God, evolutionary psychology and moral realism
August 5, 2011 — 15:49

Author: Alexander Pruss  Category: Uncategorized  Tags: , , ,   Comments: 36

Define moral realism as the claim there are objective moral truths and that we know some of them. Consider the argument:

  1. If no religious beliefs are true, moral realism is false.
  2. Moral realism is true.
  3. So, some religious beliefs are true.

I won’t argue for (2), but only for (1). For my argument I will assume a form of reliabilism. (I think the arguments here work well to establish (1) on non-reliabilist epistemologies. The present argument plugs the weakness of that argument.)

Here is the line of thought. Start with this plausible observation:

  1. If no religious beliefs are true, the correct explanation of our moral beliefs is that moral beliefs were beliefs about unobservable realities that evolved to help prevent defection in prisoner’s dilemmas in cognitively sophisticated hominids.

Now, if reliabilism is true, the question is whether the process, P, of evolutionarily forming beliefs about unobservable realities to help prevent defection in prisoner’s dilemmas is reliable: is likely to produce true belief. But now observe that if no religious beliefs are true, very likely our religious beliefs also arose out of P . Positing supernatural judges who can see if one is sneakily defecting in prisoner’s dilemmas is obviously quite helpful. Thus, we have two families of beliefs produced by P: moral and religious. If the religious ones are all false, the process is unreliable. If the process is unreliable, then its outputs are not knowledge. And so if no religious beliefs are true, we have no moral knowledge, and hence moral realism is false.

Of course, we have the usual tricky thing with reliabilism: What is the relevant level of description of the belief-forming process? Is it: “evolutionarily forming beliefs about unobservable realities to help prevent defection in prisoner’s dilemmas”, or is it something narrower that is special to the moral case, and not present in the religious case? I think it would be difficult, however, to formulate a description narrowed to the moral case without being completely ad hoc.

Comments:
  • I’m generally sceptical about all such purported explanations. But supposing something like that is right, one way someone might try to narrow the description to the moral cases is by making a distinction between the evolution of beliefs concerning which norms are correct, and the evolution of beliefs concerning why it is that those norms are correct. Perhaps the relevant process is reliable with respect to the former but not the latter.
    The thought would be that people might have reliably evolved good folk theories about what the right thing to do is (don’t defect), and bad folk theories about why that is the right thing to do (Moorean realism, divine command theory). I assume such false meta-ethical beliefs don’t threaten first order moral justification (if they do, most of us are in bad shape). So realism would remain on the table. I have no stake in this, but such a story doesn’t seem particularly ad hoc (that is, it seems no more ad hoc than any other speculation about the evolutionary origins of morality).

    August 5, 2011 — 19:37
  • Brad,
    It’s a good distinction, but I don’t think it helps, because something like this distinction can be made between first-order religious beliefs (how many gods there are, do they know what we are up to, etc.) and higher-order theological analysis (what omnipotence exactly consists in). If the evolutionary process gets wrong the religious beliefs on both levels, then even when restricted to first-order claims the process is not reliable.

    August 5, 2011 — 20:52
  • Hi Alexander,
    I intended the distinction to be not between first and second order beliefs as such, but between normative beliefs and meta-normative beliefs. And the difference was supposed to be that all religious beliefs, but not all moral beliefs, fall on the meta-normative side of things.

    August 5, 2011 — 21:02
  • I don’t see why religious beliefs are analogous to meta-normative beliefs. Consider the belief that Zeus produces lightning or that God created all life on earth. These don’t seem to be relevantly analogous to meta-normative beliefs. Rather, they seem to be relevantly analogous to first order beliefs, such as that Zeus should be worshiped or that life should be respected.

    August 5, 2011 — 21:06
  • (I should add that in my twin examples of religious and moral beliefs, I intend the first item in each pair–the one about Zeus–to be clearly false, and I believe, but am not assuming for the sake of the argument, that the second–the one about life–is true.)

    August 5, 2011 — 21:07
  • Right, I agree for those examples. But those sorts of supernatural causal beliefs are not among those explained by your process P, are they? (It’s unclear how they conduce to behaving appropriately in prisoner’s dilemmas).

    August 5, 2011 — 21:22
  • Well, believing that Zeus causes lightning certainly conduces to cooperation in prisoner’s dilemmas, at least if one also believes that Zeus cares about cooperation in prisoner’s dilemmas. 🙂

    August 6, 2011 — 9:15
  • Ha, right. So perhaps the distinction I am looking for is simply normative versus non-normative. The initial thought I was trying to articulate was simply that the process might reliably produce accurate beliefs about what should be done, but also non-reliably produce various other beliefs with which they are entwined.

    August 6, 2011 — 11:18
  • But then we’re pretty much limiting the reference class for determination of reliability to the cases in question. And that seems objectionably ad hoc.

    August 6, 2011 — 14:02
  • I don’t think so; but I’m about to be off the grid for a day or so. More soon!

    August 6, 2011 — 14:20
  • I’m not sure about (1). I assume the beliefs you think would be true are something like:
    a. Moral realism is true.
    or maybe
    b. It is really wrong to kill a child.
    If that is correct, I question whether those beliefs are accurately represented for two related reasons – John Searle (and earlier philosophers) distinguishing between sentence meaning and speaker meaning and Kent Bach’s work on impliciture.
    With those in mind, I would say the beliefs could be better represented as something like:
    a’. Moral realism is true [and is rooted in a divine author].
    b’. It is really wrong to kill a child [because God has commanded it].
    If that is the case, then (1) is not clear to me, assuming those are the sorts of beliefs you had in mind.

    August 7, 2011 — 21:19
  • Mike,
    Whose moral beliefs are you claiming have this implicit form? Those of non-theist moral realists? Those of ordinary people? Mine?
    I can only speak for myself here, and while I believe God has commanded us not to commit murder, that divine command is not in fact the primary explanation for why murder is wrong.
    You might be right about ordinary people’s moral beliefs. If so, that strengthens my argument for 1.

    August 7, 2011 — 22:21
  • Alex,
    If my assumption wasn’t correct, then I’m not sure what types of beliefs you mean in (1) and (3) of the argument. Maybe you could offer an example to clarify for me.

    August 7, 2011 — 22:36
  • “I’m not sure what types of beliefs you mean in (1) and (3) of the argument.”
    How about your example (b), but without any theological amplification?

    August 8, 2011 — 0:01
  • If that is the case, and (b) requires no semantic completion of a religious nature, then in what sense is the belief a religious one?

    August 8, 2011 — 7:57
  • I wasn’t saying it’s a religious belief.
    I was affirming the material conditional: if no religious belief is true, moral realism is false.

    August 8, 2011 — 8:35
  • I’m confused. I asked for an example of the beliefs you referred to in (1) and (3) and you specifically called those beliefs religious in the premises. So if you name an example of such a belief and do not claim it is religious, then how does that confirm (1)? Am I just missing something obvious? What is a specifically religious belief that must be true for moral realism to be true?

    August 8, 2011 — 9:05
  • First, “must be true” is an overstatement: I am only giving a material conditional.
    Second, there isn’t any specific religious belief that must be true for moral realism to be true. The argument doesn’t work like that. Let me sketch the full logical form:
    A. (Premise) Evolutionary process P produced both our religious and our moral beliefs.
    B. (Premise) If all religious beliefs are false, P is unreliable.
    C. So, if all religious beliefs are false, our moral beliefs are produced by an unreliable process. (By A and B)
    D. (Premise) Beliefs produced by an unreliable process are not knowledge. (This is a consequence of reliabilism.)
    E. So, if all religious beliefs are false, we lack moral knowledge. (By C and D plus the analytic truth that knowledge requires belief.)
    F. (Premise) If moral realism is true, we have moral knowledge.
    G. So, if all religious beliefs are false, we have no moral knowledge. (By E and F)

    August 8, 2011 — 9:17
  • Thank you. That helps. I see your point now.

    August 8, 2011 — 9:28
  • Brad:
    There may also be lots of normative beliefs about how to deal with the gods, such as that this god should have goats sacrificed to it.

    August 9, 2011 — 10:04
  • Gene Witmer

    Alex,
    Let me jump right to the formal argument you gave in reply to Mike Gage:

    A. Evolutionary process P produced both our religious and our moral beliefs.
    B. If all religious beliefs are false, P is unreliable.
    C. So, if all religious beliefs are false, our moral beliefs are produced by an unreliable process. (By A and B)
    D. Beliefs produced by an unreliable process are not knowledge.
    E. So, if all religious beliefs are false, we lack moral knowledge. (By C and D.)
    F. If moral realism is true, we have moral knowledge.
    G. So, if all religious beliefs are false, we have no moral knowledge. (By E and F)

    I confess to being quite puzzled as to why anyone would be confident in premise A. There is also an interesting problem with premise B that I will remark on, but my main worry is A.
    Regarding B: You note in the original post that there is the usual problem with reliabilism about the relevant type to which the process belongs and worry a bit about excessively narrow types. But there is a worry from another direction: if process P is understood sufficiently broadly, then premise B may be false. For instance, if P is a process that produces both moral and religious beliefs, yet it produces many, many more moral beliefs than religious ones, then it may still count as reliable even if all the religious ones are false. The point is even more obvious if P produces beliefs of other sorts — neither moral nor religious. Interestingly enough, if that is the case, the result would be that religious beliefs are justified (assuming reliabilism here) even while false, yet moral beliefs are justified while being often true.
    As I say, though, my main worry is about A. Why think the same process produces both? How could one make this claim plausible?
    One could try to do so by giving a specific evolutionary explanation and defending the claim that that is the actual explanation of both kinds of belief. This seems to be what you were doing in the original post, where you took the following claim (I’ll call it ‘E’) to be plausible:

    (E) If no religious beliefs are true, the correct explanation of our moral beliefs is that moral beliefs were beliefs about unobservable realities that evolved to help prevent defection in prisoner’s dilemmas in cognitively sophisticated hominids.

    I suspect that someone who is already convinced that the falsity of all religious beliefs requires the falsity of all moral beliefs would find this somewhat plausible on the following grounds. The explanation on offer in E is one that happily detaches the relevant facts from the belief; what explains the belief is not the fact — because the belief is in fact false — but something contingently related, namely, its relation to motivation. But if you don’t already buy the idea that if all religius beliefs are false, then so are all moral beliefs, is there any reason to think this explanation is the actual explanation?
    Consider a simple realist position according to which moral facts are facts about welfare, its distribution, and the impact of actions on such, where welfare is constituted entirely by ‘naturalistic’ phenomena such as abilities, satisfaction of desires, and so on. In this case, it seems nothing so special as the explanation in (E) needs to be invoked; one knows about welfare, its distribution, and the effects of various actions on such by ordinary perceptual means.
    I should note that I don’t think we would be reasonable even in adopting this sort of explanation for religious beliefs even given their uniform falsity. After all, there are other possible explanations — a general tendency to anthropomorphize, for instance, and the low costs of positive errors. In addition, it is worth noting that the kind of explanation you mention — whereby belief in supernatural judges who can detect defectors — can still play an important role in explaining some religious beliefs. So, for instance, it may be that general anthropomorphizing explains why people tend to believe in unseen agents, while prisoner’s dilemma situations explain why they tend to believe that those same unseen agents are apt to threaten and cajole them in certain ways.
    The same point holds for moral beliefs: it may be very plausible that such situations help explain why we have certain beliefs related to morality. For instance, the belief that if we do wrong we will pay for it. Obviously (I say!) this belief is false, but it is very pervasive and some explanation is needed. The fact that the belief is generally advantageous to the species could help explain this. But even if that process is unreliable, there is no reason to subsume all moral beliefs under this category — that is, the category of being produced by that sort of process.

    August 9, 2011 — 13:55
  • Gene,
    Thanks! These are very helpful criticisms.
    I find deeply implausible the idea that moral facts reduce to facts about the distribution of welfare. Take four actions, A, B, C and D, that you’re choosing between then take the facts about how A, B, C and D affect everybody’s welfare. Suppose, for instance, that A has the best result in terms of the sum total of individual welfare, that B has the best result in terms of the mean value of individual welfare, that C has the best result in terms of the sum of the cube roots of individual welfare, and that D has the best result in terms of the welfare of the least well-off. It seems clear that given all this we should ask for a further fact, namely which of the four is to be done. For instance, should we maximize total utility, mean utility or total cube-root-of-utility? should we maximin? or something else? (I suppose this is just a particularized version of the Open Question Argument, though it seems–perhaps fallaciously–more compelling in this case.) (I actually think that maximizing total cube-root-of-utility may match some of our intuitions about distribution of welfare better than maximizing total utility.)
    I think your point that a different explanation of religious belief can be given from the explanation of moral belief is pretty strong. It does seem to me that the social cooperation story is prima facie just as plausible on the religious side as on the moral side, and maybe more so (in that on the religious side, the evolutionary forces wouldn’t need to create completely new “moral desires”, but could simply make use of self-interested desire for personal well-being). However, you’re right that anthropomorphization does provide an alternate explanation on the religious side that isn’t available on the moral side.

    August 9, 2011 — 14:56
  • Gene:
    I should add that I don’t buy the idea that if all religious beliefs are false, then all moral beliefs are false. I only buy the idea that if all religious beliefs are false, then there is no moral knowledge. So I don’t think my reasons for accepting (E) match the ones you suggest. I just think (E) is the most plausible naturalistic story on offer.
    It is plausible (a plausibility that coheres well with what I’ve heard about the empirical and non-empirical work on cooperation and prisoner’s dilemma) that (a) cooperation can conveys selective advantages to human communities and (b) the selective advantages disappear and may turn into disadvantages once the number of clever undetected defectors rises.
    Thus there is a special fitness problem for communities that contain members sufficiently intelligent to cheat in clever ways. It is very plausible that the development of appropriate moral beliefs or of appropriate religious beliefs in fact provided such a solution: that both can enable cooperation and lower the percentage of non-cooperators to a level small enough not to disturb the basic cooperative structure of society. It does not follow logically that the moral and religious beliefs became prevalent because they solved the problem. But this still seems a plausible inference, not very different from those made in other areas of biology, where after demonstrating that some development would confer a selective advantage one infers that it probably became prevalent because of that selective advantage.
    Things, however, become more complicated when there is an alternate explanation, such as in terms of anthropomorphization.

    August 9, 2011 — 15:11
  • Much as I hate to intrude on your bizarre obsession with evolution again, I have to ask what makes you think that a cognitive process that “prevents defection in prisoner’s dilemmas in cognitively sophisticated hominids” does not yield objective moral facts, and how you know that the facts at which it arrives are not objective moral facts.

    August 9, 2011 — 21:31
  • At no point in the argument is it denied that the process yields objectively true moral beliefs.

    August 9, 2011 — 21:43
  • Conrad

    1. If no religious beliefs are true, moral realism is false.
    2. Moral realism is true.
    3. So, some religious beliefs are true.
    So I have to confess I really have no idea what (1) is supposed to amount to. To my mind it’s either tautologous or in need of deep elaboration. What do you mean by religous beliefs?
    I am an extremely devout Buddhist of the Theravada tradition and my tradition has never regarded belief in any type of supernatural beings to be essential.
    If you ask me to separate my moral beliefs from my religious beliefs, I’m frankly not going to be able to do it except for the metaphysical belief in a lack of a soul and Lord Buddha’s teaching that everything arises due to antecedent causal conditions.
    How do we distinguish religious beliefs from non-religious beliefs? It cannot simply be in terms of metaphysical content about certain supernatural person for if that is the case then neither I nor many of my fellow Buddhists can qualify as religious which is something I cannot accept as an accurate.
    To my mind what separates religious beliefs from non-religious beliefs is whether or not the beliefs in question have preeminent practical concern in one’s everyday life. It is not just that we regard these beliefs as having preeminent concern, but that we actual think about and act upon them in our everyday affairs. To my mind, not all Christians and Muslims (nor followers of the Blessed One’s Dharma for that matter) are religious believers.
    So maybe you’re not really concerned with religious beliefs. Maybe you really think the following:
    1. If no beliefs about supernatural persons are true, moral realism is false.
    2. Moral realism is true.
    3. So, some beliefs about supernatural persons are true.
    I must confess this is a weird argument to me. I have no reason to think that that (1) is true.
    Now it’s my understanding that contemporary ognitive psychologists speak of the human mind as possessing various “tools” to organize our experience, one of the most important is
    recognizing something as an agent i.e. a being capable of initiating action on the basis of mental states and being able to discern the purposes of that agent. Agency detection devices seem crucial for survival.
    Now to experience another human or nonhuman animal as an agent is nota matter of conscious inferences, but of a preconscious
    mental process that enables us to make quickly a range of judgments relevant to action.
    Our agency detection devices and our ensuing judgments are not infallible.
    So now I ask the question, why should I believe the following claim: (1) If no beliefs about supernatural persons are true, moral realism is false.

    August 9, 2011 — 23:04
  • I agree that it is difficult to demarcate religious beliefs. I don’t think it requires a belief in an agent. But it does, I think, require incompatibility with naturalism. Atheistic Buddhism is a difficult case to classify, though at least I think it satisfies the incompatibility criterion.
    “why should I believe the following claim: (1) If no beliefs about supernatural persons are true, moral realism is false.”
    Because of the argument I gave. 🙂 Whose best summary in this thread was given in my A-G argument to Mike. But I admit that Gene has hit on a serious weakness in the argument, namely that one might disagree with my judgment that the best evolutionary story about the rise of religion involves the encouragement of cooperation.
    By the way, I think that mere anthropomorphization, which both you and Gene talk about, is not enough for religious belief. The belief that, say, a river is sapient is not a religious belief. It is simply a mistaken belief about the distribution of intelligence on earth.
    I think a necessary condition on a religious belief (another? or does this entail the first?) is that it involve something of the numinous, in the Rudolf Otto sense. Something like a holy that goes beyond the morally good, an uncanny that goes beyond the strange, an awe-ful that goes beyond the scary and great, a fascinating that goes beyond the attractive. In fact, that the belief be tightly tied to a belief about something numinous might be both necessary and sufficient. (The “tightly tied” is vague, but that doesn’t worry me.)

    August 9, 2011 — 23:28
  • It sounds like Conrad might be doing the same thing I did at first. When I read the argument, I got so hung up on the initial premise that I think I didn’t really see it in context of your argument as a whole. I did read on, but I was really just mulling over that premise. As soon as you pointed out the evolutionary component, I reread it and smacked myself on the forehead!

    August 10, 2011 — 10:23
  • Gordon Knight

    I wonder why the argument’s first premise is not
    If naturalism is true, moral realism is false
    I think the argument as it stands begs a quite a few questions. What of an atheistic platonist?

    August 10, 2011 — 11:58
  • rhett

    “A. Evolutionary process P produced both our religious and our moral beliefs.” It also produces our beliefs about the minds of others, including that they have minds. So by the argument above, we have no knowledge of the minds of others.
    Can’t see how Theraveda Buddhism is incompatible with naturalism, could you say more?

    August 12, 2011 — 3:42
  • It may depend whether you’re talking of a variety of Theraveda that accepts rebirth or one that doesn’t. One that doesn’t might be pretty naturalistic, and maybe not really a religion? On the other hand, it may still posit enlightenment as involving the numinous, and numinousness is incompatible with naturalism.

    August 12, 2011 — 7:30
  • Oops. Sorry about the mangled tags. Anyway, the quotation about Comte is from the Catholic Encyclopedia.

    August 14, 2011 — 10:36
  • Good points, Tim.
    I think we might want to distinguish between religious practice and religious belief. One might have religious practice centered on non-religious beliefs. (On the other hand, the belief that this practice is required might be a religious belief, and might be incompatible with naturalism which tells us that humanity is simply constituted by people.)
    Greek religion and animism are interesting questions. Maybe I should weaken the claim that religious beliefs are incompatible with naturalism to the claim that religious beliefs are incompatible with the conjunction of naturalism and the deliverances of modern science.
    Thus, the view that trees can think is compatible with modern science, since it can be taken in a sufficiently dualistic way, and it is compatible with naturalism, since it can come along with a naturalistic theory that trees have brains, but it is not compatible with the conjunction of naturalism and modern science.
    Since contemporary naturalists generally accept modern science, this doesn’t significantly weaken my conclusions.

    August 14, 2011 — 15:48
  • One might have religious practice centered on non-religious beliefs.
    Hmm. Maybe. Religion isn’t going to be a tidy phenomenon susceptible of definitions in terms of necessary and sufficient conditions, more a family resemblance sort of thing. But offhand I think I’d start with religion as a sociological phenomenon, involving (typically communal) prescribed practices of worship and other ritual activities, bound up with attitudes of fear/reverence/awe. So make the practice and attitudes central, and then religious beliefs (where they occur) will simply be whatever ones are closely bound up with the practices and attitudes (e.g., the ones a person might cite to justify the practices and attitudes). So I’d say that a follower of Comte’s positivist religion who had certain beliefs about going to various positivist rituals and about the awesomeness of humanity has religious beliefs. If you want to say that they’re inappropriate or idolatrous or unjustified beliefs, given what a naturalist nowadays ought to think about humanity, that’s fine, but that’s a different claim.

    August 15, 2011 — 7:52
  • Fair enough.
    But I think this is a matter of detail. On my evolutionary just-so story about religious and moral beliefs, it surely wasn’t Comptean positivist religious beliefs that were produced by the processes.
    The beliefs produced by the process are presumably likely going to be ones that posit: (a) karmic systems that go beyond the connections between good behavior and good consequences that we get from modern psychology and sociology, (b) supernatural beings that have codes of conduct they enforce in this life and/or the next, or (c) claims of supernatural grandeur, beauty or majesty that powerfully sway the imagination of the majority in favor of cooperation.
    And these beliefs are such that if naturalism is true, they are false. (That’s a material conditional.)

    August 15, 2011 — 9:33