Does evolutionary psychology erase moral responsibility? The short answer is no, but the long answer is more complex than we like to think.
Scapegoating the Naturalistic Fallacy
A common reaction to hypotheses put forward by evolutionary psychologists is to worry that many human behaviors could be excused or justified by deflecting blame onto evolved psychological adaptations. The evolutionary psychologist’s stock response is often to point to the naturalistic fallacy. A common example of David Hume’s “is-ought” problem, the naturalistic fallacy is a type of appeal to nature: a fallacious assumption that something is desirable simply by virtue of being natural. For example, claiming that organic foods are inherently better or more healthy than those treated with artificial products or processes would be a fallacy through appeal to nature. But more specifically, the naturalistic fallacy is the idea that what is natural is morally good, or at least cannot be morally bad. To claim, for example, that violence is morally acceptable because humans are naturally violent would be to commit the naturalistic fallacy.
It should be obvious upon brief consideration that equating the natural with the desirable is fallacious. A clear example is the contrast between diseases and vaccines: smallpox is natural, but no one will jump to its defense on this account. Correspondingly, a sane person would recognize that the smallpox vaccine cannot be worse than the disease itself simply because it is artificial (although the ease with which anti-vaccine propaganda has spread should speak to the allure of appeals to nature). Therefore, it must be something other than naturalness that determines merit in these cases. To challenge an argument on the grounds of committing an appeal to nature is to identify a flaw in the criteria used to determine something’s worth — and in the case of the naturalistic fallacy, its moral worth.
As I said before, diagnosing moral fears about evolutionary psychology as instances of the naturalistic fallacy is an extremely common defense of evolutionary hypotheses. However, I believe this might be a misdiagnosis. If you listen closely to the objections of many people — particularly those with progressive, egalitarian ideals — their concern seems slightly oblique to whether or not we can determine moral merit from claims about human nature. Let’s take a common worry as an example: the possibility of a lawyer using evolutionary psychological arguments to excuse or diminish the actions of a rapist.
If one were to cite the naturalistic fallacy to defuse the apparent tension in this example, the underlying assumption would be that the objector is concerned that the lawyer will persuade the jury to view the moral merit of rape in light of its naturalness. But is this so? Even if the lawyer could convince the jury that rape is “natural” for humans, I doubt that such an argument would genuinely persuade most jurors to abandon the belief that rape is morally wrong. Instead, I think the objector is concerned that the lawyer may influence opinions on the rapist’s moral responsibility, rather than the moral merit of their actions.
Thus, it is in the relationship between evolutionary psychology and moral responsibility that we must seek our answers to this objection. The debate in evolutionary psychology about whether men have adaptations for rape is long and complicated (see Thornhill & Palmer, 2001) — and the scales seem to be tipped in favor of it being a byproduct, not an adaptation — but it is not relevant here. For now, let’s imagine a world in which scientists have established beyond a reasonable doubt that men do have psychological adaptations to rape women. How would this (hypothetical) fact bear on the moral responsibility of a rapist?
If a lawyer appeals to an evolved mechanism to defend the actions of a rapist, the intent of such a tactic seems readily apparent: to reduce the rapist’s punishment by making it appear that he had less control over his actions, and is therefore less morally responsible for them. I think a key unspoken assumption here is that, if a behavior is “evolved,” it is somehow more causally deterministic than other behaviors.
This assumption is profoundly misguided — but for deeper reasons than we might think.
Nature, Nurture, and Free Will in a Physical World
I’ll take for granted that anyone reading this (a) is a physicalist — believes that only the physical exists, implying there is no “soul” or anything like it — and (b) agrees that humans are one species in a lineage of organisms, all of which developed from a previous stage in the same lineage according to natural selection (and other evolutionary forces). In other words, I’ll assume we all accept that there is no dividing line that separates humans from other animals, that we evolved by the same processes and from the same materials as other organisms, and that the entire evolutionary process consisted of nothing but physical components. (If either of these are not the case for you, try to engage with the following argument using them as assumptions.)
With these assumptions in place, there are some surprising and unintuitive consequences. These consequences reveal a fallacy in the common framing of free will in the context of evolution. Even among the most educated, the “nature-nurture problem” and related ideas are discussed in fundamentally misleading terms. Lemos (2002) ends up providing an excellent example as he attempts to integrate free-will libertarianism — the belief that humans are uncaused causers, autonomous and morally responsible — with Darwinian naturalism:
The libertarian can even concede that genetic and environmental factors incline a person to act one way or the other in the face of [alternative courses of action]. But what the libertarian will not concede is that all human decisions among such alternatives are ultimately controlled entirely by genetic and environmental factors. While our genetics and environment might incline us a certain way, they cannot be completely determinative of our actions when we are acting autonomously… [M]uch of the evidence we have for the truth of the libertarian position appeals to our own subjective awareness of moral struggles against inclinations that we have acquired due to genetic and environmental factors. (Lemos, 2002, p. 475)
What arguments like this overlook is that, in a physical world, genetic and environmental factors are exhaustive. When we perceive ourselves as struggling “against inclinations that we have acquired due to genetic and environmental factors,” the source of our internal struggle can only ultimately be traced back to other genetic and environmental factors: there is nothing else.[i] If this seems like an extreme claim, consider the process of human development from the very beginning.
Accepting the earlier premises (a) and (b), from the second a human is conceived, they are solely the product of the genes they inherit and a birth environment that cannot meaningfully be said to be under their control. In fact, strictly speaking, everything that is not a given gene would be considered part of that gene’s environment, including other genes in the same genome (Dawkins, 1982). Genes can only produce phenotypic effects through the interaction between the genes and chemicals in their environment, and those phenotypic products become part of the environment in turn. Of course, the line between environment and gene is fuzzy — no surprise, given that the definition of a gene itself is fuzzy. The gene-environment divide becomes even more complex when considering epigenetics, but epigenetic factors do not constitute a departure from the purely physical processes at hand. Epigenetic markers could be classified as part of the environment, part of the genes, or a little bit of both — it doesn’t really matter. The point is simply that, at an organism’s conception, genes and environment definitionally exhaust all potential factors influencing the organism’s development. What else could there be?
If everything about an organism’s environment and genes is physical, then everything that occurs during their development is ultimately the result of natural laws. The laws of physics govern the laws of chemistry, which — along with some element of randomness — govern the stitching together of a human being from chemicals in the environment according to the chemical “instructions” of DNA. Under normal environmental conditions, this process eventually produces sense organs. Some of these interact with the environment in non-chemical ways, like using light-sensitive chemicals in the retina or kinetically gated ion channels in the skin. But these stimuli are still physical, and their transduction is ultimately also determined by natural laws. The information transduced is then carried along a series of neurons, themselves made up of chemicals. The neurons then collectively conduct calculations — what we might loosely call “cognition,” including all subconscious processes — in a network that feeds back into other types of cells in the body, ultimately generating behavior.
When each step in the process can ultimately be attributed to either a law of nature or randomness, then at what point in this developmental process can this organism be said to have developed unique “control” over anything? Is there some special undiscovered type of neuron that steps in and suddenly makes autonomous choices for the previously deterministic brain? As far as I can tell, no one has ever been able to identify some Rubicon past which an organism made up of cells made up of chemicals would break free from the long series of physical causal interactions that produced the organism in the first place.[ii]
Even if we wanted to claim that we developed (at some unspecified point) something we could call free will, how far back in the causal chain would you have to go to say that you are responsible for your behavior? If you’ve been responsible for the past 1000 decisions you made, what could have changed between the last decision that was out of your control and the first one that was within it? Trying to insert something like free will into these physical processes produces an infinite regress (Strawson, 1994).
I have read many physicalist defenses of free will, in both deterministic and indeterministic universes, but none of them seem to escape the immanent mysticism in describing an “agent” operating on the myriad factors that constitute their psychology. From what I can tell, the authors of these accounts don’t seem to realize that the agent just is that very psychological constitution. Even while attacking the existence of free will, Levy (2011) claims that:
[people] can end up with a set of dispositions and values very different from those that were the direct joint product of their genes and environment. (p. 96)
No, they can’t. That’s all there is.
Now, Levy might argue that he said “direct” in the previous quote. But there can only be one direct product of a gene and its environment: “all genetic effects are ‘byproducts’ except protein molecules” (Dawkins, 1987, p. 300). To argue that any behavioral disposition is the “direct” product of a gene and its environment is just nonsense. It might be possible to argue that some behavioral dispositions are more direct products of genes than others — a claim that I suspect would still turn out to be mostly false or incoherent. But even this would not save the idea of changing your genetic and environmental predispositions from the same infinite regress we saw before.
Contrary to intuition, in a physical world, the burden of proof falls on those who would claim that we have free will. Most people feel and act as though they have free will (myself included), but that fact alone fails to convince me that its existence should be a fundamental assumption. Humans have been proven time and again to be pretty terrible reasoners under many circumstances, and countless aspects of our conscious experiences have been exposed as systematic illusions. Until someone demonstrates otherwise, it would appear that no aspect of human behavior is freely chosen or under our own control in the way we assume when we use the vocabulary of moral responsibility. In other words, regardless of the status of determinism from a physical standpoint, free will doesn’t make sense in a physicalist world. We and everything we do are ultimately the products of a chain of (determinate or indeterminate) physical causation, with no distinct, autonomous agents to interrupt it or change its course.
Abandoning Moral Responsibility
While some people (erroneously) describe the existence of evolved mechanisms that produce immoral behavior as a “hard truth” that we must accept, the real hard truth cuts much deeper. If we reject the existence of free will — as I’ve argued is logically inescapable — the concepts of moral responsibility and desert as we know them no longer make sense.[iii] We can’t meaningfully talk about someone earning anything, because they could not have done otherwise — at least not due to anything inherent to them as a moral agent. To moralistically praise and reward an automaton for achieving desirable results through a thoroughly physical causal process is unfair, given that the other failing automata were simply built worse for that task[iv] by the same mindless interactions between genes and their environments — neither of which were ever truly up to the automata. People in a physical world would be no more or less responsible for their behavior if it were ultimately caused by “environmental” factors or “genetic” ones, because the concept of moral responsibility doesn’t make sense in the first place.
So no, evolutionary psychology doesn’t erase moral responsibility: living in a physical world already did that.
A sophomoric response to the rejection of free will is to assert that we should lie down and rot if we have no control over our actions, but few if any people would truly adhere to this course of action. Insofar as we can be said to be unitary agents in the first place, our decisions are still our decisions. We don’t know what they will be until we make them; neither do we know the outcomes. Even if we could not truly have made decisions otherwise than we have, and neither are we the ultimate source of our decisions, we still make those decisions, and we still have to live with the consequences. Rejecting the existence of free will doesn’t rob us of our desires and feelings as sentient beings. Knowing that our choices are not truly our responsibility doesn’t change the fact that we can experience pain and happiness, and we can affect the suffering and wellbeing of ourselves and those around us. From the relatively parochial perspective of human beings, we can and should go on trying to make the best world we can.
At the moment, and for the past decade or so of my life, I’ve believed everything I’ve said in this article to be true. However, to make a more pragmatically compelling argument against the idea that evolved mechanisms can reduce moral responsibility, it might be more productive to make an intellectual compromise: to find a way to properly understand developmental biology while coherently defending the concept of moral responsibility. Beyond that, nothing in this post has truly addressed why it is evolutionary psychology in particular that provokes these concerns. And honestly, what I’ve said here isn’t a very satisfying answer to the inciting question, even for myself. What lies at the heart of the moral fear surrounding evolutionary psychological hypotheses? …Well, that’s a far more complex topic — one that requires its own post.
When I initially wrote and published this post, I had not yet discovered Dan Dennett’s 1984 book Elbow Room: The Varieties of Free Will Worth Wanting. As always when I read a book by Dennett, I found myself nodding along with most of it. For instance, his claim that “whatever ‘genuine consciousness’ or ‘real intentionality’ comes to, it must lie at the reachable top of that pyramid of natural, physical processes” (Dennett, 2015, p. 41) is in perfect concert with my arguments here.
I also found some compelling points that I had not considered. For example, Dennett describes an argument similar to the consequence argument that I defended here: essentially that, if you didn’t have control over the initial causes of your personality, how could you come to have control later? Dennett responds that,
by parity of reasoning, there couldn’t really be any mammals, since every mammal must have a mammal for a mother, and if you go back far enough in the family tree of any apparent mammal, you must find a manifestly nonmammalian ancestor, whose offspring just couldn’t themselves be mammals, and so forth. Since we know perfectly well that we are mammals, we take this argument seriously only as a challenge to discover whatever fallacy it is that is lurking within it. (Dennett, 2015, p. 92, emph. added)
I do agree that this point undermines the consequence argument somewhat, so I would not have leaned on it as heavily in my post had I read this book beforehand. His overall argument against the idea “that one could not take full responsibility for something unless it was entirely of one’s making” (Dennett, 2015, p. 92) made me feel less firm in my conclusion that we could not possibly have concepts like self-determination, free will, and moral responsibility in a physical world. As a result, I think I’ll soften my conclusion a bit and hold out for an alternative interpretation.
What Dennett did not provide, in my opinion, was any satisfying conclusion to the question of whether or not it’s fair to hold people morally responsible, even if we do arrive at some satisfactory-yet-fuzzy definition of self-control. For the first time when reading one of his books, he came to a point that entirely lost my support: I think his treatment of luck fails completely.
His argument against attributing our skills to luck can be summarized as follows:
[W]hile we cannot take personal credit for the success of our ancestors, our genes can. …And since the skills of self-control and deliberation have been put to a fairly severe test over the eons, there is a real basis in fact for our having high expectations about the deliberative skill, and more generally the capacity for self-control, of our fellow human beings. (Dennett, 2015, p. 102)
I agree with this point: that we are not simply lucky in existing, since natural selection has shaped us in such a way that we continue to exist. But he seems to put a lot of emphasis on the luck of us existing in the first place, to the exclusion of the luck that determines which extant organism you happen to be. It is not a matter of luck that you exist, but it is entirely a matter of luck which combination of genetic and environmental factors happen to create the mass of cells that you embody.
Dennett obliquely addresses this issue in the following example:
Imagine a footrace in which the starting line was staggered: those with birthdays in January start a yard ahead of those born in February, and eleven yards ahead of those born in December. Surely no one can help being born in one month rather than another. Isn’t this manifestly unfair? Yes, if the race is a hundred yard dash. No, if it’s a marathon. In a marathon such a relatively small initial advantage would count for nothing, since one can reliably expect other fortuitous breaks to have even greater effects. (Dennett, 2015, p. 103)
His blithe response about the difference between a marathon and a sprint misses a central statistical concept: the standard error of the mean. What he seems to ignore is that there will always be a distribution of the number of lucky breaks that people receive. Dennett’s point that luck “washes out” can only mean that the absolute range of the distribution is narrower: it does nothing to alter the relative ranking of the number and degree of lucky breaks each person gets. And we have no viable way to quantify how big of a difference there is between the two ends of the spectrum, so Dennett’s argument cannot account for the possibility that the gap is large enough that no amount of self-discipline could possibly bridge it. Dennett’s footrace analogy can be turned against him with a similar one from Dr. Martin Luther King Jr.:
It is obvious if a man is entered at the starting line in a race 300 years after another man, the first would have to perform some impossible feat to catch up with his fellow runner. (King, 1964, Ch. 8, section III)
Though King’s analogy was intended to illustrate the absurdity of expecting outcome differences between races to simply disappear after the abolition of slavery and legal racial discrimination, the same principle applies to Dennett’s example.
Moreover, when Dennett says “one can reliably expect other fortuitous breaks to have even greater effects,” he seems to be assuming that most lucky breaks will fall in the opposite direction from where their recipients started, which is patently ridiculous. Bad and good luck can both compound. If seven billion people start a footrace staggered by birthdate, some January babies will have more lucky breaks than others, and some December babies will have worse luck than others. These two groups will fill out the tail ends of the distribution of the standard error of the mean for luck.
He does acknowledge, at least implicitly, that a lower end of the distribution must exist:
Of course some unfortunates, though born of skilled self-controllers, are defective, through no fault of their own. We do not consider them responsible. They are excused. But we do expect a lot from the rest of us, and for good reason. We are not just lucky; we are skilled. (Dennett, 2015, p. 102)
Yet he never explains why he seems comfortable drawing a line between “those who are singled out as defective” (Dennett, 2015, p. 104) and those at the lower end of the spectrum of “normal” citizens. It seems clear to me that the people he is talking about are exactly those who have received compounded bad luck. But Dennett neither directly asserts that there is a clear boundary between the two groups, nor does he attempt to justify a cutoff point on the luck spectrum. And more critically, he seems to ignore the implications of the “defective” group’s very existence in his schema.
As far as I can tell, his claims that “everyone comes out more or less in the same league” (Dennett, 2015, p. 104) and that “no one actually has more luck than anyone else” (p. 105) are based on nothing but blind faith. Worse yet, they appear to be contradicted by simple empirical evidence and mathematical reasoning. Why wouldn’t there be a luckiest and unluckiest person in the world, with a distribution between them? This is prima facie absurd, and nothing in his book seems to back it up.
Everyone starts the race of life at a different point, but the finish line is the same for everyone. For the most part, we can’t really see where a person “started” (their genetic + initial environmental endowment): all we can really see is where they are in the race now. So why is it fair to reward someone for being at the front if we don’t know where they started, or how hard they had to work to get as far ahead as they did? Couldn’t someone farther behind have worked equally hard — or harder — to get to where they are?
Ultimately, where I disagree with Dennett most is on the concept of fairness. This conflict is most apparent in the following quote:
Imagine trying to change the rules of basketball in the following way: if the referees decide that a particular basket was just a lucky shot, they disallow the points, and if they notice that bad luck is dogging one of the teams, they give that team compensatory privileges. A perfectly pointless effort at reform, of course, which would not appeal to anybody’s sense of fairness. (Dennett, 2015, p. 104)
The unspoken assumption here is that the purpose of basketball is equivalent or similar to the purpose of life. But sports are a specific category of human activity, with their own specific goals, agreed upon by the participants: victory, an outlet for competition, entertainment, personal fitness, proving superiority, demonstrating skill. While many of these can be components of a fulfilling human life, it would be much harder to get people to agree that any of them were the purpose of life, let alone the purpose of society as a whole. In fact, from more socially left-leaning perspectives, many of these are in conflict with one of the primary “purposes” of society: maximizing the wellbeing and minimizing the suffering of sentient creatures.
So yes, in the context of sports, compensating for bad luck is unfair. But contrary to what some people seem to believe, sports are not life. The game of life has different goals than sports, and therefore potentially very different standards of “fairness.” Maybe if you’re an extreme libertarian, or the personification of natural selection itself, the purpose of life is competition — but it doesn’t have to be, and using that as an assumption to judge what is “fair” is completely at odds with my moral intuitions.
Dawkins, R. (1987). The extended phenotype: The long reach of the gene. Oxford University Press.
Lemos, J. (2002). Evolution and Free Will: A Defense of Darwinian Non–naturalism. Metaphilosophy, 33(4), 468-482.
Levy, N. (2011). Hard luck: How luck undermines free will and moral responsibility. OUP Oxford.
Strawson, G. (1994). The impossibility of moral responsibility. Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition, 75(1/2), 5-24.
Thornhill, R., & Palmer, C. T. (2001). A natural history of rape: Biological bases of sexual coercion. MIT Press.
Van Inwagen, Peter (1983). An Essay on Free Will. Clarendon Press.
Dennett, D. C. (2015). Elbow Room, New Edition: The Varieties of Free Will Worth Wanting. MIT Press.
King, M.L. (1964). Why We Can’t Wait. New York: New American Library.
[i] Yes, this does imply that I think akrasia is oxymoronic, but that discussion would be a little bit of a tangent, and potentially better explored in a different post.
[ii] I am far from the first to make this argument, best known as the consequence argument (van Inwagen, 1983). My point here is largely to illustrate that comparisons between genetic and environmental contributions to behavior can’t do much to settle the issues of free will and moral responsibility. More crucially, I have not seen anyone make one specific point that I make here: that genes and environment exhaust the possibilities of factors influencing and ultimately determining behavior, precluding the existence of a causal “agent” in compatibilist and libertarian accounts.
[iii] There may be arguments in favor of moral responsibility in the absence of free will, but that may be a topic for another post.
[iv] This statement seems to imply a level of essentialism and inevitability in our dispositions that I plan to debunk and undermine in my next post. I can understand how the statement that some people are “built worse” for a task than others may sound very elitist and fatalistic, but I hope to eventually decouple such conclusions from the argument I’m making here. Thinking of ourselves as unchangeable automata only works at the high level of analysis at which I’m operating in this particular post: in our everyday lives (and particularly in policymaking), I think we should behave as if we can change the course of the future, as I explain in my second-to-last paragraph above. But operating as though we have power over the future does not necessarily entail holding people morally responsible (favorably or punitively) for their actions in the past.