Religious beliefs: Matter of fact or of preference?

In the public sphere, religious beliefs are often considered to be a matter of private sentiment or preference, not as matters of fact. While this may be helpful for the maintenance of a pluralistic society, religious individuals often regard their beliefs as true in an objective sense. Attempts to incorporate fictionalism into religious practice, such as the Anglican Sea of Faith, have met only with limited success.


There is thus a tension between the large diversity of religious beliefs, which prompt a more subjectivist understanding, and the appraisal by individual religious believers, who seem to have a more fact-like understanding.

How do we intuitively conceptualize religious beliefs? In an article entitled “The Development of Reasoning about Beliefs: Fact, Preference, and Ideology” (forthcoming in the Journal of Experimental Social Psychology [1]) Larisa Heiphetz, Elizabeth Spelke, Paul Harris, & Mahzarin R. Banaji investigated how children and adults view religious doctrinal and faith statements. They made a psychological distinction between three kinds of beliefs: factual beliefs (beliefs concerning states of affairs, of things that are believed to be true in some objective sense); preference-based beliefs (incorporating cognitive appraisals, and varying across individuals and contexts), and ideology-based beliefs (such as religious beliefs) which contain elements of both fact and preference.

To investigate how children regard religious beliefs, Heiphetz et al presented children of various ages (youngest 5 years) and adults with disagreements about these three kinds of statements. For example, a disagreement on a fact was: “this child thinks that germs are very small and this child thinks they are very big.” They were then asked whether only one of these two children could be right, or whether they could both be right. They also used disagreements about preferences, and about religious doctrines and faith statements (e.g., “This child thinks that there are many Gods, and that child thinks there is only one”).

Interestingly, participants of all ages were more likely to respond “only one child is right” when asked about factual beliefs, rather than religious beliefs, but they still responded “only one child is right” more frequently for the religious beliefs compared to preferences. So, all participants judged religious beliefs to be less fact-like than facts, but more fact-like than preferences. Even the youngest children (five years old) already discriminated religious beliefs from facts and preferences.

To rule out the role of background information, a second experiment presented participants with beliefs in the three categories of inhabitants of a fictional planet. The results were very similar to those of experiment 1.

In sum, concerning matters of religion, the participants were not entirely subjectivist or entirely objectivist, but intuitively regarded such beliefs to fall somewhere in between subjective preference and objective knowledge of facts.

These results conform well with Paul Harris’ earlier research on how children regard religious testimony [2]. Even young children (age 5) seem to think that religious testimony is different from factual testimony. They are, for instance, more confident that germs exist than that angels exist (even religious children), despite the fact that both are invisible, and that both are learned about through testimony.

Harris (in a lecture of his I attended at Oxford last spring) speculates that this may be due to children’s receiving subtle cues from adults that there is something distinct and not entirely fact-like about religious beliefs. For instance, adults can be heard saying “I really believe God exists” but not “I really believe germs exist”. Or they may be sensitive to the greater amount of disagreement about religious beliefs, compared to factual beliefs in their community. Also, perhaps religious beliefs do not figure in explanatory contexts in the same way as factual beliefs do (a child may regularly hear “wash your hands to get the germs off” in a matter-of-fact manner, whereas religious explanations “well, it was God’s will” are perhaps only common in the most religious households). Perhaps disagreement itself can be a guide for perceived objectivity.

In this respect, it would be interesting to replicate Heiphetz et al.’s study with people from rather isolated, religious communities, like the Amish. My prediction is that people from such communities would regard religious beliefs as more fact-like than preference-like.


[1] Heiphetz, L., Spelke, E. S., Harris, P. L., & Banaji, M. R. (2013). The development of reasoning about beliefs: Fact, preference, and ideology. Journal of experimental social psychology49(3), 559-565.

[2] Harris, P. L., Pasquini, E. S., Duke, S., Asscher, J. J., & Pons, F. (2006). Germs and angels: The role of testimony in young children’s ontology. Developmental Science, 9(1), 76-96.

5 Comments

  • comment-avatar
    Dan Sperber 9 October 2012 (12:15)

    Thanks Helen, for this very interesting post. I had argued a long time ago (in “Apparently Irrational Beliefs”, 1982) that religious beliefs are not treated as simple facts by believers -if only because their factual content is often relatively mysterious or, more technically, ‘semi-propositional’ – and involve a metarepresentational commitment to the belief itself that can take a variety of forms. The study of Heiphetz et al. that you discuss provides some experimental evidence of the range of attitude to religious beliefs and is very welcome. Two points however:

    1) The authors somewhat flatten the issue by proposing the single dimension fact-preference. All the possible metarepresentational attitudes that underlie religious beliefs do not sit equally well on that dimension. Consider, for instance, a commitment to the belief based on a sense of duty to one’s community rather than on a preference in the ordinary sense of the term. A believer might feel that people in another community have, in some sense, ‘preferable’ beliefs, and nevertheless prefer following her own community’s commitments.

    2) Your suggestion that people like the Amish might differ from the participants of the experiment reported is indeed insightful. I would like to extend the experimental challenge to anthropologists around the globe and have them both address the issue in their ethnography and help themselves in doing so by means of experiments not necessarily identical to those you discusss, but drawing inspiration from them.

  • comment-avatar
    Helen De Cruz 9 October 2012 (14:17)

    Hi Dan,

    With respect to your (1) I think that this is a very good idea, and I think it is particularly pertinent to ideology and religion. I am currently involved in an ongoing study where we look at deference to religious beliefs, and we find a strong effect of church attendance frequency as a predictor of assent to those beliefs. Commitment is an important aspect of belief. I think something similar would be true of political beliefs (e.g., a socialist might give assent to beliefs that are socialist, and oppose, say, liberal beliefs even if on the face of it, they are attractive).

    (2) I was at an intro to theology class yesterday, where the lecturer remarked that he never read in sources prior to the 18th century, “Of course, as a Calvinist I believe…” or “As a Roman Catholic, I endorse…” So not only is there variation on the way we conceptualize beliefs across cultures, but also across times. It’s interesting that even five-year-old children are sensitive to this relativized understanding of religious beliefs in our culture.

  • comment-avatar
    Jonathan Mair 19 October 2012 (18:44)

    Thanks for the post and for bringing this paper to my attention, I’m looking forward to reading the whole paper.

    My initial thoughts, based on your summary here, are that the prospect of delineating different forms of relation to belief as the authors of this paper try to do is very exciting, and that it is an approach that might begin to address the many religious traditions that speak about having special or marked relationships to particular bodies of knowledge that are different to everyday believing or knowing.

    On the other hand, the idea that some beliefs are held by preference seems very culturally specific to me — specific to a particular post-reformation/enlightenment religion==freedom-of-conscience kind of world. In that particular world it makes sense to take for granted that beliefs are transparently intelligible to individual, ordinary believers. As Dan points out in the comment above, frequently, in religious and non-religious contexts, that is not the case.

    I must declare an interest — I’ve just had a paper accepted for ‘Anthropological Theory’ on cultures of belief. My argument is essentially that the categories of belief that are available within particular (especially but not only religious) cultures are often very important in shaping the ways in which people act in relation to bodies of belief. In relation to my key case study, based on research among Mongolian Buddhists in northern China, I argue that for them getting the form or style of belief right is much more important than the content.

    The significance of my argument for this paper would be that before we reach any conclusions about fundamental ways of thinking on the basis of these experiments, the categories they investigate ought to be interrogated to see to what extent they are trans-culturally evident. We might also ask to what extent the researcher’s intuitions about these categories match those of their subjects. On the other hand, I’m certainly open to the idea that there may be a number of fundamental cognitive processes that provide the raw material that cultures of belief deal with — but I expect Dan’s meta-representation/factual representation distinction is more likely to be universal than the distinction proposed here between objective and preferential beliefs, which seems very particularly….Anglican(?) to me.

    In relation to Dan’s second point, in which he encourages anthropologists to investigate these questions experimentally in various cultural settings — I endorse that, but it’s also worthwhile pointing out that categories or styles of belief are something that people often have a lot of reflexive knowledge about, and this is something anthropologists can usefully report on and something that can be usefully applied by psychologists in formulating experiments. Unfortunately, as Talal Asad has pointed out, rather than doing this, anthropologists (such as Geertz and his students) and other scholars of religion (such as Wilfred Cantwell Smith) have often elaborated general schemes in which religion is given a mythic status (related to deep-seated motivations….or preferences?) and opposed to ‘objective’ spheres of politics and economics.

    In religious contexts, the kind of reflexive knowledge I’m speaking about is often tied up with the fact that achieving specific styles of belief in respect of specific bodies of knowledge is a condition for acquiring some non-cognitive good — believing in the right way might be a condition for salvation or might be a virtue whose acquisition is desirable in its own right or is a moral duty. When believing in the right way is incentivised in this way, people speak and think about it and police it a great deal, so there is plenty for an ethnographer to observe. Obviously Christianity is big on this — but Christian views on proper belief are a lot more complex than they have been portrayed as recent work by e.g. Susan Friend Harding and Tanya Luhrmann on Evangelicals in the States shows. In my paper (sorry for the plug!), as well as discussing different categories of styles of belief in relation to my Buddhist case study, and in relation to work with Evangelicals, I also draw on work on ancient Greek pagan religion and on mediaeval Judaism. In each case the distinctions in forms of belief and the value placed on them were/are quite different.

  • comment-avatar
    Dan Sperber 19 October 2012 (19:17)

    Thank you Jonathan Mair for this great comment! More generally, while I have been pushing over the years for a more cognitively informed anthropology, most of the positive response has come from people who were more at ease with cognitive than with ethnographic research, and we obviously need both. When I was taking Helen’s suggestion to do experimental work among people such as the Amish and suggesting myself to extend it to a variety of ethnographic fields, similarly, I meant this as a useful complement to the expertise of anthropologists like yourself, who have, to begin with, payed detailed attention to the people’s own understanding of their beliefs. Experiments are a complement, not a shortcut to ethnography!

  • comment-avatar
    Don Gardner 6 November 2012 (09:42)

    First, it’s very nice to have ICCI back again. I’m sorry it took me so long to discover that the problems had been corrected and to learn that Helen’s interesting post, and Dan and John’s responses had been up for so long.

    A propos of those, in line with Dan’s point about flattening, I found myself wondering about “the metarepresentational attitudes” that underlie scientific or conceptual beliefs too. How, for example, might Heiphetz et al categorise beliefs about the adequacy of the divison between “three kinds of beliefs: those that concern matters of fact (e.g., dinosaurs are extinct), preference (e.g., green is the prettiest color), and ideology (e.g., there is only one God)”: would those beliefs concern matters of fact, preference or ideology?

    Similarly, in regard to the division used in the other paper Helen drew our attention to (Harris et al (2006)), between “real, scientific and endorsed” entities, I worried where pholigiston would fall in this classification, and how the differences between Priestley’s and Lavoisier’s beliefs about “dephlogisticated air” would be handled with respect to the three kinds in Heiphetz et al.

    One might also pose some of the issues in simpler terms by considering our beliefs about the Müller-Lyer Illusion, for although we cannot avoid seeing (and, proverbially, isn’t seeing believing?) the lines as different in length, nobody who’s experienced the phenomenon would, for example, on another presentation of the illusion bet against the lines’ being of the same length when measured with a ruler.

    It sometimes seems that certain psychological experiments involve something very like the idea that our sub-personal cognitive machinery is committed to a Comtean distinction between religious-metaphysical versus positive claims, just as it sometimes seems that social scientists (like Edmund Leach) use a distinction of that sort to constitute the realm of the religious/ritual/symbolic.

    Jonathan’s response refers to Talal Asad’s critique of Geertz on religion, and we might do the same in pleading that the epistemic commitments of social beings be viewed as almost always the joint product of cognitive and ecological factors, notwithstanding the demonstrated capacity of either kind to trump the other in respect of a given judgement, on a given occasion.