The opinion of epistemic superiors as evidence for/against theism?
June 26, 2012 — 6:09

Author: Helen De Cruz  Category: Atheism & Agnosticism Religious Belief  Tags: , ,   Comments: 20

A few months ago, I met a grad student who has a prominent philosopher of religion (X) in his department. The graduate student was a theist when he started grad school, but soon realized that this is a minority position in philosophy. Disparaging remarks about theists (and specifically Christians) fueled his insecurity. Although there are no longitudinal surveys on this, it seems that atheism increases as people climb the academic ladder. For example according to the PhilPapers survey, in philosophy, only 14.6% of philosophy faculty believe in God. When postdocs are included, this percentage rises to 16.3%. Graduate students have the highest percentage of theists, 20.8%. So it seems plausible to me that at least some graduate students lose their faith as a result of the majority opinion in academia.
As the graduate student and I discussed X and what a wonderful scholar she is, the graduate student said that one important reason he is still a theist is the fact that X is a theist. He said “X is one of the smartest academics I know. The fact that X is a theist, even though she considered counterevidence carefully (like the problem of evil), is for me strong evidence for theism, and for me it’s a good enough reason to remain a theist”. Is the graduate student rational?

Now, treating X’s opinion that p as evidence for p is not as problematic as it may seem at first. Given the importance of the evidential force other people’s opinion in recent social epistemology, it does not seem all that implausible. Although we should not straightforwardly defer to the opinion of others, even epistemic superiors, it seems rational to accord at least some evidential weight to their views.
However, there are many smart academics who are not theists. Purely percentage-wise, our graduate student, were he to adopt the principle “Accord some evidential weight to the beliefs of epistemic superiors” would find his opinion swamped by the majority of non-theistic academics. But the graduate student does not statistically aggregate opinions of others. Rather, he takes *X’s beliefs* in particular as evidence for theism, at least as a good enough reason to remain a theist.
My question is: is the graduate student being rational, and if so, on what basis? I am thinking of the following suggestions:

  1. He knows X’s work and views really well, so he can better assess the reasons why she is a theist and does not have this access to other academics’ religious opinion. However, if this is his heuristic, it is not quite clear why the *opinion* of X, rather than straightforwardly going for the reasons for why she holds this opinion, should count as evidence? Can’t he evaluate those reasons independently of her opinion?
  2. X is already a theist, and, at least under some epistemological views (e.g., Kelly’s belief polarization paper) it seems not unreasonable to pay more attention to evidence that supports one’s views. X’s opinion might carry much weight in the graduate student’s total body of evidence that is significant for theism.
  3. X has turned out to be a reliable guide to justified true beliefs in other domains for the graduate student, so in the absence of other information, it is not unreasonable to also trust that her theistic belief is justified.

It seems that in everyday life, we do treat the opinion of some people as having more weight than the aggregated opinion of the community. Such people are typically epistemic superiors, i.e., people whom we think of as smarter or better-informed than we are. I am wondering if doing so is justified, or just a form of biased reasoning (confirmation bias, misleading vividness etc).

  • In social psychology experiments (the Asch conformity studies) it has been shown that people will conform to the false opinion of others even when it is obvious the opinion is wrong. People would rather agree than be right. However, if there is even one confederate in the group who points out the obvious falseness of the group opinion, conformity drops dramatically. As such, perhaps this graduate student was just being psychologically honest and a little intuitive too.

    June 26, 2012 — 6:37
  • Helen De Cruz

    Thanks for mentioning the Asch studies. The fact that one minority voice can make such a difference provides a good psychological explanation of why we pay attention to dissenting voices, and why it’s sometimes rational to not just aggregate opinion. Of course, the majority atheist view in philosophy and academia more generally is not the result of an experimental agreement to hold this view.

    June 26, 2012 — 6:49
  • In a sense doesn’t this kind of argument have some strong links to religious experience arguments (sharing a genus, perhaps)? For instance, while it’s not the primary reason, if I listed all the reasons I’m a theist, one of the reasons would be Teresa of Avila, since I think (1) she’s awesome in general and (2) all purely nontheistic attempts to account for her experiences and her interpretation of them that I’ve seen or can think of end up requiring us to believe her more unreasonable or irrational than I think the evidence actually shows her to be (there are a few other supplementary points). This isn’t really that much different from the graduate student’s case; it’s just that the experience that is doing the work in that case is not a purported direct experience of the divine but an experience of assessing evidence and argument of the divine. But that’s an experience as much as anything else, and it makes just as much sense to consider the veridicality of experiences of rational assessment as any other experiences. And there’s a pretty straightforward sense in which even in arguing our own case we are really arguing for the veridicality of our experience of ourselves as rational evaluators (which is why in many kinds of arguments people have difficulty not taking contemptuous dismissal personally). But there’s no obvious reason why we would be able to do this for ourselves but not also for other people, just as there’s no obvious reason why we should only appeal to our own sensible experiences and not those of others as well. So it’s an argument from rational experience, closely analogous to arguments from religious experience.
    (I should I say that I think the justified vs. biased reasoning is actually a false dichotomy; I think unbiased reasoning is impossible — i.e., justified reasoning is not unbiased reasoning but reasoning in which, whatever the biases may be to which it is subject, certain standards are met. I suspect that this would affect my assessment of these kinds of arguments.)

    June 26, 2012 — 7:27
  • Helen De Cruz

    Hi Brandon: thanks for these comments. I agree that the justified vs biased reasoning is not a helpful dichotomy. All reasoning is biased to some extent, and the Cartesian project of trying to start without any past beliefs affecting our judgment is doomed.
    But we can strive to reduce bias in our reasoning, and perhaps paying proportionately more attention to people we hold as epistemic superiors is a form of bias that we should reduce – I’m not sure.
    It is interesting to consider whether it makes a difference if our epistemic superior is an epistemic superior because she is academically better-versed (this was the case for the graduate student) or because she has privileged access to information, as in the case of mystics like Teresa of Avila).

    June 26, 2012 — 7:36
  • xyg21

    I had episodes like this as an undergraduate and continue to do so as a graduate student. The thought (speaking for myself) isn’t so much that the fact that prof X believes p makes p particularly likely. It’s that since prof X is a much better philosopher than me, her reasons for thinking p are probably better than any reasons I could come up with on my own. Of course, disappointment can be easy to find (“Prof X, why do you believe p?”), but more often it dawns on me that prof X has a point. I don’t know what I’d do if prof Y (who thinks ~p) were to join the faculty.

    June 26, 2012 — 8:39
  • General comment:
    Well, first of all, X’s theism might reasonably count for a lot more to the graduate student than the theistic and atheistic views of philosophers X only knows by academic reputation. After all, it is people one interacts with that one can better judge the moral and intellectual virtue of. And being a really good philosopher–in the truth-apt, not just the professional, sense of “good philosopher”–tends to require both moral and intellectual virtue. Why moral virtue? Well, for one, being morally better is central to making one wise, and wisdom is needed for getting deep and correct insights in moral philosophy, and philosophy has such an interconnectedness that one needs to bring to bear insights from moral philosophy to other areas of philosophy–it is, indeed, a criterion of adequacy on a metaphysical view of the world that it support or at least be compatible with a satisfactory view of what is important, what is good and what is right.
    So if the graduate student reasonably seems more virtue in X than in other philosophers that the graduate student is close to, and X is less well-informed or less intelligent, then there is a straightforward story as to why the graduate student should trust X’s opinions more: X has more of the prerequisites for being truth-apt philosophical investigation.
    But suppose not: suppose X as reasonably seen by the graduate student is no more virtuous than X’s colleagues. Then X can at least function as a counterexample to universal generalization like: “One cannot be reasonable, fairly virtuous, highly intelligent and well-informed contemporary philosopher and a theist.” And hence X’s existence defeats at least one argument against theism.
    Without X, the graduate student might think like this: “Yeah, I’ve heard that there are reasonable, virtuous, highly intelligent, well-informed contemporary philosophers, like Al Plantinga, who are theists. But I haven’t actually had any significant interaction with them, and I don’t know if people’s opinion of their reasonableness, virtue and intelligence is justified. All the reasonable, virtuously, highly inteligent, well-informed contemporary philosophers I know well–i.e., my professors–are nontheists, and that makes me sceptical about the people I’ve heard of. Moreover, even if there are a few such, I bet their philosophical views are really strained. They may be clever at answering objections, but their views probably don’t hang together very naturally.” The presence of X not only lets students judge for themselves about reasonableness, virtue and intelligence, but also lets them tell better whether the person’s view is strained–whether the person has to struggle to make things fit with theism.
    Now this doesn’t make X’s characteristics be the graduate student’s reason for being a theist. But it makes the characteristics defeat defeaters against theism.
    Specific comment:
    “it is not quite clear why the *opinion* of X, rather than straightforwardly going for the reasons for why she holds this opinion, should count as evidence”
    This assumes that the reasons are all graduate-student evaluable. But in fact a good professional philosopher also has reasons that aren’t graduate-student evaluable. A good professional philosopher can sometimes just see that a theory “isn’t going to work out” (and when one asks “Why?” the philosopher just says “Because views like that never do”), that a principle will have a counterexample, that an objection is “merely technical”, or that the alternatives to a position will be strained and epicyclic. Such intuitions are crucial to the work of a good mature philosopher, and they are the result of skills that are refined over many years after graduate school. And not everyone develops them.
    Of course there is a question whether the graduate student can tell that the philosopher has such skills. But at least some such skills can be evaluated. “She’s told me three times that that wasn’t going to work out, and gave me a vague hint as to why, and each time, by golly, she was right.”

    June 26, 2012 — 9:12
  • What percentage of post-docs are theists?
    That’ll be skewed, perhaps, because post-docs aren’t that common in philosophy, and there are several specifically theistic post-doc programs, though.
    It would be interesting to see what percentage of young faculty are theists.

    June 26, 2012 — 9:48
  • Helen De Cruz

    Hi Alex, looking back at the Philpaper survey, I noticed that I misinterpreted some of the figures (it says non-target faculty for the percentage I thought to include postdocs).
    There is some information on academia in general in a study by Ecklund and Scheitle (here:, paywalled alas).
    They interviewed faculty from prestigious US universities in physics, biology, the social sciences. They write “Similarly, although … older individuals express higher levels of religious belief and practice when compared to younger individuals, we do not find this relationship among scientists. … the middle age cohorts are more likely than the youngest scientists (18 to 35) to say that they do not believe in God. A similar pattern is seen with the attendance measure, with those 56 to 65 more likely than the youngest group to report not attending religious services over the past year.” (pp. 301-302).
    I find this really intriguing, especially since it goes against the grain of the general trend of increased religiosity with age (This general trend is not caused by older people suddenly becoming more religious, but just a cohort effect, apparently: younger generations are just less certain about God, as Trent would put it, and attend religious services less often). So given that this is a different trend, what could explain it? The explanations might be that (1) academia is now friendlier to religion than it used to be, (2) religious individuals are for various reasons less likely to stick around in academia, and so get (self)selected out (3) religious individuals become less religious as they get more versed in science – a sort of instance of the effect Gervais & Norenzayan were describing or (4) the fact that one is in an overwhelmingly atheist work environment leads, through conformist bias, one to abandon religious beliefs.

    June 26, 2012 — 10:50
  • hiero5ant

    Definitely, as a general proposition one ought to take one’s epistemic superiors’ beliefs as evidence for their truth. In fact, I might be tempted to go so far as to say that this is definitionally true.
    However I would dispute the assumption that academic ascent in philosophy comprises specifically epistemic superiority, defined in terms of awareness of a greater body of extra-textual facts, or superior calculating power which makes a PhD a more generally reliable truth-finder than an MA. I am not aware, for instance, of any degree program which grants one “epistemically superior” access to the reality of suffering not possessed by the illiterate migrant farmworker who has just had to bury his son after he was killed by a drunk driver.
    I suspect the turgid, unreadable Twilight novels are less popular among those with MFAs in creative writing than the population at large, but it seems a category mistake to refer to this as the result of “epistemic” differences. Rather, as in philosophy, I would want to attribute the superiority of elites to a combination of innate character dispositions leading to initial self-selection, combined with acquired taste.
    Or perhaps it is more like the decrease in pack-a-day smokers among elite athletes. They have no generalized “epistemic superiority” about the causes of lung cancer and heart disease compared to the average man on the street. They have simply found certain habits inimical to success at their chosen project.

    June 26, 2012 — 11:40
  • JJ

    Perhaps it would be helpful think about the question in terms of how knowledge of the form “The majority of members of truth-apt group S hold belief A” generally functions as epistemic support for A. It seems to me that a supplemental premise (in addition to the general truth-aptness of individual members of S) like this is usually involved: If more of the individual members of S hold A than hold not-A, then although the ascription of general truth-aptness applies to the latter individuals as well, it is prima facie more reasonable to presume that truth-aptness has been “interrupted” by other factors in the minority of individuals than in the majority.
    So, knowledge about majority opinion seems to function (in the absence of detailed knowledge and adequate understanding of all the reasons/arguments of every member of S) by way of a supposition that, other things being equal, it makes more sense to think the belief held by the majority of truth-apt individuals is correct than to think general truth-aptness has been interrupted for the majority of individuals in a particular case.
    It seems evident, however, that any number of factors may constitute reason to judge that all things are not equal in a particular case, and thus that majority opinion among generally truth-apt individuals does not, in that case, function reliably as evidence for A. If the grad student in question derives, from his close interaction with X and particularly from his observation of how she does philosophy, reason to suspect that it is highly unlikely that truth-aptness has been interrupted for her in the case of theism, then it may no longer make more sense than not for the student to presume that truth-aptness has probably been interrupted in the case of theism for each not-A-holding member (including X) of S.
    If so, then the student’s opinion of X seems at least to justify his rejection of the use of “The majority of good philosophers hold atheism” as evidence for atheism. But more to the point, because it seems truth-aptness must have been interrupted in the case of either A-holders or not-A-holders (or both), the student’s judgment that it is highly unlikely that truth-aptness has been interrupted in the case of A for not-A-holder X may indeed constitute (in the absence of defeaters) positive reason for his believing that it has probably been interrupted in the case of the A-holding majority and expressed in the not-A-holding minority of which X is part, and thus positive reason for his theistic belief.

    June 26, 2012 — 14:19
  • BTW, are there any data about past philosophers’ and graduate students’ religious beliefs? Could it be that nowadays graduate students are more likely to believe in God because of, you know, the Plantingas, the Swinburnes, the Craigs and the Prosblogions? I tend to think that that is actually the case.

    June 26, 2012 — 14:48
  • Let me throw another wrench into this.
    In the case of completely rational epistemic agents, one needs to take into account not what view is the majority view but also the degrees of belief as well as the degrees of dependence between their evidence and differences, if any, in priors. It can get pretty complicated.
    Let’s say that A, B and C are all privy to the same body of evidence. Each assigns a prior of 0.5 to some theory T. But each evaluates the evidence differently, so that A and B end up assigning 0.2 to T while C assigns 0.99. Under those circumstances, I think the correct aggregate probability for T is given by letting the log-odds be the average of the log-odds, which gives 0.65.
    On the other hand, let’s say that A, B and C are all privy to different and conditionally (on T and on ~T) independent bodies of evidence. Again, suppose each assigns a prior of 0.5 to the theory T, and they assign probability 0.2, 0.2 and 0.99. In that case, if you want to aggregate their evidence, you simply need to generate an aggregate Bayes’ factor which is obtained by simply multiplying the Bayes’ factors that got them from 0.5 to 0.2, 0.2 and 0.99. This basically means your log-odds are the sum of the log-odds of A, B and C, which gives 0.86.
    On either take, then, the near-certainty of p that C has ends up outvoting the moderate certainty of ~p that A and B both have. Presumably in real life, A, B and C have some evidence that’s the same, some that’s independent and some that’s mutually dependent, and if the priors were 0.5 initially, then the aggregation will presumably be between 0.65 and 0.86 (this isn’t a theorem, just a gut feeling).
    I know that opinion polls sometimes ask how strongly people believe something, but I expect it would be hard to convert this to probabilities (or log odds).

    June 26, 2012 — 17:12
  • Matthew Mullins

    there are many smart academics who are not theists. Purely percentage-wise, our graduate student, were he to adopt the principle “Accord some evidential weight to the beliefs of epistemic superiors” would find his opinion swamped by the majority of non-theistic academics
    We should be careful about how we use epistemic superiors here. The class of individuals with more cognitive horse power than yourself, isn’t the class of epistemic superiors. Someone who is an epistemic superior will have to be someone who, among other things, has spent a lot of time thinking about the matter at hand. On the issue of theism, this isn’t true of most academics or even most philosophers. So that most academics, or even most philosophers, aren’t theists shouldn’t worry the graduate since most of these individuals aren’t in fact epistemic superiors.

    June 26, 2012 — 21:49
  • Helen De Cruz

    Hi Eduardo: There is only anecdotal data, suggesting that it’s now more intellectually respectable to be a theist. I think prior to the 1960s, especially given the intellectual climate, theist grad students in philosophy were probably more circumspect about their beliefs.

    June 27, 2012 — 1:26
  • Helen De Cruz

    Matthew: interesting point. This relates to an earlier discussion we had here about the fact that the supermajority (70 to 75 %) of PORs are theists. If they started out agnostic, and became theists as a result of their investigation, I think their belief would be strong evidence for theism. However, it seems more plausible to say that in most cases, PoRs start out as theists and *maintain* their position throughout their investigations.
    What do you think can we make of that? Is it purely confirmation bias or entrenchment?

    June 27, 2012 — 1:31
  • Kenny Pearce

    The following principle is surely correct:
    If it is the universal opinion of those who are genuine experts as to whether p, that p, then it is unreasonable for a non-expert who is aware of the universal agreement of the experts to believe that ~p.
    This principle is, I say, surely correct, but it is vague (because it’s not clear who the genuine experts are), and its antecedent may never be satisfied in any interesting case. But, if the graduate student you talked to regarded himself as a non-expert, then he may have regarded X as a counterexample to the antecedent of some principle like this, and that seems right.
    A more informative principle, that applied to more cases, would refer to expert consensus, but figuring out what a consensus is just as difficult as figuring out who counts as an expert. It seems to me, however, that in trying to determine whether there is a consensus, we should weight opinions by the level of expertise of the person who holds them, rather than just classifying the world into experts and non-experts and treating all experts alike. Now, no matter how great an individual’s expertise is, one individual’s dissent does not necessarily break a consensus – for instance, there was a consensus of experts in favor of quantum mechanics even when Einstein was still disputing it, but Einstein had as much (or more) physics expertise as anyone then living. Still, if the grad student judges that X has an extremely high level of expertise this might contribute strongly to a rational judgment that there is not a consensus among experts on the question of God.
    One other thing that might be of interest here. Leibniz argues in the New Essays and Theodicy for a presumption in favor of the experts within one’s own community. He’s worried about the case of a European peasant who knows what his own priest teaches, and is also aware that, out in the wide world, there are other communities who regard other people as experts who teach things contrary to what his priest teaches. We are to suppose the peasant doesn’t know the content of these other teachings. Leibniz says it’s rational for the peasant to presume (in the legal sense, as in ‘presumption of innocence’) that his own priest is correct unless and until he has the opportunity to inquire further. (A peasant is not likely ever to have such opportunity, but similar considerations apply to all of us, of course.) The reason for this, I think, is that he has better evidence for the claim that his own priest is an expert than for the claim that these other alleged experts exist, are genuine experts, and really give contrary teachings. This sort of idea seems to me to be similar to some other suggestions that have been made on this thread.

    June 27, 2012 — 12:14
  • John H.

    I haven’t been able to read all the comments, so forgive me if this has been suggested already, but even if one does not think that a knowledgeable philosopher’s opinion that God exists is evidence for God’s existence, that philosopher’s opinion could still have epistemic significance for the graduate student. Perhaps the grad student believes that most theists are theists for non-evidential reasons, and that non-evidential reasons can be epistemically justifying; what the grad student wants to know is if there is strong counter-evidence to theism. The fact that most philosophers are atheists would have no bearing on this: those philosophers may not be atheists on the basis of counter-evidence to theism, but simply because they don’t have non-evidental reasons to justify theism. Thus the grad student does not need to hold his model philosopher’s opinion in higher regard than the atheist ones.
    Of course, some atheist philosophers do think there is strong counter-evidence to God’s existence, so the question of the grad student’s rationality reemerges. Nonetheless, if the grad student believes that only a minority of atheist philosophers hold that there is strong counter-evidence to God’s existence, that makes thinking that his model philosopher is in some superior epistemic position – perhaps for reasons that others here have provided – more plausible.

    June 29, 2012 — 13:42
  • John H.

    I should probably amend my previous comment in one way – the grad student does hold the model philosopher’s opinion in higher regard in at least one way, namely thinking that it is true while the atheists’ opinions are false.
    What I meant is that the grad student does not need to think that the model philosopher is more rational, better at evaluating evidence, or something of that ilk, than the atheist philosophers.

    June 29, 2012 — 14:17
  • Carneades of Ga.- Skeptic Griggsy

    The student should just ponder from wide reading of atheist and theist materials and weight the pros and cons and then make up his mind.

    June 29, 2012 — 16:57
  • What percentage of senior faculty are theists?

    July 27, 2012 — 19:33
  • Leave a Reply to Syphax Cancel reply

    Your email address will not be published. Required fields are marked *