Abstract Recent advances in the evolutionary and cognitive sciences of religion have raised questions about whether the assumptions and findings of these fields as applied to religion conflict with belief in gods. Specifically, three scientific approaches to religion (Neurotheology, Group Selection, and Cognitive Science of Religion) are sketched, and five arguments against theistic belief arising from these approaches are discussed and evaluated. None of the five arguments prove formidable challenges for belief in gods.
Key words: Belief; Cognitive science; Evolution; Religion; Theism
The long and adversarial relationship between Darwinism and theism continues. Not satisfied to explain the nature of living things, evolutionary scientists now aim to explain religion itself. Pascal Boyer’s Religion Explained raised a few eyebrows with subtle suggestions that religion is “airy nothing” but a byproduct of evolved human minds.1 Scott Atran’s In God’s We Trust: The Evolutionary Landscape of Religion even more directly affronted some religious sensibilities by purporting to explain theistic beliefs as irrational and counterfactual thought propped up by evolved mental capacities.2 Repackaging these same types of arguments, Daniel Dennett has finally managed to gain considerable popular attention for the movement with his recently released Breaking the Spell: Religion as a Natural Phenomenon.3Dennett’s book and its apparently lethal implications for religious belief have been discussed on multiple occasions on National Public Radio and in wide-circulation publications such as the New York Times.4
That this area of scholarship, captured under the umbrella bio-psychological theories of religion, appears to be growing in visibility and influence suggests a need to consider carefully the current and potential implications of the evolutionary and cognitive sciences of religion on religious belief. Will evolutionary sciences finally win the war and defeat theism once and for all? Some atheists within and outside these sciences hope so. Likewise, the fear that the bio-psychological theories of religion finally hold the right weapons to destroy theism torments some believers observing the field and even working in it. In this essay, I offer a brief overview of recent bio-psychological theories of religion and then present a number of arguments against theistic belief based on this scholarship.
The bio-psychological theories of religion come in three different varieties: neurotheology, group selection, and cognitive science of religion. All three subfields represent attempts to explain religion reductively by appealing to evolved properties of the mind-brain and their impact on behavior. That is, religious beliefs and practices arise because of the activity of evolved functional properties of human brains. Nevertheless, the three schools take importantly different approaches to explaining religion.
Neurotheology primarily concerns identifying which components and dynamics of the brain underlie religious experiences and subsequent beliefs.5 The suggestion is that religious phenomena can be identified as the (perhaps accidental) output of evolved neural circuitry. Evolved brains have components that have arisen because of their usefulness to survival that happen to interact in such a way as to generate religious experiences, including experiences of the presence of an invisible, disembodied being. These experiences are shared and may be codified into common religious beliefs, particularly in the existence of gods that account for these experiences.
The most dramatic and controversial element of the neurotheology school is work by Michael Persinger. Reportedly, Persinger has been able not only to identify brain regions responsible for certain types of religious experiences, but has also succeeded at artificially inducing such experiences using electromagnetic fields, experiences that appear similar to being in the presence of a supernatural being.6 If such experiences may be artificially induced, could it be that god concepts arise from misfiring of these same brain regions?
David Sloan Wilson’s Darwins Cathedral represents the group selection perspective. In short, Wilson argues that religious systems encourage pro-social behavior, and groups that exhibit pro-social behavior (cooperation, lack of cheating and stealing, etc.) will tend to out survive and reproduce groups that do not exhibit these traits. So, religious communities will tend to survive better than non-religious communities do.7
As religious communities will have stronger pro-social tendencies, they will cooperate better. As they cooperate better, they will survive and thrive better than competing communities. Hence, religious communities—and whatever genetic information accounts for their religiosity and pro-sociality—will tend to survive and expand at a greater rate than non-religious communities. Over time, then, people with the biological disposition to be religious will increasingly outnumber non-religious people. This selection process thereby accounts for the widespread existence of religious people.
By this account, religion does not persist because it makes any special truth claims or inject lives with meaning. Rather, religion exists because it helps people survive. Religion possesses utility.
The most developed bio-psychological field dealing with religion has come to be called the cognitive science of religion. This camp includes Atran, Boyer, and a growing number of others.8 This work has been strongly associated with evolutionary psychology, particularly that of Atran and Boyer, but an evolutionary theoretical foundation is not strictly necessary.
Though research has concerned religious rituals, religious community morphology, god concepts, prayer, morality, after-life concepts, scripture use and a number of other areas, what unifies these projects are a number of fundamental commitments. The cognitive science of religion begins by acknowledging that:
(1) In general, the basic functional processes of human minds are the same regardless of cultural environments. That is, by virtue of a common human biology living in a remarkably uniform natural world, minds develop similarly everywhere as regarding, for instance, intuitive causal reasoning.
(2) Further, as the cognitive sciences have shown, human minds are not general-purpose information processing devices but are highly specialized conglomerates of many functional subsystems that solve particular problems.
(3) These subsystems importantly color or shape perception and cognition regarding the natural and social world. They do not receive passively and indiscriminately whatever is “out there,”
(4) These contours of human minds inform and constrain recurrent patterns of human thought and action including religious thought and action.
(5) Hence, recurrent features of religious thought and action (e.g. belief in gods) can be explained (or predicted) by appealing to requisite conceptual structures. Particular thoughts and actions will occur more frequently among humans than other possible thoughts and actions by virtue of their foundation in the basic dynamics of human minds.
To illustrate, we can confidently predict that in no yet-to-be-studied religious tradition will people practice a ritual system that demands a sacrifice of sparrows 13 times the square-root of the number of days past since the last full moon. We can also predict that the gods of the system will not be five-dimensional amoeboids that experience time backwards, only know what is not true, and behave beneficently on every third day. Why not? Because such a religious system incorporates reasoning completely foreign to the naturally occurring preferences and capabilities of our minds. The difficulty does not lie in the complexity of the information required per se—primitive computers could easily calculate the appropriate sparrow sacrifice—but because our minds are receptive to a very narrow subset of theoretically possible information. Find out what information we thrive on consuming and producing, and you find out why religions (or any cultural phenomena) tend to look the way they do.
The cognitive science of religion also includes a number of scholars interested in the formation of children’s religious beliefs. Young children’s understanding of death and the possible persistence of persons after death,9 children’s affinity for intentional design explanations of the natural world,10 and children’s acquisition of various aspects of god concepts receive attention.11 What these cognitive developmental psychologists offer to the more general cognitive science of religion is the observation that religious concepts for which children’s minds seem to have a natural tendency to entertain and generate will tend to become more common. Showing, for instance, that children find creationist accounts of animals more compelling than evolutionary accounts regardless of parents’ beliefs, helps explain why creationism is so resilient in the face of alternative accounts.12
Though scholars in these fields have largely avoided explicit published treatments of what their science might mean for religious belief, popular reactions clearly demonstrate that believers fear behind titles such as Religion Explained and Breaking the Spell and Is God an Accident?, lurks a genuine assault on religious belief at the experienced hand of science. Authors in this area typically do not directly attack religious belief but leave plenty of reasons for believers to be suspicious. Psychologist Paul Bloom’s Atlantic Monthly article serves as a fine case in point.13
In Is God an Accident?’, Bloom elegantly summarizes some central observations from the cognitive science of religion including his own work concerning what appears to be a natural and early-developing tendency in people to be mind-body dualists.14 Such a tacit belief makes acquisition of various life-after-death notions quite easy indeed. Bloom offers that the organization of human minds accounts for why religious beliefs are so persistent in the face of science. Nowhere in his article does Bloom offer an argument for how the bio-psychological approaches show religious belief to be false. Rather, Bloom gently asserts the incompatibility of science and religion as when he imagines what would happen if religious folk allowed for insights from science to permeate their belief system: “Scientific views would spread through religious communities. Supernatural beliefs would gradually disappear as the theologically correct version of a religion gradually became consistent with the secular world view.”15 In addition, characteristic of many authors in the cognitive science of religion, Bloom casually assumes the falsity of religious beliefs in his presentation of the scientific evidence. For instance, he explains that natural systems “go awry” and consequently give rise to religion, by “inferring goals and desires where none exist.”16 Here we have no direct argument or evidence that gods or other entities do not exist. This conclusion is assumed in the presentation. These subtle rhetorical moves give observers the impression that these new treatments of religion are fundamentally dangerous or even outright hostile to theism.
To make explicit just where the potential points of danger for theistic belief might lie, in the following, I discuss five possible arguments against theism on the basis of the bio-psychological explanations of religion. These arguments represent positions that I suspect (but cannot prove) some authors in the area hold based on my conversations with these authors and exposure to both their written and spoken ideas. These arguments also represent those that believers fear lie behind the texts based on the conversations I have had with theists both inside and outside the field.
Arising from advances in neurotheology, one implicit attack on religious belief could go this way:
We have identified the regions of the brain responsible for religious experience and can artificially induce religious experience. Therefore, its causes are entirely natural and so, we have no need to appeal to supernatural to account for them. Hence, theistic belief is unjustified.
To get this argument off the ground we have to grant the claim that neurotheology has actually identified the brain regions and functions that give rise to those religious experiences that are fundamental to belief in gods (broadly construed). This supposition is certainly debatable, and its full achievement might prove impossible, for now, let us assume that neurotheology will eventually be successful at detailing all the biological functions at play when someone comes to believe in a god.
Even with this enormous concession, this argument remains a non-starter. To get from a perfect biological specification of a belief to the belief being wholly explained by natural processes, we need a premise that the supernatural does not regularly causally act upon the neural substrate. Otherwise, a god could be directing neuronal networks to experience it. I imagine some theists would resist granting such a premise, and certainly, science cannot marshal requisite evidence to the contrary, as it is notoriously bad at measuring supernatural activity.
However, suppose the theist grants the further premise that the supernatural does not causally act upon the neural substrate. Then does the argument manage to defeat this form of theism? Not yet. The accommodating theist could simply maintain that a god or gods put into place the natural order—including the organization of human brains via evolution—such that human brains naturally give rise to religious experiences under particular situations. That natural biological processes have been found to correspond to religious experiences and beliefs does not entail that gods do not exist.
I find no way to salvage this argument even with the concessions made. Without the concessions it is even more fragile. Its pivotal flaw seems to be a misunderstanding about explaining epistemic states in terms of biology. Suppose I believe I see a robin outside my window. You tell me you can exactly specify the neural pathways responsible for generating my belief. Does that mean the robin is not really there? Hardly. That religious experiences can be artificially manufactured is irrelevant in the same way. Neurologists have found that by stimulating the cortex they can create various perceptual experiences. No one wants to argue that if scientists can use electromagnetic fields to make me believe I see a robin that suddenly I am not justified under normal conditions in believing I see robins.
For some theologies that see minds as wholly separate from bodies (including brains) the findings of neurotheology may present difficulties for this view of humanness. That particular bodily activities correlate perfectly with particular experiences should give one pause to suppose that all mental life is causally separate from bodily processes. Nevertheless, aside from complete dualists, finding biological substrates bears little impact. How many contemporary theists doubt that their brains are active during religious thought or experiences? The only news is just which parts of the brains are active. That brains might be active during religious thought or experience is no affront to religious belief so why would the insight that this or that particular brain area might be active matter?
Another implicit argument appears to be that
Selection pressures (operating on either groups or individuals) have led to various dispositions or propensities in human minds that happen to give rise to religious belief. Religious beliefs are, therefore, accidents or byproducts of evolution. As such, religious beliefs cannot be trusted. Belief in gods amount to cognitive illusion.
This argument, sometimes attributed to group selectionists but more in line with the cognitive science of religion, is rich in rhetoric but poor in substance. Words like “accident,” “byproduct,” and “illusion” sound damning, but what do they really mean in this context?
“Accident” and “byproduct” both capture the idea that evolution did not select for people to believe in gods. Rather, it selected for other human behaviors (and the genetic material that gives rise to the behaviors and any prerequisite beliefs), and religious beliefs just happened to spring out of the way humans are put together. Religious beliefs themselves did not confer any selective advantage; rather they ride on beliefs and behaviors that did confer advantages. Music, too, may be an accidental byproduct of evolution in this sense. Natural selection did not favor music lovers and producers over others, rather, other functional units of the brain that did confer selective advantage happen to promote music.
The religion as byproduct position is fairly common in cognitive science accounts of religion, but what follows from this observation, assuming it is true?
Theists generally do not believe in gods, nor do they justify their beliefs, on the grounds that such beliefs conferred a selective advantage in our evolutionary history. As no weight rests on this foundation, to remove it does no harm to these beliefs.
Further, many beliefs and values that the scientists of religion themselves hold dear likely would be weakened by the same argument if it applied to theistic commitments. Contemporary beliefs and behaviors bestowed by science and technology arose far too late in our history to have played a role in natural selection of humans. Evolution did not select for calculus, quantum theory, or natural selection. Are these beliefs then suspect for being “accidents” or “byproducts” of evolution? With this line of reasoning, Darwinism would face the ax alongside theism. Suggesting that when beliefs arise as byproducts of evolution they must be jettisoned would force abandonment of the premises regarding selection in the argument—more on this Suicidal Tendency of such arguments below.
A different variation of the Evolutionary Byproduct argument suggests that theistic beliefs are not simply accidents or free riders on adaptive systems, but byproducts in the way that perceptual illusions arise from properly functioning evolved perceptual systems. Accusing religion of being a “cognitive illusion” (or “delusion”) fares no better as a serious attack but nonetheless raises interesting comparisons. Psychologists have documented numerous illusions in which our minds tend to tell us something is the case that, upon closer inspection, turns out to not be so. For instance, people tend to see random events (such as series of coin tosses) as more orderly than they actually are, and see correlations where none exist.17 These count as conceptual illusions brought about by how our minds naturally function. Perceptual illusions similarly tell us that something in the world is different than it actually is. For instance, when viewing a rainbow, people typically perceive several bands of color. Science has demonstrated, however, that a rainbow does not display bands of color but a perfect continuum of light wavelengths. The illusion of bands occurs by our visual system’s tendency to categorize stimuli into meaningful units (termed categorical perception). The cognitive capacities that give rise to conceptual and perceptual illusions conferred selective advantages but occasionally produce relatively harmless illusions or mistakes.
One might argue that religious beliefs have much in common with these other cognitive illusions. The evolved, natural functions of human minds operating in the ordinary natural world prompt the belief in and spread of religious ideas, much as the evolved, natural functions of human minds operating in the ordinary natural world prompt the belief in illusions such as illusory correlations and bands of color in rainbows. Again, this comparison amounts to a conventional claim in cognitive science of religion. What the term “illusion” adds is the evaluation that a perception or belief is in error. On the basis of closer inspection (often using scientific or statistical methods) we can discover that a correlation we thought existed does not actually exist, or a sequence of coin tosses that looks unlikely could in fact arise by chance, or a rainbow does not actually present bands of color. To be able to call genuinely religious beliefs “illusions” we need to be able to demonstrate that they too, upon further examination, are in error. However, this task is not aided by the evolutionary or cognitive sciences of religion. To determine that a theistic belief amounts to an “illusion” requires a metaphysical commitment. To call theism “cognitive illusion” is a premise and not a conclusion of this argument.
The third argument to consider is the flip side of the preceding argument. Instead of emphasizing religious beliefs as unnecessary free riders on natural selection, it emphasizes the value of religious beliefs and subsequent behaviors.
Religious beliefs have utility (e.g. in social arrangements) leading to their natural selection. As religious beliefs are selected for their utility and not their truth, they should not be trusted in terms of truth-value.
Though it starkly diverges from the previous argument, it carries its own fatal flaw. Nevertheless, such an argument is valuable to consider as it expresses an important but often overlooked element of evolution by natural selection.
This argument, derivable from the group selection account of religion, begins with an interesting empirical claim about religious beliefs, that they have utility in the sense of conferring advantages in reproduction and survival. Obviously, this claim must be a general claim about belief in superhuman intentional beings with moral concerns (i.e. gods) and not a claim about any and every specific religious belief. (A belief that universal celibacy is the true path to salvation would likely lead to negative reproduction effects.)
However, even taken as a claim about religious beliefs generally, it is not clear how believing in gods—even those that promote pro-social behavior—can be encouraged through natural selection. Natural selection, after all, selects for behaviors and not beliefs or attitudes. A person that generously shares with neighbors and never cheats because she believes in a punitive god is treated the same by natural selection as a person that behaves identically with no particular beliefs (let alone theistic ones) surrounding the behaviors. As it is pro-social behavior that allegedly is favored by natural selection, the argument should be recast with pro-social behavior taking the place of religious beliefs.
Pro-social behaviors have utility (e.g., in social arrangements) leading to their natural selection. As pro-social behaviors are selected for their utility and not their truth, they should not be trusted in terms of truth-value.
As behaviors do not have truth-value, clearly such an argument is confused.
However, suppose we fortify the argument in another way. As much as any particular cognitive structure typically gives rise to behaviors, natural selection may reward those behaviors and the underlying cognitive structures that give rise to the behaviors. Hence, indirectly, evolution may select for cognitive devices and any beliefs they might generate (that then motivate behaviors). Replacing pro-social behaviors with cognitive devices that generate religious beliefs in the argument, we get
Cognitive devices that generate religious beliefs have utility (e.g. in social arrangements) leading to their natural selection. As cognitive devices that generate religious beliefs are selected for their utility and not their truth, they should not be trusted in terms of truth-value.
This version of the argument more closely resembles what a cognitive scientist of religion might suggest instead of what is claimed by Wilson’s group selection account. Substantively, it does not differ from the Evolutionary Byproduct argument. Both claim religious beliefs free ride on naturally selected cognitive structures. The difference is only in the introduction of a premise concerning what should and should not be trusted for belief. The Evolutionary Byproduct argument suggests that if natural selection did not select for a particular belief, but only its underlying cognitive mechanisms, it is not to be trusted. (However, note that, strictly speaking natural selection does not select for any beliefs but only underlying cognitive mechanisms and these only indirectly.) The Religious Utility argument takes the other side and suggests that if natural selection did select for a cognitive device, in part because of the beliefs and subsequent behaviors it generates, then the beliefs cannot be trusted.
If a belief is an accident of evolution, it cannot be trusted. If it is a legitimate product of evolution, it still cannot be trusted. Why not? Here is an important but often neglected feature of evolution by natural selection: it only favors minds that generate survival behaviors and not necessarily true beliefs. As already discussed, natural selection does not care about beliefs. If true beliefs lead to the extermination of a gene pool, so be it. If systematic errors or a particular profound disconnect with reality encourage survival and reproduction, then those genes win. In fact, the illusions discussed above are part of the vast psychological literature demonstrating that human minds seem systematically to get things wrong for the sake of survival. When it comes to natural selection, Truth is expendable.
However, the twice-revised version of the Religious Utility argument still fails as a defeater for religious belief. Granted, just being a product of evolution surely does not give us confidence in the truth of religious beliefs. However, as discussed in relation to the Evolutionary Byproduct argument, to say that being the product of evolution undermines religious beliefs is overstating the opposition between natural selection and Truth-representing minds. Natural selection does not discriminate against Truth; it just does not attempt to preserve it.
At a conference I was asked if children simply acquire belief in gods (1) because such beliefs cannot be proven wrong (i.e. they are not falsifiable) and (2) because of the persuasive and coercive power of parents and communities. A cognitive scientist of religion answered the question by explaining that regardless of the coercive or persuasive techniques of adults, children will not believe just any claims that cannot be proven wrong. Religious ideas are readily acquired because of their particular fit with a large number of cognitive devices that normally develop in humans. Persuasion or coercion can only augment acquisition of this particular subset of beliefs that enjoy the support of cognitive devices.
The questioner could have responded (but did not) with another proposed argument against belief in gods.
People are credulous recipients of theistic beliefs (e.g. from parents). Natural selection provided people with the cognitive faculties that make us credulous recipients. As we now know why people so readily believe in gods, continuing to believe is irrational.
An example may illustrate the strength of this argument.
Suppose someone persuades you of some proposition P so that you believe that P is true. Later you discover that at the time someone convinced you that P was true, you were under the influence of a drug that makes you particularly gullible. Are you then still rationally justified to believe that P is true? At the very least, you should reevaluate P while sober. Perhaps P will still be rationally justifiable on other merits but perhaps not.
The difficulty with applying this illustration to the Inherited Belief argument against religion is that we are never “sober.” That is, the same factors that made us vulnerable to theistic beliefs are still operating once we are aware that they are there and from where they have come. One might suggest that these, too, are grounds for rejecting P or theistic beliefs outright, just to be sure, we do not adopt an unmerited belief. After all, how could theistic beliefs be fairly evaluated if we have these evolved capacities pushing us toward belief? Theism will always seem plausible. Therefore, theism should be rejected.
However, when do we ever use the plausibility of a proposition as grounds against believing the proposition? Something clearly is not right in this line of reasoning.
One problem with the Inherited Belief argument as presented is that it tacitly assumes that theistic beliefs are only credible because of the operation of these evolved mechanisms. That is, it assumes that no alternative reasons for belief are available outside of the cognitive mechanisms specified by cognitive scientists of religion. As essentially all of the mechanisms cognitive scientists point to as supporting religion are intuitive or even implicit, explicit reasoning such as that done by philosophers and theologians falls outside the pail. Arguments and evidence gathered through these other avenues are safe from the disturbing influence of the cognitive science of religion.
Perhaps, then, a weaker version of the Inherited Belief argument is in order.
People are credulous recipients of theistic beliefs (e.g. from parents). Natural selection provided people with the cognitive faculties that contribute to making us credulous recipients. As we now know why people so readily believe in gods through these faculties, their contribution toward belief should be discounted. Hence, fewer reasons exist for theistic belief.
While not defeating religious belief outright, this version purports to weaken theism by cutting away some (perhaps most) available reasons for belief. Some individuals might be able to marshal enough justification to continue to believe but many believers would likely no longer have enough rational ground on which to stand.
Before feeling too satisfied with this vicious blow against theism, let us consider the drug illustration anew. A revision to parallel the renewed Inherited Belief argument might be the following. Suppose someone persuades you of some proposition P so that you believe that P is true. Later you discover that at the time someone convinced you that P was true, you were under the influence of a drug that makes you particularly gullible regarding many reasons to believe P. You should then reevaluate P in light of this information and disregard the reasons for believing P accounted for by the drug. Belief in P should be weakened.
However, consider another alternative. Someone persuades you of some proposition P so that you believe that P is true. Later you discover that at the time someone convinced you that P was true, you were under the influence of a “smart drug” that makes you particularly clear-headed and likely to believe only true propositions. Does this information undermine your belief in P? Surely not. If anything, it bolsters your confidence in P. If, on the other hand, you learned that you had been under the influence of a “stupid drug” that not only made you gullible but also especially gullible concerning the most patently false propositions imaginable, you would be especially anxious to reevaluate and probably reject belief in P.
The cases of the smart drug and the stupid drug illustrate that the ability for the Inherited Belief argument to undermine or encourage religious belief lies in whether the evolved mechanisms that allegedly support theistic belief are prone to produce accurate beliefs or inaccurate ones. If they generally produce accurate beliefs, such an argument actually favors theism. Hence, by itself, the Inherited Belief argument fails. I now turn to a fifth and final argument that explicitly builds in the charge that these evolved capacities are error prone.
The cognitive science of religion demonstrates that cognitive mechanisms of our evolved minds provide most impetus for believing in gods. As these mechanisms are error-prone, they cannot be trusted to give us Truth. As we cannot trust these aspects of our minds to give us Truth, we cannot trust most of our impetus for believing in gods. Therefore, cognitive science of religion weakens belief in gods.
Having learned from the mistakes of its predecessors, this argument avoids trying to defeat theism outright but aims only to weaken theistic belief. Its opening premise is a reasonable inference from theoretical claims and relevant evidence from the cognitive science of religion. Certainly, if a belief arises largely because of evidence (or reasons) from sources known to be untrustworthy, the belief is in danger unless sufficient other evidence (or reasons) exist apart from the dubious source. Assuming that the cognitive science of religion is on the right track, the only question seems to be whether the cognitive mechanisms that support belief in gods are error-prone to an extent that they cannot be trusted.
To illustrate, one mechanism implicated in theistic belief has been dubbed HADD—the Hypersensitive Agency Detection Device.18 This device scans the environment for intentional agents and their activity. It has been suggested, and experimental evidence seems to support, that this device has a tendency to “detect” agency given fairly ambiguous evidence. The reason given for this hypersensitivity is a “better safe than sorry” strategy.19 As other intentional agents (including people and animals) represented our ancestors’ greatest threat and greatest promise for survival and reproduction, those with an agency detection device that did not often miss agency would have been at a selective advantage. Better to assume that a noise in the brush is an agent than to miss the chance of detecting an agent and miss out on capturing prey, or miss detecting an agent and being eaten. Such a strategy would lead to relatively harmless false-detections. HADD’s ability to find agency given scant evidence would make HADD a strong candidate for detecting the activity of ghosts, spirits, and gods and thus promote belief in their existence. If HADD is tuned to find agency a bit too easily, then HADD is an error-prone mechanism promoting belief in gods. Being error-prone, HADD cannot be trusted and so, the evidence HADD contributes to belief in gods should be discounted.
As HADD’s role in religious belief illustrates, this argument carries some potential merit. Nevertheless, here I focus on two weaknesses of the general argument (including HADD’s role), and not additional criticisms specific to HADD.
Though HADD might appear to be the epitome of an error-prone device, fairly judging the accuracy of the entire concert of cognitive mechanisms involved in promoting belief in gods is no mean feat. If HADD worked alone in determining when or where we discovered the existence of agents, we would never be able to tell definitively when it was wrong. It would be analogous to trying to determine how many colors are in a rainbow only by looking at it over and over again. We would likely get the same answer repeatedly. Fortunately, HADD does not work alone. Other cognitive mechanisms, including our abilities to consider evidence reflectively, can override HADD or any other single cognitive mechanism that tries to generate a belief. This realization helps us decide that HADD is indeed error-prone, but it creates another problem. As HADD and the other cognitive mechanisms that promote belief in god do not work alone to generate beliefs, their accuracy cannot be evaluated in isolation.
What evidence, then, do we have that our total system (including HADD) that generates belief in intentional agents (such as gods) is error-prone to the extent that it cannot be trusted? Here we might be able to show that when people swear that computers are malicious intentional agents, upon further examination, computers turn out to be simply complex machines without intentions (or malice). The believer in computer agency himself might recant. That would be a case of mistaken agency detection. Alternatively, a person might be convinced that an intruder is in the house and lock herself in her bathroom for an hour as a precaution only later to discover that the sinister noises came from a radio. That would be a case of having mistakenly detected an agent. These sorts of episodes could be tallied up against the cases of having accurately detected agency (e.g. I thought a person was in front of me and I was right; I thought this letter was sent to me deliberately and I was right), and we could see how accurate agency detection typically is. However, note that we still have the problem of not definitively knowing when agency has been accurately detected. Are people really agents? How do we know apart from the cognitive systems responsible for agent detection? Do we count detections of ghosts, spirits, and gods as mistakes or as accurate detection? If as mistakes, we beg the question at hand. The same problem confronts other hypothesized cognitive systems responsible for promoting belief in gods.
These observations concerning determining the accuracy of our cognitive mechanisms to give us accurate beliefs raise another problem with the Error-Prone Mind argument. I call this problem Collateral Damage.
Collateral damage is a euphemism for killing unintended civilians in the course of warfare. I am afraid that in this war with belief in gods, several of the arguments presented here including the Error-Prone Mind argument run the risk of causing severe collateral damage. If it can be successful at damaging belief in gods, it will likewise damage other, unintended, beliefs.
Suppose we can agree that our agency detection system is indeed so inaccurate that it cannot be trusted. It gets things wrong more often than right. Then, if it encourages us to believe in a supernatural agent, we have every reason to ignore its encouragement. Belief in gods is certainly weakened, but so is belief in any and all intentional agency.
On what basis do we believe that other people have minds? At least in large part by virtue of the same cognitive mechanisms responsible for encouraging belief in superhuman minds. H ADD and other mechanisms converge on the intuition that other people are intentional agents with beliefs and desires and not merely automatons or machines inside of human shells. We have no direct, conclusive evidence to support the belief that people are intentional agents with minds. Minds cannot be directly observed. We have no empirical evidence for their existence.20 If belief in gods is weakened by the Error-Prone Mind argument, so too is belief in all intentional agents. This collateral damage may be tolerable to some cognitive scientists, but few others.
The cognitive sciences (including experimental psychology) have given us plenty of reason to doubt the veracity of our minds on a number of fronts. Above I mentioned perceptual and conceptual illusions as examples. Some of these cognitive errors are persistent and systematic and influence how we perceive the world, think, reason, and conduct ourselves in everyday life. Suppose our agency detection system is profoundly error-prone. So too are a host of other cognitive systems.
If the Error-Prone Mind argument is well formed, it should allow for innumerable additional substitutions.
The cognitive science of X demonstrates that cognitive mechanisms of our evolved minds provide most impetus for believing X. As these mechanisms are error-prone, they cannot be trusted to give us truth. As we cannot trust these aspects of our minds to give us truth, we cannot trust most of our impetus for believing X. Therefore, cognitive science X weakens belief X.
I regard such an argument as yielding not only intolerable collateral damage. In fact, it suffers from a Suicidal Tendency. Not only does it potentially weaken belief in a host of different domains, but it also casts doubt on its own foundations and presuppositions.
As noted above, evolution by natural selection has no regard for Truth. We can only be confident that it has given us minds helpful for survival in our ancestral environments. Further, the cognitive sciences have given us evidence that our minds—for the sake of survival—can be systematically fallible, trading survival and reproduction for accurate representations of reality. On what basis then do we trust our minds at all? Our minds cannot be trusted to tell us that gods exist, that other human minds exist, that our memories are reliable, or that natural laws remain the same from moment to moment, or that cognitive science can produce accurate findings or that evolution is true. The Error-Prone Mind argument proves to be self-defeating—it has a Suicidal Tendency.
The way out of this problem is to reject the claim that our mind and its component mechanisms are too error-prone to be trusted. Do they make mistakes? Sure. But do they make enough mistakes to completely undercut them as sources of beliefs? I hope not. They are all we have; but on this point, the evolutionary and cognitive sciences do not offer any assurances. Natural selection offers no truth guarantees and the cognitive sciences seem to give us plenty of reasons for doubt. One will have to build an epistemological foundation using pre- or extra-scientific timber.
The theist may build such an epistemological foundation by appealing to the divine as a trustworthy source of Truth that has imparted the ability to conceive Truth (at least under some conditions) through cosmic fine-tuning or supernatural selection or supernaturally generated mutations that then were naturally selected to produce human minds. I leave the details and the coherence of such a theology up to the individual theist.
For the moment, it appears that theists have nothing to fear from the bio-psychological explanations of religion. Although these scientific endeavors may grant new insights into the mechanisms that play a role in shaping religious thought and action, whether or not belief in gods is rationally justified remains a question outside of science.
This work was supported by a grant from the John Templeton Foundation.
1 Pascal Boyer, Religion Explained: The Evolutionary Origins of Religious Thought (New York: Basic Books, 2001), 2, 4.
2 Scott Atran, In Gods We Trust: The Evolutionary Landscape of Religion (Oxford: Oxford University, 2002). In keeping with the comparative character of this area of scholarship, I use theism and theistic to refer to beliefs in non-natural intentional agents generally including God, gods, ghosts, and ancestor-spirits. Monotheism and polytheism are regarded as subclasses of theism in this sense. As the arguments here equally apply to belief in any of these intentional agents, I use the term gods as an inclusive umbrella. The Judeo-Christian God, Muslim Allah, Hindu Shiva and Vishnu, and any number of other gods are all objects of the arguments presented here.
3 Daniel Dennett, Breaking the Spell: Religion as a Natural Phenomenon (New York: Viking, 2006).
4 Leon Wieseltier, ’The God Genome,” New York Times, (February, 19, 2006).
5 Eugene G. D’Aquili and Andrew B. Newberg, The Mystical Mind: Probing the Biology of Religious Belief (Minneapolis: Fortress, 1999); Rhawn Joseph, ed., NeuroTheology: Brain, Science, Spirituality, Religious Experience (New York: University Press, 2002); Michael A. Persinger, Neuropsychological Bases of God Beliefs (New York: Praeger, 1987).
6 Michael A. Persinger, “The sensed presence within experimental settings: Implications for the male and female concept of self,” Journal of Psychology: Interdisciplinary and Applied, 137:1 (2003): 5–16.
7 David Sloan Wilson, Darwin’s Cathedral: Evolution, Religion, and the Nature of Society (Chicago: University of Chicago, 2002). Whitehouse suggests that a particular configuration of social morphology and ritual practices he terms the “imagistic mode of religiosity” might likewise produce group cohesion, pro-social behavior, and resultant competitive advantage over other groups; but Whitehouse’s account does not appeal to genetic, adaptationist selection but cultural selection. Harvey Whitehouse, Modes of Religiosity: A Cognitive Theory of Religious Transmission (Walnut Creek, Calif.: AltaMira, 2004).
8 Justin L. Barrett, Why Would Anyone Believe in God? (Walnut Creek, Calif.: AltaMira, 2004); Brian Malley, How the Bible Works: An Anthropological Study of Evangelical Biblicism (Walnut Creek, Calif.: AltaMira, 2004); Robert N. McCauley and E. Thomas Lawson, Bringing Ritual to Mind: Psychological Foundations of Cultural Forms (Cambridge: Cambridge University Press, 2002); Ilkka Pyysiäinen, Magic, Miracles and Religion: A Scientist’s Perspective (Walnut Creek, Calif.: AltaMira, 2004); D. Jason Slone, Theological Incorrectness: Why Religious People Believe What They Shouldn’t (New York: Oxford University Press, 2004); Todd Tremlin, Minds and Gods: The Cognitive Foundations of Religion (New York: Oxford University Press, 2006); Whitehouse, Modes of Religiosity.
9 Jesse M. Bering, “Intuitive conceptions of dead agents’ minds: The natural foundations of afterlife beliefs as phenomenological boundary,” Journal of Cognition & Culture, 2:4 (2002): 263–308; Jesse M. Bering, Carlos Hernândez-Blasi, David F. Bjorklund, “The development of ’afterlife’ beliefs in secularly and religiously schooled children,” British Journal of Developmental Psychology, 23:4 (2005): 587–607; Paul L. Harris, and Marta Gimenez, “Children’s Acceptance of Conflicting Testimony: The Case of Death,” Journal of Cognition & Culture, 5:2 (2005): 143–164.
10 E. Margaret Evans, “Cognitive and contextual factors in the emergence of diverse belief systems: Creation versus evolution,” Cognitive Psychology, 42:3 (2001): 217–266; Deborah Kelemen, “Why are rocks pointy? Children’s preference for teleological explanations of the natural world,” Developmental Psychology, 35:6 (1999): 1440–1453; idem, “Functions, goals, and intentions: Children’s teleological reasoning about objects,” Trends in Cognitive Sciences, 3:12 (1999): 461–468; idem, “Are children ’intuitive theists?’ Reasoning about purpose and design in nature,” Psychological Science, 15:5 (2004): 295–301; Oliveria Petrovich, “Understanding of non-natural causality in children and adults: A case against artificialism,” Psyche en Geloof 8 (1997): 151–165; idem, “Preschool Children’s Understanding of the Dichotomy Between the Natural and the Artificial,” Psychological Reports, 84:1 (1999): 3–27.
11 Justin L. Barrett, Roxanne M. Newman, and Rebekah A. Richert, “When seeing does not lead to believing: Children’s understanding of the importance of background knowledge for interpreting visual displays,” Journal of Cognition & Culture, 3:1 (2003): 91–108; Justin L. Barrett and Rebekah A. Richert, “Anthropomorphism or preparedness? Exploring children’s concept of God,” Review of Religious Research, 44:3 (2003): 300–312; Justin L. Barrett, Rebekah A. Richert, and Amanda Driesenga, “God’s beliefs versus mother’s: The development of non-human agent concepts,” Child Development, 72:1 (2001): 50–65.
12 Evans, “Cognitive and contextual factors in the emergence of diverse belief systems.”
13 Paul Bloom, “Is God An Accident?,” Atlantic Monthly, (December 2005): 1–8.
14 Paul Bloom, Descartes’ Baby: How the Science of Child Development Explains What Makes Us Human (New York: Basic Books, 2004).
15 Bloom, “Is God An Accident?,” 8.
16 Ibid.
17 Thomas Gilovich, How We Know What Isn’t So: The Fallibility of Human Reason in Everyday Eife (New York: Free Press, 1991).
18 Stewart E. Guthrie, Faces in the Clouds: A New Theory of Religion (New York: Oxford University, 1993).
19 If we assume that people and sophisticated animals do qualify as agents, such a strategy of tabulating accurate versus inaccurate agency detections would likely lead us to the conclusion that our cognitive equipment is extremely reliable and not so error-prone after all. Such a finding would encourage rather than discourage confidence in theistic beliefs.
20 Alvin Plantinga, God and Other Minds: A Study of the Rational Justification of Belief in God (Ithaca, N.Y.: Cornell University Press, 1990).
Justin L. Barrett is Senior Researcher at the University of Oxford’s Centre for Anthropology and Mind. He earned degrees in psychology from Calvin College (B.A.) and Cornell University (Ph.D.). He served on the psychology faculties of Calvin College and the University of Michigan (Ann Arbor). Dr. Barrett is an editor of the Journal of Cognition & Culture and is author of numerous articles and chapters concerning cognitive science of religion. His book Why Would Anyone Believe in God? (AltaMira, 2004) presents a scientific account for the prevalence of religious beliefs.