The Cognitive Science of Religion (CSR) has emerged in the last twenty years as a fundamentally new approach to the study of religion. Using ideas from cognitive psychology foremost, CSR has overwhelmingly focused on individual aspects of cognition in trying to explain the recurrence of patterns in religious thought and behavior. In general, this has been a fruitful approach and a pack of ingenious theorists have developed these ideas to study everything from initiation rituals to possession. But for all the utility of CSR, it is time to reiterate criticisms about it being overly essentialistic and too concerned with the individual at the expense of the culture in which the individual develops. Critique can be a useless enterprise, though; the criticism noted here should be considered only in light of an underlying admiration for the field of CSR and the hope of its increased sophistication in the future.

Though CSR has not borrowed the paradigm of natural selection wholesale—as, for instance, Richard Dawkins does in his theory of memes (1976)—it cannot be denied that selection processes remain central to most CSR theories. In "Explaining Religious Ideas: Elements of a Cognitive Approach," Pascal Boyer discusses this framework: "Applied to cultural ideas, the notion of a selective model means that, given certain circumstances and a variety of mental representations entertained by a population of subjects, some of these representations are more likely than others to be stored in the subjects' memories and communicated to other subjects. Selective models generally focus on transmission processes as the main cause of recurrence" (1992:32). What Boyer stipulates here is a distinctly "bottom-up" approach; this model rejects that higher-level processes may affect lower-level processes in favor of lower-level processes effectively determining higher-level processes. In short, many higher-level processes are epiphenomena that need not play any role in the causal loop. But if we take the analogy of CSR to natural selection seriously then the approach that Boyer and others affirms is one that does not truly respect the system of natural selection; a system that involves both generative speciation and ecological selection. By focusing on one to the exclusion of the other, CSR rejects complex causation. Boyer posits a "bundles of features" analysis to be superior to one that favors causal complexity (Boyer 2005, cf. Boyer 1996). This supports the idea that religion, as a whole, congeals around a handful of cognitive attractors; the multitude of details that seem to connect these privileged attractors are little more than "spandrels," to borrow from Stephen Jay Gould and Richard Lewontin (1979). To study the spandrels might be a descriptively meaningful project but does little to elucidate the core elements of a religion. Boyer eschews critiques of CSR that favor holism of one sort or another (Boyer 1996). His only concession to a holistic paradigm is using it as a means to understand a particular religious tradition. In "Religion as an Impure Subject," he writes:

There is in fact only one situation in which it is possible, and even desirable, to add heterogeneous causal factors into an integrated description. It is when we are considering a single historical situation of which we want to identify all the particular aspects. That is to say, we are then doing the "history" of a particular situation. As far as general objects are concerned, adding viewpoints hampers our understanding of causal processes rather than enriching them. (Boyer 1996:209)

I challenge this last statement: "As far as general objects are concerned…" The statement reminds me of another proviso I have seen repeated time and again in the CSR literature, namely: "All else being equal…" To both these, I respond: when is a religion ever general and when is all else equal? If really understanding a religion is doing "history" then do we give up on the idea of cognition being involved once we turn to the specific details of a given religion? I don't find this a reasonable point of view. Understanding a religion and being able to relate it to other cultural processes is not just history; this quest lies at the heart of contemporary cognitive anthropology. If we continue to use natural selection as an analogical framework for CSR then we need to consider it seriously; as it is it seems as if CSR would remove the "nature" from natural selection. For instance, to favor something like a minimally counterintuitive concept (MCI), which owes its optimized transmission to increased attention and memorability, is to hold in check the ecology which selects the MCI. There is logic behind this, of course, namely that the ontological categories which an MCI "tweaks" were fixed in the human nervous system sometime during the Pleistocene (Tooby & Cosmides 1992). But to overemphasize the microprocess of an MCI being selected based upon these ontological categories is to effectively ignore the larger, and much more contemporary, ecology of representations.

An analogy to this would be something like the co-evolution of the coffee plant and the insects which fed upon it eons ago. Over time, based upon this single co-evolutionary relationship, the speciation of the coffee plant led to an organism with a high content of the xanthine molecule known as caffeine which caused those insects that fed upon the plant to bug out. And indeed—if we hold all else to be equal—this might be the end of the story; the coffee plant stochastically happened upon a strategy to prevent insect infestation. But of course that's not the end of the story. A curious evolutionary byproduct effectively changed everything. A certain primate species chanced upon the coffee plant as a way to exacerbate its own anxious mentality. The consequence of this co-evolutionary relationship was the cultivation of the plant on a scale unimaginable were the plant to remain in its original ecology. Natural speciation transformed, in a flash, into artificial speciation. To "explain" the coffee plant by reducing its causal history to the co-evolution of its caffeinated strategy vis-à-vis insects is correct. Undoubtedly, the reason the coffee plant survived to the present can be found in this ancient tale of bugs and plants. But to ignore the integration of coffee into human culture, to reject its epiphenomenal "second-life," a life that helped bring about the Industrial Revolution and endless iterations of colonial projects, would be quite sad and certainly lead to an enervated version of ethnobotany.

Moreover, there is something essentially antiquarian about this whole take on MCIs and their relationship to our Pleistocene developed ontological categories. I really can't imagine Julia Childs, for instance, being content to explain thousands of years of developed gourmet delicacies from India to France as the mere combination of salt, fat, and sugar which our remote ancestors evolved to appreciate, though that is certainly an important part of the story. Nor can I imagine a sommelier admitting that his knowledge of wine is of no essential consequence since its crucial ingredient, ethyl alcohol, can be found as readily in moonshine and Schlitz beer as it can in a good Bordeaux. While it is important to recognize that ethyl alcohol universally inebriates human beings and that without this property alcoholic beverages would not have spread so readily, we would be remiss to reduce drinking customs, bars, Catholic Communion, and alcoholism to this property alone. Again, this is not to deny the very interesting reductionism that CSR has favored so far, it's just to say that CSR's purview need be expanded to some higher levels of analysis. We need, at the very least, a good "middle-range theory."

In my opinion, Jesper Sørensen has pointed out the direction for the reconciliation of CSR and cultural models theory with his notion of an immunology of cultural systems (2005). This approach responds to and elaborates upon Sperber's earlier work. In "Anthropology and Psychology: Towards an Epidemiology of Representations," Sperber wrote that "Cultural phenomena are ecological patterns of psychological phenomena" (76). This was an exciting paradigm, one that seemed to echo Bateson's ideas about an "ecology of mind" (1972). But instead of progressing with an ecological approach, Sperber went on to use epidemiology as his dominant heuristic. This was unfortunate since it emphasized the transmission of representations over and above the selection with which transmission is coupled in an ecological system. Epidemiology, after all, relates to the spread of microbes, so this framework inherited a focus on the spread of representations even if Sperber attempted to distinguish his notion of representations from simple meme-like representations. Sørensen's immunological perspective clarifies how representations, like microbes, must come up against co-evolving challenges to their proliferation. Animals possess dynamic immune systems which readily evolve to counter novel microbes. Similarly, minds and cultures, or "socio-cognitive systems" to use Malley's phrasing (1995), possess mechanisms which dynamically respond to novel representations with the same alacrity. Sørensen's immunology of cultural systems meets Sperber's epidemiology metaphorically, but it also opposes many underlying assumptions of CSR. He writes: "My two objections can be summarised in (a) a rejection of the notion that culture is only a statistical phenomena without any causal efficacy, and (b) a rejection of the epiphenomenalism involved in the argument that all aspects of culture and religion can be explained by reference to cognitive processes of individuals alone" (Sørensen 2004:62). What Sørensen suggests, in stark contrast to the assumptions of CSR, is that cognition needs to be defined in a more expansive way. Cognition is not individual minds transmitting and receiving representations, or even transmitting and reconstructing representations pace Sperber, but something much larger than individual minds altogether. Perhaps the definition Jean Lave offered in Cognition in Practice is most relevant:

The point is not so much that arrangements of knowledge in the head correspond in a complicated way to the social world outside the head, but that they are socially organized in such a fashion as to be indivisible. 'Cognition' observed in everyday practice is distributed—stretched over, not divided among—mind, body, activity and culturally organized settings (which include other actors). (1)

Human cognition is not the Cartesian mind, it is the workings of an individual nervous system—body included—within a very robust and extensive cognitive ecology. Much of our memory, for instance, can be located in extrasomatic sources or depend upon them as mnemonic cues. Representations are found in internal processes, artifacts, external processes and institutions, and achieve their patterns of distribution based on a complex interplay of their appeal, their transmissibility, and their compatibility with extant representations. With this understanding of cognition it becomes clear why an immunological approach is necessary. Immunology erases the "all else being equal" proviso and encourages the study of the cognitive ecology already in place. Sørensen offers an entirely new level of analysis: …an immunology of cultural systems will propose a third selective principle based on the degree of fitness of new cultural representations into what has been referred to as the existing ecology of representations (Malley 1997), in short, a selection based on how compatible new representations are with already existing public and mental representations organised in systems of mutual reference. (2004:62) Sørensen's immunology is the meaningful antithesis to Sperber's epidemiology. I should hope the two would lead to a synthesis of ecology as the dominant metaphor as Bateson initially suggested.

To clarify the immunological stance, I would offer the example of European religious history. In spite of an always prolific human imagination and in spite of trade networks that connected Europe to many other lands, Christianity remained the unchallenged institutionalized religion of the European continent for nearly two millennia. CSR can respond to this in one of two ways; either Christianity in Europe and, say, Mahayana Buddhism in China are essentially the same, both having evolved cognitively optimal concepts like beliefs in saints, miracles, and the like; or these two traditions are indeed quite different but their difference does not relate to meaningful cognitive processes but to historical contingency. I think these two traditions are quite different in spite of sharing some cognitively optimal concepts and practices but I also think that each faced true challenges which given an "all else being equal" playing field, would have transformed each of these traditions altogether. In fact, I believe these traditions developed institutions (councils of bishops, the Inquisition, theological schools, etc.), external representations (holy books, icons, and the like), and tied themselves into culturally specific models which helped to create a persistent tradition in the face of internal and external challenges. In short, each tradition—though not being monolithic by any means—developed immunological techniques to contend with novelty and innovation.