Could one of the foundations of radical behaviorism be wrong?
That’s the prospect raised by research published in Personality and Social Psychology Bulletin. Hold on, don’t get your britches in too much of a defensive bunch: Neither behavior analysis nor its philosophical touchstone, radical behaviorism, are directly mentioned in the article, and its purpose isn’t to critique a behavioral perspective. But it merits the attention of serious behavior analysts because of its counterintuitive perspective on a concept that is central to the behaviorist position: free will. Read on to find out more.
How would you react if something at the very heart of behavior analysis — something you’ve accepted as self-evident and incontestable since you began your journey in our discipline — was called into question? Would you double down and insist that, no, what you have always believed must be correct? Or, given evidence to the contrary, would you be open to the possibility that some of the things you’ve long embraced might need to be reconsidered? Would tugging on this one thread of doubt possibly affect your confidence in behavior analysis overall?
The Fallacy of Free Will
A foundational idea of the behavioral perspective — in fact probably the most essential ingredient in behaviorism’s scientific and practical successes over the past 100+ years — is a firm commitment to environmental determinism. Behavior analysts think of behavior — in fact define it — in terms of its relation to environmental events. We stubbornly resist other kinds of explanations, such as those appealing to fixed individual differences like inherited personality/intellectual traits (see Note 1), or to inner/mental drivers of outward actions. The classic case of the latter is “free will,” the purported ability to choose what to do, regardless of what the environment may demand.
Note 1: Radical Malleability
For a lovely treatise on the power of assuming behavior is always malleable, regardless of any assumed inherited tendencies, see Pat Friman’s “There is no such thing as a bad boy.”
What both of these views have in common is at least a partial rejection of environmental determinism: They rely on factors hypothesized to reside within an individual that limit how much behavior will adapt to situational factors. And without picking apart either view in detail, we can say this much in rebuttal: Behaviorists’ refusal to see behavior as unchangeable has led to many important breakthroughs, such as effective treatments for the supposedly “untreatable” condition of autism.
To a behaviorist, committing to the notion that behavior can change with environmental circumstances means embracing that under most conditions it must, i.e., that behavior only occurs, and changes, as a function of the environment. This is something Skinner set out to explain from very early in his career. He was particularly adamant that there are no mysterious, self-initiated inner causes like free will. You can see Skinner working to dismantle the myth of free will in Walden Two (1948) and Science and Human Behavior (1953), and of course his greatest period of public recognition followed the publication of Beyond Freedom and Dignity (1971), a book devoted entirely to defending determinism and critiquing free will.
Skinner’s position on free will was fairly simple: “There is no place … for a self as the true originator or initiator of action,” he wrote in About Behaviorism (p. 248), with the explanation boiling down to this (paraphrased):
| IS FREE WILL REAL? No, because an organism (with or without a guiding “self”) can’t will itself into action any more than a boulder on Mars can decide to spontaneously rocket toward Earth. Free will is a circular concept that muddles cause and effect. In any coherent account, cause and effect need to be separate entities (see Note 2). |
Note 2: Diversity of Opinion (click to expand)
Not every behaviorist has hewn tightly to the Skinnerian line when it comes to explaining the notion of free will — for instance, see here. However, Skinner’s perspective is the most frequently invoked in behavior analysis, and also the one most closely aligned with the philosophy of science applied to various better-established sciences.
This stance hardly originated with Skinner — it’s essentially what scientifically-thinking folks have been saying for as long as there’s been scientific thought. For instance, the 5th Century BCE Greek philosopher Leucippus proposed that the entire universe (including human actions) reflects mundane environmental cause-and-effect in which neither chance nor personal autonomy play any part.
But of course not all theorists have rejected free will. Physicist Arthur Eddington famously argued that quantum mechanics leaves space for free will (though for a rebuttal of that notion, download here). Psychologists have been especially enamored of free will or something like it — think of William James, Edward Tolman, Alfred Adler, Abraham Maslow, Carl Rogers…. Quite obviously fascination with free will is a habit that the rise of behaviorist environmental determinism in the early 20th Century never quite snuffed out.
A Baneful Belief
Widespread fascination with free will also, arguably, was bound up in behaviorism’s later fall from grace as a popular scientific movement. It’s often said that behaviorism, once a dominant force in Psychology, was displaced by a “cognitive revolution.” And it’s true that cognitivism enjoyed a massive surge of interest, circa 1950-1980, that left decreasing intellectual space for behaviorism. But behaviorists, led by Skinner, simultaneously worked to hoist themselves on their own petard. Just as the “cognitive revolution” was coming into its own and making claims about human nature that more or less equate to free will, Beyond Freedom of Dignity earned Skinner the cover of Time magazine, and with it considerable societal and intellectual blowback.
Beyond Freedom and Dignity “quickly become notorious” because, rightly or not, readers saw Orwellian overtones in its deterministic theme. That’s not my theatrical turn of phrase: Noam Chomsky wrote that the book is “widely regarded as pointing the way to 1984” (download here). Public relations like that quickly transformed the public’s opinion of behaviorism. What once seemed a quirky practical movement that gave the world teaching machines and the Air Crib (download) and pigeon-guided missiles and birds playing ping pong now seemed an alarming existential threat. In the void this created cognitive science was happy to swoop in and offer more palatable free-will-laced alternatives. Behaviorism’s public reputation never recovered, as evidenced by the fact that today we’re still fighting the free will battle: For instance, recent critiques of ABA from neurodiversity advocates boil down to the same old concerns about Orwellian threats to individual freedoms.
Which raises the question: Why was Skinner so willing to pick this particular fight so publicly, to risk everything to persuade everyday people that they are not really “free”? The answer (again, paraphrased) is that he saw beliefs about personal autonomy as the root of many of society’s biggest failings:
| WHY IS BELIEF IN FREE WILL HARMFUL? Because this belief provides people with the illusion that they understand behavior, and when people think they understand they stop being curious. All of the world’s great ills are problems of human behavior, and those problems have been allowed to unfold because humans don’t know how behavior works. Belief in free will distracts from finding out. Truly wise humans don’t wait around for other humans to make smart choices (history shows this to be a fool’s errand!). They strategically construct an environment in which wise choices are high probability responses. |
As I’ll explain shortly, this is where the radical behaviorist position might not square with available evidence, but let’s first acknowledge that Skinner meant well and had good reason for adopting his position. He saw a world rife with Big Problems tracing to destructive human behavior (any reasonable contemporary observer will agree that he was right about this much), and he hoped to persuade fellow humans to design societal systems that would promote more constructive behavior.
Overpopulation, the depletion of resources, the pollution of the environment, and the possibility of a nuclear holocaust – these are the not-so-remote consequences of present courses of action. But pointing to consequences is not enough. We must arrange contingencies under which consequences have an effect (Skinner, 1971).
Yet in hindsight this is an argument that was never destined to land. The thing is, if you believe in free will, you don’t think society’s ills trace to flawed social systems (a type of environmental context). No, you blame people’s misfortunes on their poor individual choices, which leaves you disinclined to do anything about Big Problems beyond railing about those terrible people and their bad choices. For example:
Up to 83% of Britons think that ‘virtually everyone’ remains in poverty in Britain not as the result of social misfortune or biological handicap but through choice…. Because of their belief in the fairness of ‘deserved inequalities’, such respondents were discovered to have become almost completely unconcerned with the idea of promoting greater equality while at the same time asserting that Britain was a beacon of fairness that offered opportunities for all….. There was a ‘clear sense’ across all of the groups surveyed that an individual’s situation ‘is largely of his or her own making’. Anyone can make it if they ‘really try’, the great majority of participants asserted. (Miles, 2013, p. 212)
This illustrates the kind of disordered thinking that Skinner had hoped to correct with Beyond Freedom and Dignity, and reminds us that what most concerned Skinner was less whether free will is real, and more what behavioral consequences follow from believing in it. The reason Skinner worked so hard to argue against the concept of free will is that he thought those consequences unfavorable: It was evident to him that when you believe in free will you act in damaging ways. Which brings us to the point of this post.
But… Positive Side Effects of “Free Will”?
Skinner arrived at his conclusions by extrapolating from everyday observations. He did not conduct experiments on human behavior nor did he analyze formal descriptive data about human affairs. In short, he offered no systematic empirical evidence for his assertions about the adverse effects of believing in free will. As I’ve written elsewhere, interpretations like Skinner’s can be surely thought provoking, but they do not and cannot serve as scientific evidence.
By contrast, Baumeister and colleagues (2009, Personality and Social Psychology Bulletin) conducted experiments to address the critical question of What are the consequences of believing in free will? Previously Vohs and Schooler (2008) had reported that undermining people’s belief in free will led them to engage in socially prohibited behavior such as cheating on a test. Baumeister et al. extended this finding by showing that when belief in free will was experimentally reduced people were less likely to help a stranger in need and more likely to aggress toward another person. They concluded that believing in free will (which people in every culture apparently do), might be a functional adaptation that spawns a variety of prosocial behaviors:
We suspect that most cultures will have found beliefs in free will to be socially beneficial and hence will tend to favor and promote those beliefs….The present results have nothing to say about whether free will is an objective reality. All of our findings could be perfectly valid even if free will is a complete illusion…. Our results do indicate, however, that belief in free will contributes to socially beneficial actions and outcomes. Indeed, if it is an illusion, those benefits may well explain why society continues to find it useful to promote it. (p. 268)
As with any empirical report, you should check the methodological details of this one for yourself and decide whether you believe the authors’ claims hold water (as food for thought, here’s one rather rambling critique). But let’s grant that the basic premise of the Baumeister et al. investigation feels pretty behavioral. After all, you’d imagine that a good behaviorist would begin by proposing that:
Belief in free will is nearly ubiquitous in human beings (download evidence).
and
Believing is behavior.
and
Frequently-occurring behavior must serve some practical function.
Further, you’d expect a good behaviorist to follow up these observations with experiments to test the presumed function of the behavior of interest. That’s what we do, right? At our core, we are behavior-function whisperers. And one of the great contributions of behavior science has been to clarify, through relevant experiments, that the function of behavior cannot necessarily be gleaned from its topography. Many different behaviors can have the same cause — for example, tantrums, feces throwing, and the delusional experience of one’s head falling off all can be essentially a plea for attention (i.e., they are reinforced by others’ reactions). And a given behavior topography might have different causes in different contexts — for instance, self-injury might occur because it’s reinforced by attention, because it terminates unpleasant circumstances, or even as a side effect of a mundane medical condition.
In principle, then, there’s nothing “un-behavioral” about proposing that what looks like foolish entanglement with a fallacious idea could in fact serve a useful purpose (see Note 3), even if the underlying mechanisms could be different from what Baumeister and colleagues suggest (see Note 4).
Note 3: Selection (of Belief) by Consequences (click to expand)
How could believing in something that’s not real become a widespread cultural practice? Here it helps to remember that phenomena of the natural world are subject to selection by consequences, and what’s selected is often “good enough” rather than ideal.
For instance, in biology natural selection may favor a characteristic like sickle cell trait that creates both suffering and survival advantages. That trait creates potentially life-threatening symptoms for some but protection against disease to the species as a whole. Another example: the human spine, an evolutionary improvisation that can be rickety and painful but also facilitates speedy locomotion on two feet. Something similar happens in the world of behavioral world of selection by consequences, where a response like tantrumming can be reinforced in the short run but create harm to social relationships in the long run. Sometimes the short term/long term script is flipped: Relying on punishment to control others creates considerable unpleasantness; but punishment also may cement cooperative repertoires that make human interactions more productive and safe in the long run.
The hypothesis forwarded by Baumeister et al., that believing in free is beneficial, regardless of whether free will is real, has ample precedent in cultural anthropology. For instance, as Marvin Harris (1989) wrote in his lovely book Cow, Pigs, Wars, and Witches: The Riddles of Culture, Hindu cow worship is about resource management, not bovine divinity, regardless of what its practitioners may say and think. In the South Asia environments where Hinduism arose, those who believed cows were sacred survived (cows provide protein-rich milk and other resources), while others starved, until cow worshippers were the dominant culture. Amidst this selection pressure it matters not at all whether cows contain one iota of divinity. Sure, treating cows as sacred is a factual mis-step, but selection by consequences doesn’t care about that.
Skinner agreed about this general process of cultural selection:
A [cultural] practice arises as a mutation, and it affects the chances that a group will solve its problems, and if that group survives, the practice survives with it. It has been selected by its contribution to the effectiveness of those who practice it. (Skinner, 1974, pp. 223-224)
Note 4: Mechanisms of Prosociality (click to expand)
Baumeister et al. could be correct about free-will belief in broad strokes without getting the underlying mechanisms right. Here’s their proposed explanation for why belief in free will is related to prosociality:
Volition and self-control require the person to expend energy, and these expenditures enable them to act prosocially. Apparently disbelief in free will subtly reduces people’s willingness to expend that energy. Hence, disbelief in free will serves as a cue to act on impulse, a style of response that promotes selfish and impulsive actions such as aggressing and refusing to help. (p. 267)
Other explanations are possible. For instance, when people are taught about free will and what they are putatively meant to do with it, perhaps this harnesses rule governance (“should” rules for behavior) and the threat of social opprobrium contingent on antisocial behavior. Skinner suggested that cultures which embrace free-will beliefs are likely to establish legal consequences for people who use their “free will” in service of antisocial behavior:
Man has been said to be superior to the other animals because he has evolved a moral or ethical sense…. But what has evolved is a social environment in which individuals behave in ways determined in per by their effects on others. Different people show different amounts and kinds of moral and ethical behavior, depending on the extent of their exposure to such contingencies. Morals and ethics have been said to involve “attitudes toward law and government that have taken centuries in the building,” but it is much more plausible to say that the behavior that is said to express such attitudes is generated by contingencies that have developed over the centuries (Skinner, 1974, pp. 214-215).
So, perhaps what a cultural embrace of free-will beliefs generates is not some internal self-control process but rather enhanced social opprobrium to certain kinds of behavior, in which case the threat of disapproval and legal consequences gives people a key reason to behavior prosocially. And of course nothing rules out the possibility that a behavior like believing in free will could have multiple consequences — some beneficial, some not, and some immediate, some delayed. The Skinner and Baumeister views are not necessarily contradictory.
In both the Baumeister and Skinner-based perspectives, belief in free is proposed as a tool for short circuiting the control of unwanted behavior. What’s not explained in the Baumeister et al. model is why, if belief in free will is widespread (according to one survey I saw, more than 80% of Americans are all in on self-determination), and belief in free will promotes prosocial behavior, the world is such a prodigious mess. Skinner’s “social opprobrium” perspective is a better fit for this datum. In behavior analysis circles it’s well understood that the threat of legal punishment isn’t equivalent to the process of operant punishment, and actually works only weakly, thus failing pretty miserably as a deterrent of unwanted behavior. But maybe a society in which free-will belief is dominant will have a more elaborate set of legal sanctions for unwanted behavior, and maybe these sanctions help a bit to dampen antisocial tendencies. But as Skinner often argued, the real solution to the world’s problems is to structure society to teach prosocial behaviors early and directly, and reinforce them often.
Fellow Travelers
One of the difficulties in discussing free will in the first place is that it is an ill-defined concept. You can’t properly evaluate a claim until you specify precisely what is being claimed. At the core of all discussions about free will is the observation that sometimes organisms act without a proximal environmental initiating stimulus. Well, nothing controversial here — this is in fact inherent in the definition of an operant. There’s no specific “trigger” for an operant in the way there is for respondent behavior; merely an organism finding itself in an environment that connects to behavioral history in a particular way.
And, importantly, you certainly don’t have to invoke free will to explain such effects, because Nature is rife with similarly deterministic autopoietic systems:
An autopoietic system is a network of processes which produces all the components whose internal production is necessary to maintain the network operating as a unit. (Razeto-Barry, 2012, p. 549)
An autopoietic system is said to be self-sustaining, in the sense that, once created, it seems to seek out its own “fuel.” As a wildfire oxidizes plant matter, generating heat and sparks, it grows in size and becomes more likely to spread to adjacent plant matter. Obviously the fire has neither intention nor volition; its spread is a natural consequence of erupting in the sort of environment that supports burning. In such circumstances the existence of fire makes more fire likely. Similarly, operant behavior may be said to be self-sustaining because once it contacts reinforcement it occurs more often in the environments where reinforcement is contingent on it. This higher frequency of behavior then causes reinforcement to occur more often, which results in even more of the same behavior. Thus, the existence of a particular behavior makes more of that behavior likely.
Behavior analysts have developed solid theoretical accounts of such processes generally (e.g., behavioral momentum) as well described specific instances in which it would seem, from the outside looking in, that an individual is exerting control over future behavior (e.g., self-management, problem solving). These are precisely the kinds of phenomena that cause non-behavior-analysts to speculate about free will, which highlights unexpected commonalities between behavior analysts and folks like Baumeister and colleagues: We’re trying to understand some of the same things. Maybe, just maybe, we have something to teach each other.
Heavy Lifting
Now, as a committed behaviorist you might not like the Baumeister et al (2009) conclusions, and if not I have a question for you: Where are YOUR data on this topic? In the 80 years or so since Skinner began speculating about the human condition, as far as I’m aware nobody in behavior analysis has bothered to experimentally test his claims about the adverse behavioral consequences of believing in free will. That’s taking a lot on faith, and nowhere in the manual for “science of behavior” does it say you should buy a claim just because some famous person offered it up. Even if that person is Skinner.
So, kudos to Baumeister and colleagues for doing some of the difficult work that we ought to have undertaken. But clarification: I’m not saying that Baumeister and colleagues are right, only that if there’s a disagreement over whether belief in free will is beneficial or detrimental, there’s currently hard evidence only for “beneficial.” And that invokes another foundational touchstone of the behaviorist movement: objectivity. We claim to always let the data be our guide. Well, with relevant data now on the table, belief in free will can no longer be treated merely as a philosophical/conceptual issue. Moving forward, everything in our discipline’s discussions about free will (or at least the consequences of believing in it) should take the available data into account… and, I hope, prompt us to get more and better data.
Separate from the specific issue of how believing in free will affects other behavior, I hope it is obvious that we cannot be content to let other people ask, and answer, questions that cut to the heart of our science of behavior. When we do this, we miss opportunities to grow our science, we give the impression that our science has nothing to say about some really compelling matters, and we cede control of the societal message about our discipline. To me, that sounds way more detrimental than believing in free will ever could be.
UPDATE 9/7/26: For a sort of elaboration on Skinner’s worries about advrse effects of believing in free will, see this essay. The author does a good job of elaborating on how believing that others operate through free will may put us in what she calls the “blame and threat detection” mode. Translated into behaviorspeak, that means we are primed to employ punishment when other peoples’ behavior doesn’t meet our expectations or standards. This is a thought piece, not an empirical report, but at the least it makes an impassioned argument for applying a deterministic perspective to everyday life.
THEME MUSIC