The Zeus problem revisited – or is it the Jedi problem?

In their recent paper (available here) in Journal of Cognition and Culture, Will M. Gervais and Joseph Henrich call attention to the Zeus problem. If religious belief is solely guided by representational content biases (as many scholars in the cognitive science of religion have argued), why do people generally not come to believe in the gods of their neighbours, or indeed, in gods of the past such as Zeus? Zeus has all the features that are characteristic of successful religious agents, but he is no longer a target for widespread belief and commitment. Of course, what Gervais and Henrich do not mention is that there are in fact modern believers in Zeus and other members of the Greek pantheon, namely adherents to Hellenic Polytheistic reconstructionism. As can be seen in the movie here, Zeus is still an object of worship today. There are about 2000 adherents to this form of paganism in Greece today.



So is there in fact a Zeus problem? I am not so convinced, since it turns out that even religions that make no secret of their purely fictional origins are quite successful.



Carole Cusack's recently published Invented Religions. Imagination, fiction and faith (2010, Ashgate) provides a thorough and fascinating overview of religions that have been invented in the course of the 20th century. In several cases, these religions have their roots in fictional works, including Jediism (which is based on the fictional religion in the successful Star Wars franchise) and the Church of All Worlds, which is based on a fictional religion that was invented by Robert Heinlein in his novel Stranger in a strange land.




Although it remains unclear what the long-term success of these invented religions will be, at present they do have adherents. What explains the success of religions that do not invoke revelation, a continuity with older traditions, or holy books (as other recent religions such as Mormonism)? Perhaps content biases can provide a compelling answer. As Cusack (2010: 4) mentions "In the case of the Church of all Worlds, its founders thought Robert A. Heinlein's Stranger in a strange land was a fiction so good it should actually be true".

It would be interesting to pit the religious agents from invented religions against Justin Barrett's (2008, Why Santa Claus is not a god. Journal of Cognition and Culture, 8, 149-161) five features that according to him are typical of representations that have godly status – i.e., they violate ontological expectations, they are minimally counterintuitive, they possess strategic information, they have detectable interaction, and they have representational content that motivates ritual and other religious practice.


  • comment-avatar
    Olivier Morin 26 November 2010 (16:32)

    I am not sure that so many scholars have argued that “religious belief is solely guided by representational content”, but if they have, well, their position seems obviously wrong to me… Who would deny that the kind of religious beliefs you have access to, is the most important factor in determining which religious beliefs you end up adopting? In a world where Judeo-christian beliefs are widespread, people are more likely to adopt them, not necessarily because they are conformist, but because you choose from the pool of existing beliefs. It seems just as obvious to me that this dynamic would make some religions spread much more than others, even in the absence of cognitive factors, through a series of historical accidents. The same applies to flu epidemics: some are very sucessful, some are not. Some varieties of flu are more contagious than others, but this does not explain the huge differences in success between different epidemics. Chance does. So, yes, of course, Zeus is much less popular than Jesus. And, no, fictional religions are not successful if you measure success by the standards of today’s big religions. And I can see no obvious link between the cognitive properties of Zeus beliefs and their current lack of success. I thought Gervais and Henrich were attacking a straw man when they made their point, but maybe I was wrong.

  • comment-avatar
    Helen De Cruz 26 November 2010 (19:30)

    Of course, Olivier, the pool of religious ideas you have access to (next to more specific conformist bias) is perhaps the most important factor in the epidemiology of religious representations – that is after all why 98 % of Greeks today are members of the Christian Orthodox church, and not of Hellenic neopaganism. But I do think that representational content biases can have some explanatory power in accounting for why some new religions are more successful than others (perhaps there are other factors in terms of biased cultural transmission, such as prestige bias which might play a role in the success of scientology). What might the allure be of invented religions? Cusack pointed out that people who are members of such religions are typically quite individualistic; they try to find a religion that ‘works for them’, i.e., that is somehow appealing in its ideology. This looks like content biases at work.

  • comment-avatar
    Jesper Soerensen 30 November 2010 (22:28)

    Even though I agree with Olive Morin that the standard cognitive explanation does not exclude other factors (its an all else being equal-argument), Gervais & Henrich do point to a important problem often alluded to, when you present cognitive theories to scholars of religion. One question, often appearing, is why certain ideas successfully spread in new areas, whereas others fail miserably. I agree that reputations systems involved in tracking the source of reliable information necessarily must be part of the answer, but here I would like to point to another way of understanding this problem. A problem in the standard account seems to me to be the, relatively unfounded, conceptual atomism involved. The epidemiological spread of gods such as Zeus, Jesus, Ganesh is explained as a result of the internal content, i.e. the cognitive optimum, totally independent of their relation to other concepts. I believe this is flawed for a number of reasons. First of all, concepts are not remembered independently but in relation to other concept that gives them a locally embedded, pragmatic meaning. Thus, ’Jesus’ has an altogether different meaning when uttered as a curse than when said by a priest during mass. Further, the evocation of one concept will potentiate the evocation of other concepts as a simple priming effect. ’Jesus’ will give rise to a number of predictions as to the following relevant concepts, and thus automatically relate a concept to other concepts. So the argument would be that concepts form relatively stable clusters, and that the adoption of a concept depends on the relative strength of related concepts forming such pragmatically relevant models. Let me give an example: Lord Krishna has been proselytized intensively in Western Europe by ISKCON (also known as Hare Krishna) for almost 40 years but with a very low rate of success. In contrast, the equally Indian concept of reincarnation has become hugely popular despite ardent opposition from majority churches. Why? Besides the reputation systems obviously at work tracking the source of new ideas, the two concepts relates very differently to other prevalent conceptual structures in the Western world. Whereas reincarnation got hook up with positive concepts of progress, evolution and individualism when introduced in the Western world by theosophist in the late 19th Century, Krishna has no such conceptual underpinning, and adopting Krishna as your god with full commitment will force you to give up on a long series of other concepts and introduce wholly different ones instead. This is not only cognitively burdensome, but it will also automatically segregate you from other individual not sharing your new conceptual structure. Thus in the first case we find a smooth transition allowing for a new concept to emerge as it slightly alters its content to fit into the existing ecology of ideas, in the other case the concepts demand a whole new conceptual ecosystem to attain everyday relevance – a relevance that will be countered in numerous situations due to its strangeness. In short, the epidemiology of representations is held in check by an “immunology of representations” that stabilizes clusters of concepts in structures of cross-reference. I have presented this argument more fully in the now defunct journal Evolution and Cognition (

  • comment-avatar
    Konrad Talmont-Kaminski 4 December 2010 (16:03)

    I might just point out that while Evolution and Cognition was discontinued under that name, the Konrad Lorenz Institute is publishing a journal – Biological Theory – that is the continuation of that journal. Indeed, just earlier this year Atran and Henrich published in it a paper on their dual inheritance account of religion: