaviorism fueled better progress, as even non-behaviorists have acknowledged (see, for instance, an affectionate tribute to modern behaviorism by Henry Roedger, accomplished memory researcher and President of the Association for Psychological Science).
Behaviorists have, for the most part, assumed that better methods automatically rule out mentalistic theory and interpretation, but this view is not shared by everyone, like Roediger, Tom Zentall, and comparative researcher Clive Wynne. They all would say that mentalistic questions can have heuristic value when pursued in the context of squeaky-clean experimental design. I take no strong position on this but, having read fairly extensively in the literatures on human and animal memory, will allow that I have seen some incredibly creative experiments, with striking results, that it’s hard to imagine arising from a classical Skinnerian behavioral framework.
Postscript 2: Mental Maps, Part 1
Here’s a simplified example of the heuristic value of mentalistic thinking in research. Tolman et al. (1946; see here for 1992 reprint) trained rats in the simple runway device shown below at left. The rats entered onto a platform from which a single exit led, after a few turns, to food. The red arrow shows the direction (left) of the first 90 degree turn in the runway. If you take a simple chaining perspective, perhaps what the rats learned in training was a sequence of turns (Left-Right-Right).
The green arrow shows the direction of food from the platform “as the crow flies.” Tolman didn’t say his rats really formed a mental map during training, but if they did, he said, they would understand this spatial relationship.
In a test condition rats were released into the same initial runway and platform, but the path taken during training was now blocked, and the platform now opened to a variety of runways. Of interest: Where rats would go upon backing out of the blocked runway. The red arrow shows where they would go if they simply took a left turn, as per the original sequence of movements. The green arrow shows the runway that corresponded to the as-the-crow-flies location of food during training. Before proceeding, do your own very-behavioral analysis and predict what YOU think the rats did. Then go to Postscript 4 to see the actual results.
Postscript 3: About Popular Attention to Scientific Studies
Most behavior analysis researchers don’t receive the kind of attention Dinet did for his report of The Cooper’s Hawk Incident. For example, as of this writing the four of the five animal studies appearing in the May, 2025, issue of Journal of the Experimental Analysis of Behavior (thus published around the same time as Dinet’s report) had accrued exactly zero altmetric mentions. The fifth had a single mention on social media (X). So, clearly, a focus on animals alone are not enough to rally an audience. To get a sense of why, try interesting your accountant neighbor in those JEAB articles’ titles:
- Contributions of delay, duration, and intensity of shock on rats’ choices involving conflicting-valence consequences
- A quantitative analysis of the effects of target- and alternative-reinforcement rate on resurgence
- Training numerousness to numerosity in the dog (Canis lupus familiaris)
- An analysis of renewal following fading of reinforcer type
- Low occasion setter salience results in learning conditional stimulus partial reinforcement instead of occasion setting
Tough sell! I mean, none of these articles is exactly an “Osprey drops hammerhead shark on disc golf course” classic 😀. And though I’m no expert on popular culture, when it comes to harnessing peoples’ pre-existing interests and conceptual frameworks I’m pretty sure a hawk’s technology-aided sneak attack tops all of these.
You might wonder which JEAB animal-research articles have received a decent dose of popular attention. Since you’ve read this far it should not surprise that 8 of the all-time top 10 JEAB articles, in terms of Altmetric Attention Scores, focus on animals that interest people (goldfish, octopi, sea lions) and/or behavioral topics that are transparently related to the human condition (procrastination, predation, “art appreciation”). It’s not hard to explain what these articles ask or what they find, and all of them can be pitched in terms of mentalistic concepts that lay people think they understand.
| (Rank) article title | POSSIBLE SALES PITCH | attention score | news | blogs | social media |
|---|---|---|---|---|---|
| (1) Pigeons’ discrimination of paintings by Monet and Picasso | art appreciation | 487 | 31 | 12 | 159 |
| (2) Dogs don’t always prefer their owners and can quickly form strong preferences for certain strangers over others | love, fickleness | 252 | 28 | 2 | 19 |
| (4) Procrastination by pigeons: Preference for larger, more delayed work requirements | procrastination | 157 | 17 | 2 | 15 |
| (6) Sea lions and equivalence: Expanding classes by exclusion | knowledge, insight | 142 | 15 | 3 | 0 |
| (7) The execution of planned detours by spider‐eating predators | planning, mental maps | 133 | 10 | 4 | 62 |
| (8) Temporal discrimination learning of operant feeding in goldfish (Carassius auratus) | internal clock | 122 | 13 | 1 | 9 |
| (10) Some observations on an operant in an octopus | intelligence | 108 | 11 | 1 | 3 |
Postscript 4: Mental Maps, Part 2
So, what did Tolman’s rats (Postscript 2) actually do in their single opportunity to choose an alternate path in the test phase? The most commonly chosen runway was that oriented in the as-the-crow-flies direction of food during training (roughly 45 degrees east of north). Nearly 6 in 10 rats chose a runway oriented in this general direction. Almost no rats turned left (red). About a quarter of the rats, however, chose a strict right-turn runway. Wrote Tolman et al. (1934):
It is evident that at least in our experimental situation practice on a specific route, or response sequence, produces in some rats a disposition to take the shortest Euclidean path to the goal [location of food], whenever this path is available and the practiced one is blocked (p. 432)….If the goal location had been recognized merely as the terminus of the original path, or the place of the terminal response in the original response sequence, then our rats would have been helpless on the test trial. The fact that they selected the shortest path indicates that what was learned during the preliminary training was not a mere response sequence…. They learned, instead, a disposition to orient towards the physical location of the goal (p. 433).
Tolman’s interpretation of the results ran along these lines:. If the training created mental maps, those for some rats were quite precise (corresponding to the Euclidian orientation of food from the start position), while those for others were more general (they at least got it that food was to the right of the start position, not the left). Also, if there were mental maps, in most cases during testing they superceded the influence of simple left-right-right chaining that might have been acquired during training [see below].
Now, you don’t have to buy into mental maps to admit these results are interesting, perhaps even counterintuitive. More to the point, you should ask yourself (perhaps guided by the results hypothesis you generated in Postscript 2), whether your strict behavioral training would ever have led you to design an experiment that could show these effects in the first place. You may be capable, having seen the results, of throwing together a persuasive behavioral account of them, but that is is beside the point. The heuristic value of theory concerns its capacity to stimulate new experiments that anticipate and document cool effects. You are most likely to see behavior that appears to be guided by a mental map if you go looking for it. Observing that behavior doesn’t verify the existence of a mental map, of course.
THEME MUSIC A:
What’s So Funny ‘Bout Peace, Love, and Understanding?
THEME MUSIC B:
Ode to the Delicate Dance Between Environment and Behavior
Since our topic is elaborate behavior sequences, check out this video that was shot in a one continuous single-camera take. Pretty impressive.