Darwin’s Cathedral versus The God Delusion

Awhile back I read Richard Dawkins’ The God Delusion and more recently I finished David Sloan Wilson’s Darwin’s Cathedral. Both books provide a view on religion from the perspective of a prominent evolutionary biologist, and the contrast between these views tells us a lot about the culture of evolutionary biology as well as the nature of religious culture.

Given that both authors are scientists, it is not surprising that each book comes with its own working hypothesis. Wilson’s hypothesis is that religion, like many other cultural technologies (language, tools), is an adaptive product of evolution by natural selection whose function is to facilitate cooperation at larger scales of the human population. Dawkins’ hypothesis is that religion is a parasite on the human mind that does far more harm than good and ought to be replaced by other forms of culture, in particular scientific rationality. From the outset the difference between these hypotheses should make it clear that these books take very different approaches to the “problem” of religion. Wilson’s hypothesis says something potentially positive about religion (although I will argue below that this positivity is quite limited) while Dawkins’ take is unequivocally negative. But more importantly, the hypotheses are of a very different nature. Wilson’s is a purely scientific hypothesis, one whose predictions he attempts to test via comparison to anthropological and sociological data. Although underpinned by a variety of scientific theories and some data, Dawkins’ hypothesis isn’t really scientific: his argument is fundamentally political.

There are several reasons for this fundamental difference between Wilson’s and Dawkins’ approaches to religion. First, I think that one can argue that Wilson has remained more of a scientist throughout his career (despite becoming a popularizer of scientific ideas of late) whereas Dawkins has thoroughly committed to his role as (self-appointed?) ambassador of evolutionary biology. As such Wilson is still looking at hypotheses at a smaller scale, even when they tie in to larger theories about the way that natural selection works. Dawkins, on the other hand, has spent the last three decades providing a theoretical overview of how evolution works, really at a more philosophical than scientific level. While he often comes up with some pretty interesting ideas, those that are not obvious conclusions drawn from synthesizing existing scientific work can only be validated if we scale down to more testable hypotheses.

We can see this difference in the approaches of each book to religion as a cultural trait.
Wilson’s fundamental approach to religion is to ask whether or not it serves some purpose. Although he does address the fact that some aspects of religious belief are superstitious, he does not assume that ideas that are irrational from a scientific perspective are necessarily maladaptive. In fact, Wilson gives a pretty healthy “benefit of the doubt” to most religious contentions, assuming that if they have survived for long periods of time they probably have some function. Dawkins takes a much more narrow view of religious culture, spending a good portion of his text making a strong argument against the scientific/rational validity of religious ideals, most prominently the existence of an omniscient god or gods. For Dawkins, cultural traits that are inconsistent with scientific rationality and logic are assumed to be maladaptive.

The debate over whether to assume that all traits are adaptive has raged within the field of evolutionary biology for decades, and its most famous argument was posited by Stephen J. Gould and Richard Lewontin in a 1979 paper flamboyantly entitled “The spandrels of San Marco and the Panglossian paradigm: A critique of the adaptationist programme”. In the “spandrels” paper, Gould and Lewontin argued that evolutionary biology had fallen into a default assumption that all traits observed in organisms serve an adaptive purpose, making the chief task of scientists to uncover the nature of these purposes. They argued instead that some traits are the byproducts of other traits (themselves adaptive), and that the assumption of the “adaptationists” should be dropped. I think this is a valid point, and perhaps the culture of evolutionary biology needed to be pushed away from unproven or unprovable “just-so stories” that carried the science of that day. But the “spandrels” observation in of itself does not do much to advance our understanding, as it does not suggest how to differentiate between adaptive traits and byproduct traits through hypothesis testing.

Wilson’s view of religion is clearly that of the adaptationist. He assumes that a trait as prevalent and historically-ancient as religion must have an adaptive value, in particular because it comes with so many costs. In contrast, Dawkins clearly views religion as a byproduct of other evolved traits. In fact, Dawkins’ view of religion belongs to a special category of byproduct explanations, those that argue that a trait is in fact parasitic. Not just a neutral consequence of other adaptive neurological traits, religion is instead a costly cultural parasite that capitalizes on existing (presumably adaptive) features of our brains. Cultural “traits” are particularly amenable to this sort of “parasite argument” because culture can evolve so quickly, thus allowing cultural ideas (just like parasitic organisms) to evolutionarily outrun the defenses of their human hosts. These sorts of byproduct arguments are quite fashionable these days: in another prominent example, Steven Pinker has argued that creative expression and all forms of the arts are a byproduct of our neurological architecture rather than adaptive cultural “tools”.

In the context of the adaptation versus byproduct argument there lies a burden of proof problem. Whose burden is it to prove their hypothesis? What Gould and Lewontin brought to light was that many hypotheses positing an adaptive value of a particular trait were being accepted without being tested, and the science of evolutionary biology has been improved by their demand for greater rigor. Now, in order to support the idea that a trait is adaptive, we must test the prediction that its benefits outweigh its costs. Countless studies have provided this sort of hypothesis testing for a great variety of traits. But what about hypotheses that suggest that a trait is a byproduct of another trait, and especially hypotheses that suggest that a prevalent trait is actually maladaptive? How is the burden of proof distributed in this case?

I will skip the more difficult case of providing support for the presence of neutral byproducts and focus on the kind of hypothesis that Dawkins posits: that a trait is maladaptive and therefore parasitic. It seems to me that the burden of proof here is as heavy as that which must be lifted by the adaptationist. If we believe that a cultural trait is parasitic, we need to show that there are higher costs than benefits associated with possessing the cultural trait and then show one of two outcomes in groups of human “hosts”: either the trait should quickly die out (suggesting a temporary susceptibility to the trait followed by acquired immunity) or those “infected” with the trait should do more poorly than those without it (suggesting variation in the susceptibility of the population to the parasitic trait). Notice how these two cases perfectly track the two extremes of biological parasites: they can either be relatively low-impact parasites that exact short-term costs and mutate frequently (like the common cold) or pandemics that wipe out large segments of the population, radically transforming the species (like the plague). The science of epidemiology has already provided very strong theoretical and empirical methods for tracking different kinds of infections, and there is no reason why such methods cannot be retrofitted to apply to the spread of cultural traits. Strangely, this kind of hypothesis testing does not seem to be applied to the current crop of “parasitic byproduct” hypotheses used to explain prevalent human cultural traits like music, art, and religion. To my knowledge, no one has formalized an argument against the unfounded use of the hypothesis that cultural traits are parasitic on their human hosts. It seems that we are still waiting for another pair of evolutionary biologists as influential as Gould and Lewontin to come along and write a metaphorically-rich treatise against the “parasitic byproductist programme”.

So how do each of these authors approach testing their very different hypotheses about religion? In Darwin’s Cathedral, Wilson uses a two-pronged strategy to test the predictions made by his hypothesis that religion is a group-level adaptation. First, he uses anecdotal evidence to suggest that many religions appear to be “designed” to minimize internal competition and maximize the competitive advantage of the group, a pattern consistent with what we would predict about a group-level adaptation. His chief case study is of Calvinism, and he makes a pretty compelling argument that the characteristics of this religion were designed to benefit the group as a whole and prevent abuse of power among church leadership. Two additional chapters on the “Secular Utility of Religion” go on to provide historical and present-day anecdotal accounts of religion’s instrumental value to members of its group. While we often frown upon this sort of “story telling” in the sciences, these anecdotally-supported trends are usually the precursors to more rigorous tests of adaptationist hypotheses. An appreciation for natural history leads to more rigorous models of what mechanisms underlie that natural history. Wilson acknowledges that making the transition from analyzing anecdotal trends to testing hypotheses about the group-level instrumental value of religion is difficult, but he makes the attempt. In Darwin’s Cathedral he describes (albeit inconclusively) his process for applying the comparative methods of evolutionary biology to a database describing world religions. This analysis was completed and published later in Human Nature as “Testing major evolutionary hypotheses about religion with a random sample”. Even in this more completely realized analysis Wilson admits that further work is required to validate his hypothesis, but he has clearly shown the way in which such work can be completed, opening the door for further scientific inquiry.

Dawkins’ approach similarly uses anecdotal evidence to support his hypothesis that religion is a parasitic byproduct of other adaptive features of the human species. The “higher level” nature of Dawkins’ hypothesis is revealed by the fact that he actually entertains an entire family of hypotheses that explain religion as a byproduct of other functions. This plurality makes for interesting thought experiments but doesn’t do much to narrow down the reasons why religion parasitizes us. Is religion taking advantage of our tendency towards adaptive irrationality, or is it our tendency to fall in love that makes us worship a fictional god? Does our “intentional stance” towards the world (Daniel Dennett’s idea that we ascribe intentionality to organisms and sometime objects) prime us for religiosity, or is it that we are vulnerable to religion’s promises of eternal life because we are the only creature on the planet capable of the abstract thought necessary to consider our own mortality? Are these mutually exclusive hypotheses that can be distinguished by their distinctive predictions or should we treat these as equivalent hypotheses all supportive of Dawkins’ meta-hypothesis? Dawkins never takes on the burden of answering these questions; nor does he bother to even suggest much less show us how these hypotheses might be tested. The byproduct argument is just that: an argument, rhetorical and questionably testable, perhaps not even scientific. This is not to suggest that these hypotheses never could not be tested via scientific methods, it is just that Dawkins does not engender any confidence that he knows how to test them.

I gave Wilson the benefit of the doubt for using a “natural history” approach to providing anecdotal evidence for his hypothesis, but it is hard to extend this same allowance to Dawkins because he has done so little rigorous analysis of his major hypothesis. If religion is a parasite, it needs to fit somewhere along the continuum that I laid out above. Is religion like the common cold, a temporary burden that manages to mutate frequently enough to hang on to our attention, albeit in an ever-fluctuating form? Or is it more like the plague, massively infecting and therefore selecting out a large segment of the population? Given how much he discusses the social costs associated with modern-day religion and advocates for its extinction, I am guessing that Dawkins would support the latter scenario. But the fact that we don’t see major die-outs of religions historically, and that many religions have persisted for hundreds of human generations, makes this idea pretty inconsistent with the available anecdotal evidence. And if religions are more like the common cold, where are the constant mutations that enable it to re-infect those who gained immunity to its previous forms? Wilson’s account of the commonalities of religion make a pretty good first-approximation argument against both of the possible extremes of Dawkins’ religion-as-parasite argument. If Dawkins has bothered to think this rigorously about his hypothesis it isn’t apparent in the haze of his rhetoric.

Another clear distinction between Wilson and Dawkins, well-established in both of their other writings, is in the levels of selection they consider reasonable to consider in relation to the evolution of religious culture. Wilson’s explanation is fundamentally group selectionist, and he makes no apologies for this. He suggests that the major function of religion is to scale up the size of the cooperative in-group from a small tribal scale to a much larger and more complex regional or even global scale. The egalitarian ethics of small tribes has been well-documented, but the scale of this mechanism for suppressing intra-group competition is limited. Religions allow for the scaling up of these mechanisms, often through the propagation of ideas that seem irrational or scientifically unverifiable. In his anecdotal and comparative analyses, Wilson discusses a great number of religious beliefs and values that seem to have the clear function of suppressing competition with other believers. He also points out that this hardly removes competition from human society, it simply increases the scale at which competition is most strong, an observation that elegantly explains the prevalence of large-scale religious conflict in our recent history.

Dawkins does not believe that group selection has played a significant role in evolution. Belonging to a large group of evolutionary biologists who feel that group selection was dispatched by George C. Williams and others in the 1960’s, Dawkins is famous for advocating that selection occurs at the level of the gene rather than the individual or any higher level of organization. Dawkins does mention — in passing — explanations of religion based on higher-level selection, but he dismisses them without much substantial argument. He even briefly mentions Darwin’s Cathedral but does not directly engage any of its arguments, derisively referring to Wilson as the chief “apostle” of group selection. That Dawkins carries a serious bias for a particular level of selection is not news, but in the context of discussing religion this bias presents him with a difficult challenge. He has no intention of ascribing to religion the ability to suppress selection within large social groups, both because he doesn’t believe that group fitness matters and because he doesn’t see religion as having such a benign effect. But he also needs to explain why human beings don’t spend all their time ripping each other to shreds in a competitive orgy.

What he suggests is that human beings have a strong moral side that is inborn, that we are kind because of our human nature and not because of the culture that we live in. To a point this seems reasonable, and seems to be consistent with cross-cultural studies of social behavior. But if selection on the group is not responsible for the prosocial features of human nature, what is? For Dawkins, it is reciprocity, the “I’ll scratch your back if you scratch mine” exchange seen in many of our mammalian relatives and even some non-mammalian species. Dawkins suggests that our self-interest motivates us to rationally choose to play nicely with others. He suggests that the invention of money was an important cultural technology facilitating a flowering of reciprocity in larger societies. This also seems reasonable to a point, but in order to really explain the benefits of scaling up reciprocity we need to also consider the importance of the rule of law, the ethics of equality and justice, and the policing provided within a society. Why these should not be considered group-level traits allowing some societies to outcompete others is not clear, and the distinction between secular rules designed to prevent intra-group cheating and religious rules designed to prevent intra-group cheating is never made. Again, we see a explanation by Dawkins that is poorly elaborated.

Given these vast differences in their approaches, is there any point in trying to compare these books? I think so. Although you can probably guess that I think that Wilson’s analysis is a lot more valuable than Dawkins’, I appreciate both books — especially when placed alongside of each other — for their power to illuminate the key issues that challenge any explanation of religion made from an evolutionary perspective. While one has to wade through a lot of diatribe to get to Dawkins’ main points, they are an important part of the larger discussion of religion’s origin and function.

Clearly the overall purpose of each book is different. As I stated from the outset, I believe that Dawkins’ main goal in writing The God Delusion is political. Although he couches his argument in (supposedly) scientific terms, he really is making an argument for dropping religion, perhaps one believer at a time. This is Dawkins at his best as a meme warrior, trying to spread the ideas that have captured his mind’s attention. Unfortunately, Dawkins pulls a blunt political pronouncement from his under-supported hypothesis. If science is meant to be in service of his politics, Dawkins needs to do a little more to contribute to the science that might inform our understanding of religion. In contrast, Wilson’s primary goal is to contribute to and forward the science aimed at explaining religion. Some may suggest that the hypotheses that he forwards are heading off in the wrong direction, but that sort of disagreement is a part of healthy scientific debate. Wilson also does not offer suggestions for how we might translate his hypothesis into social policy, in part because the prognosis is mixed: his understanding of religion suggests that it has positive prospects for intra-religious relations but perhaps dire predictions for inter-religious relations. But this challenge at least comes with information and understanding, and should Wilson’s hypotheses be further validated, the resulting scientific knowledge could be used to better inform social policies.

I think that both Wilson’s and Dawkins’ arguments uncover a fundamental problem that has not been addressed in a satisfactory manner by either the natural or social sciences. This problem is the problem of scale. What is clear is that human societies have scaled up immensely in the past ten thousand years, expanding from small tribal groups to complex, nested, international sets of allegiances. The modern human belongs to many groups, including families, local communities, religious orders, and states at various scales. What is the role of religion at these various scales? Clearly it can, in of itself, define a particular scale: all members of any particular religious order claim affinity with an international group of fellow believers. But religion also mediates the exchanges that occur at other scales, both those smaller (such as family or community interactions) and larger (such as the laws of countries). What is the function of religion at each of these scales?

Part of Wilson’s idea that religion increases the scale of cooperation is that human nature needed a little prodding to get beyond the very local scale. If Dawkins and Wilson agree on something, it is probably that “human nature” has many built-in mechanisms to facilitate mutual exchange at the very small scale of a local community or tribe. Wilson suggests that religion allowed for greater cooperation, competing larger groups against other larger groups as well as those humans whose cooperative behavior remained limited to the most local scale. But since the development of the cultural technology of religion, other technologies have also developed which massively scale up our interaction. The governing mechanisms that allow for the formation of countries and global markets both create a level of organization (and perhaps competition) that at least vies with — if not supersedes in scale — the organization created by religion.

I do not claim to have a full handle on how religion functions at these various scales, but I do wonder whether its functions vary from scale to scale. At the very small scale religion is not necessary, as the reasons for behaving cooperatively with one’s family or local community are self-evident, so much so that we probably have features hard-wired into our brains that compel us to be pro-social at a local level. But the benefits of local cooperation are limited, and larger-scale cooperation can yield far better survival and reproductive outcomes for individuals and so (if Wilson is right), religion was invented to scale up cooperative behavior. It is a cultural technology which augments already-existing human behaviors in the service of mutually-beneficial exchange. But what about the function of religion within the larger organizational scales imposed by countries and global markets? Is it possible that once we move beyond the intermediate scale of cooperation empowered by religion that religion itself ceases to serve its positive purpose? Are Dawkins’ claims that religion is a “parasite” valid within the context of our complex modern society? I think these are questions that require further investigation.

Anyone who is religious in an orthodox vein — who takes religious ideals as literal truth — will dislike both these books, but each book has its own way of offending. Dawkins’ offense is pretty clear: he believes that religions are wholly negative and ought to be rapidly phased out from human culture (through a revolution of rationality, not a firestorm). He goes so far as to suggest that teaching children religious “truths” is abusive, as it interferes with their ability to think rationally. Oh, and he also goes to great lengths to disprove the existence of any god. Dawkins goes so far in his critique that one certainly would not have to be extremely orthodox in order to be offended by him.

Wilson’s analysis is much softer, but still should offend the orthodox. Although he does not question the validity of spiritual beliefs such as a belief in a god, he also uses his analysis to reduce religious practices to their instrumental value. Why do religious people behave morally? Wilson would argue that the proximate reason is that religious adherents hold cultural values which make them believe in a higher power and/or a set of universal moral principles, and these spiritual values compel them to behave morally. But the ultimate reason for their adherence to religious values is far more secular and completely of this world: Wilson believes that religion serves the function of increasing the competitiveness of adherents by increasing their ability to cooperate. Although also benefiting other adherents along the way, religious belief as a cultural trait ultimately comes down to the self-interest of the religious adherent, and it is imagined as a creation of the human mind. For those who believe that they are surrendering themselves to God’s plan or to some universal goal, this explanation of religion has to seem sterile, flat, and maybe even offensive. Add to this Wilson’s proper observation that religious cooperation simply scales up the competition between people, often leading to more bloody conflicts than one might see within a smaller group, and his take on religion is probably not one that true adherents will embrace with open arms.

For those of us without an ax to grind in religious debates, the origin and function of religion still must be confronted. Whether religion represents some set of universal truths which scientific analysis can never uncover, serves a utilitarian purpose for its adherents, or is the largest cultural parasite to ever infect the human population, we need to consider its reason for being. What Dawkins and Wilson both clearly and rightly acknowledge is that religion is far too influential a force in human culture to simply ignore.