Wednesday, March 12, 2008

Is There an Ethics of Belief?

Recently I’ve been coming across some thinkers promoting an ethics of belief. Initially the idea sounded a little strange to me: ethics concerns actions—intentions to act, dispositions to act, consequences of action—and beliefs matter ethically only insofar as they are relevant or involved in action, right? As I said, this was my initial reaction, and after some reflection, I still hold to it, although with an important caveat. So let me say what I think this notion gets right, and what it misses.

Take for example Colin McGinn’s claim that his atheism results in part from an ethics of belief. I’ll let him say why in his own words:
“It is often forgotten that atheism of the kind shared by Jonathan [Miller] and me (and Dawkins and Hitchens et al) has an ethical motive. Or rather two ethical motives: one is ethical repugnance at the cruelty, tyranny and oppression of organized religion over the course of human history; the other concerns the ethics of rational belief—how we are obliged to form our beliefs about the world. The first motive is familiar and needs no commentary from me. The second is less widely appreciated, but for some of us it is crucial to the whole discussion. We believe, as an ethical principle, that beliefs about what reality contains should always be formed on the basis of evidence or rational argument—so that “faith” is inherently an unethical way to form your beliefs.”
I will focus on the latter principle. McGinn asserts as a ethical principle of belief (formation and retention) that:
“beliefs about what reality contains should always be formed on the basis of evidence or rational argument”
Let’s call this principle ‘MG’. I see some problems with MG as stated. For one thing, it is too vacuous for serious universal application. Most theists, after all, claim that their belief is informed by serious evidence and argument. Thus, in an empty way, their beliefs satisfy MG. Another problem concerns the concepts of ‘evidence’ and ‘rational argument.’ What are we to make of these? By ‘rational argument’ are we committed to making as many of our beliefs as possible consistent with one another? If so, MG is surely too strong. Is it irrational, and therefore, unethical, for someone to be a dualist? As stated by Descartes, for instance, I take it as a truth of reason that dualism simply cannot be correct. Was Descartes—worse than wrong—immoral? By MG, it would seem so. A similar problem emerges with the notion of evidence. I’m not sure exactly what McGinn means by this term, but presumably he means something like verification, demonstration or proof. But many of our beliefs, especially the most interesting and profound ones, nevermake it to the stage where evidence in that sense applies. I, for instance, don’t think that a belief in God is a false belief—it is a confused belief. Most of the new atheists with whom McGinn aligns himself spend way too much time going straight away to arguments about why the belief is false, and skip over the really hard work of unpacking the confusions involved in the concept of God itself. Let me get personal for a moment and say why I think that one argument for the existence of necessary being is very hard to get around—it is one of Aquinas’ five proofs.
From nothing, nothing is caused. Something must always have existed if anything exists. This universe exists (assumed). Thus, something must always have existed, ie something exists necessarily.
Now, of course this is only a proof that something must necessarily exist, not that a personal God with intentions and an interest in human affairs exists. But so what? It is a theological claim, and even though I don’t believe it, I have a hard time saying why. Am I unethical? Not for this reason, I think.

We should also consider that for many of our beliefs we simply lack any decisive evidence or demonstration either way. Is it consistent with MG to form a belief on such matters anyway? For instance, should I believe that quantum mechanics is a correct description of reality? After all, we have yet to solve the puzzling fact of apparent particle-wave duality, and we still haven’t unified gravity with the electromagnetic forces. More importantly, we have no really good idea how this might be possible. Is it wrong to have beliefs on these matters, then? Perhaps McGinn would suggest that we append a rider to MG, such that in cases where sufficient evidence is lacking decisively for or against the truth of a belief, one ought to withhold judgment. Let’s call this MG’. Well, ok, but how do we apply MG’ to cases like the axiom of choice? Doing set theory requires that one take a stance on the axiom of choice--a 'leap of faith,' as it were--without any rational justification. Is this an example of the intrinsic conflict of goods in the domain of belief? Maybe, but this conclusion seems way overwrought.

Finally, what are we to do about counterfactual beliefs? Maybe there are some counterfactuals that we could exclude from reality, but surely not all of them. So what about someone’s belief that, had Napoleon not been defeated, then Germany would have democratized much earlier than it did? It’s surely fun to consider counterfactuals like this, and we can consider them knowledgably, but of course we really have no decisive way of determining their truth value. Again, is it immoral to believe a claim like this?

The fact is that most of our beliefs have no obvious actionable consequences, and because of this, I have a hard time thinking of them as ethical at all, and this in turn is why I think that my initial reaction is basically the right one. As of now, I can think of no morally sensitive consequence of either the belief that the universe is bounded or that it is infinite, and thus, this belief simply has no moral import. Now, we might suggest one further reworking of MG, to something like the following (MG’’):
Beliefs about what reality contains should always be formed on the basis of evidence or rational argument WHEN those beliefs potentially lead to morally sensitive consequences.
But this is just to say, again, and redundantly, that it is the actions or consequences that matter, not the belief per se. Suppose for instance that someone has a completely vacuous belief in a God—believing that some God exists, but who takes no interest whatsoever in human affairs and has prescribed no rules or norms. This belief, basically, is totally irrelevant to how one behaves and lives. It’s difficult for me to understand how such a belief could be moral or immoral at all--how one could be committing the 'sin of atheism' to believe it.

But let me conclude with my caveat. Despite all that I’ve said above, I do believe in a basically Kantian project of enlightenment, and this makes it requisite that I submit my moral and political beliefs to public, rational scrutiny and, when the public argument is persuasive, I ought, morally, to change my mind. So perhaps we could amend MG one final, less problematic way (MG’’’):
Beliefs about what reality contains should always be formed on the basis of evidence or rational argument WHEN those beliefs are morally or politically salient.
I have no problem with MG’’’, but I’m not sure that it says anything except that, insofar as beliefs are relevant to moral action or consequences, we have an obligation to ensure as best we can that those beliefs are, morally, right. But that’s obvious. I think. And again, this means that there is no such thing as a distinct ethics of belief; there is just ethics.


  1. I think that you are not using the term 'ethics of belief' in the same way people in the literature do. Ethics of belief is used just to signify the study of what normative requirements apply to belief. People in the literature don't think that necessarily when you violate a normative constraint on belief you are acting immorally. You are just violating a normative requirement.

  2. Errol - I take it that you're accusing Michael of confusing two senses of ethics. But it's actually McGinn who confuses them, and thoroughly so. (If you look at his post, he actually says that belief in God, because it involves insufficient evidence, is "the sin of theism.") I think McGinn is just wrong here, and trivially so. Attempts to show that belief in God is, as such, unethical are mostly childish nonsense. Sure, accusing anyone who believes in God of being unethical is good for a soundbite, but when you're making ethical accusations against people like Albert Einstein or Martin Luther King, you'd better have your rhetoric checked. So I think your criticism is fully warranted; but the mistake is really McGinn's.

    Michael - I don't know if this gets you off the hook. McGinn mentions two kinds of norms involved: first, there is an ethics of belief that pertains to the consequences of those beliefs. Second, there is an ethics that is internal to belief-formation itself. But, though you seem to recognize the distinction, you still treat the second as subsumable under the first. You go through various ways of understanding the normativity of belief formation and basically claim that there can't be anything wrong with forming beliefs in non-rational ways in these cases, because holding these beliefs may not have immoral consequences. But this completely begs the question on the issue of whether there is an ethics internal to belief formation itself.

    Second: Your argument is predicated on the notion of inert beliefs. As a thought experiment, this looks like it could work: imagine that you have an irrational belief, but one that has no consequences whatsoever. Is there anything wrong with the belief itself? Well, cute as this is, it's at a level of abstraction where you may think twice before going. If you accept some sort of mental holism--and you do, don't you?--then you must recognize that there are no inert beliefs. More significantly: even if there were inert beliefs, a believer could not possibly know which of his beliefs are inert. So if there is an ethics pertaining to the rationality of beliefs that might have causal consequences, then there is an ethics pertaining to all beliefs, since any of them could potentially have consequences.

    Furthermore: It's not clear to me that inert beliefs can't be unethical. At least in some cases, that seems wrong. Imagine that you think all Latino people are lazy, but this belief is completely inert because, e.g., you never actually meet any Latino people. It still seems like there is something unethical about having the belief itself, and not just because, counterfactually, the belief could influence your future actions.

  3. Errol, I take your point, and I think that I agree with you, actually. However, if by 'ethics of belief' we mean the 'normative requirements that apply to belief,' then aren't we just saying 'reason' or 'rationality' or 'knowledge.' If all that people mean by 'ethics of belief' is, be rational, then there would have been no reason to come up with a new term--epistemology works just fine. I hope that this came across as the point of my post. I guess I could have been clearer.

    Roman, also good points. Let me make this suggestion: why is it not enough to think of an irrational belief as just stupid and wrong? I would say of your example that the person held a very stupid belief, remind him or her that it's not polite to be stupid,and worry that if the person persists in the stupid belief that it might lead to morally wrong consequences should s/he in the future come into contact with Latinos or influence someone who might. But again, its just the consequences and actions that matter morally, not the belief per se.

    As to your point about belief formation--yea, I should have been more clear above, but I would respond similarly to how I would answer Errol: why not just call that reason? There certainly are norms 'internal to belief formation itself,' but those are rational, not ethical norms, and the study of those norms epistemology, not ethics.

    As to the implications of holism to this entire issue, it's an issue I'd have to think more about

  4. M-

    Good points. On whether a belief can be ethically wrong to have (quite apart from the consequences), I think we have differing intuitions. It's possibly that my intuition that racist beliefs are ethically wrong to hold regardless of consequences might just be skewed from some inability on my part to separate them from the consequences, but I'm not sure. Maybe we should hand out surveys?

    The issue of ethical vs epistemic norms becomes important in light of some contemporary approaches that try to derive obligation (of the moral sort) from the rational norms involved in belief or intention formation. I should try to post about this...

    Lastly: if you managed to bring the word "stupid" into respectable philosophical circulation, I'd be thrilled. But I wonder if it would get boring too quickly...

  5. As always, Bayesianism resolves the issue. Your examples strain moral intuitions largely because they are presented as stark dichotomies. If one simply either acts morally or immorally in holding a belief, then many beliefs that seem innocuous somehow make one immoral, which seems quite strange and wrong-headed. However, if we recognize bias in a Bayesian sense as a species of immorality, then most of the intuitive difficulties fall away, or at least there aren’t any problems that an ethics of belief has that an ethics of anything else doesn’t have. First of all, everything is relative to one’s priors: no particular belief is “immoral,” only a failure to make one’s beliefs as accurate as possible—given constraints of time, intelligence, ecetera—is immoral. Similarly, we don’t condemn Plato as a pederast just because he liked to diddle the occasional boy youth; he was just doing what he thought was right. Secondly, immorality exists along a continuum, and being only a little bias is like being only a little immoral: just as it’s wrong to commit adultery, but only a little wrong to look at hot women who aren’t your spouse, you can be a little bit bias but still be a pretty good person. Lastly, you should appreciate Bayesianism as a moral philosophy because it is very Kantian in a sense: since according to Bayesianism your beliefs ought to be the same as the beliefs of the person that you would characterize as the perfect truth-seeker, you must believe that everyone ought to believe as you believe because you ought to believe that everyone ought to believe the truth. There, now I’ve settled the issue for you.


  6. Bro, you write: "everything is relative to one’s priors: no particular belief is “immoral,” only a failure to make one’s beliefs as accurate as possible—given constraints of time, intelligence, ecetera—is immoral." I guess this is the issue in regard to an ethics of belief: is it a moral, in addition to rational, failing to not make one's beliefs as rational as possible, ie, in face of counter evidence, not to adjust one's priors. My point in the post is that I don't think that this is a moral issue. Now, being stupid may potentially lead to morally hazardous action, so, for instance, I probably do have some sort of obligation to make sure that my beliefs about the likely affects of my actions are accurate. IE, if I were to persist in the belief that everybody wishes for my interests to take preference to their own, then I will almost certainly end up doing some wrong things. Thus, I do have an obligation adjust my priors for this sort of belief, but again, only in light of potential consequences of my actions. However, as there are no obvious actions or consequences that follow from whether or not I believe in the axiom of choice, or belief in a bounded universe, I have no MORAL obligation to form true beliefs on these matter. As a rational agent i will no doubt be motivated to form true beliefs, but this is not a moral motivation or obligation.

  7. I thought that you were more or less a Kantian, so I'm surprised to see you taking a consequentialist line on morality. If you go that route, you have to deal with all the problems that consequentialists have to deal with. But more importantly, it's impossible to say with 100% certainty
    (given certain assumptions about priors) that any particular belief won't have consequences for future actions, so it's not clear that your criteria for what beliefs one has an obligation to make accurate don't apply to all beliefs. Also, I'll reiterate that Bayesianism seems to me to be a perfectly logical and attractive extension of Kantianism: Just as you ought to do to others as you would have them do to you, you ought to believe as you would have others believe.

    Me again.

  8. Brau, I don't think that that the type of ethics matters for my argument (consequentialist, deontological, virtue, whadeva). My claim was just that beliefs matter morally only insofar as they affect action or the consequences of an action. There is a Kantian idea here: there's just reason, which has either a theoretical or a practical manifestation. As for the problem posed by globalism and holism, though, yea, I've not thought through what this might mean--because you're right, in a certain sense, how I act and respond in the world is influenced by the entire set of my beliefs as a whole, so that, taken globally, in some sense one could argue that my particular beliefs about, say, Pluto (I might truly believe that it should still be categorized as a planet, for instance) might actually have some relevance for action. Maybe.

  9. On the foundation (feel free to shake me of it) of the Good being happiness and individual "goods" being things that bring about overall happiness, would it be fair to suggest that "belief" itself (as the grammar suggests) is an action. From your initial doubts I can see that you're looking at belief as a foundation for action, or infrastructure in the way that arms and legs facilitate action.

    If I asked you, "Can belief bring about happiness without any action flowing from it?" how would you avoid admitting that belief itself brings us happiness? I would thereby challenged you to find whether the happiness derived from belief actually comes from some action closely connected to it (providing certainty, praying, submitting to higher power, etc.) or admit (I don't at all take this as a given) that belief itself is an action.

    If we find that belief is an action, I suppose there should be no question that ethics apply directly to belief in the same way that they apply to murder or charity.

    (Sorry, I did read over McGinn's thoughts, but if I ignore them, it is because I'm more interested in what is in your head!)

  10. I think McGinn read this, because he has apparently also decided to bring back the word "stupid". See this post.

  11. Thanks. I commented on it (twice- oops)

    I always hesitate to say "stupidity" because it's so vague. I'm probably wrong to use it, because what I'm really talking about is ignorance; stupidity communicates ignorance but it's not clear what else it communicates. Incapacity to know (McGinn's understanding of it)? Apathy about knowing?

    The reason I used it anyways is that it's irrelevant to this discussion why we are ignorant, we just are. It's useful to use such familiar lingo in guiding less sophisticated minds about ethics, but as a general rule I would probably avoid saying "stupid" in most other settings, however. It's only a word, after all!

  12. Anonymous, Michael -

    Thanks. A good exchange.

    To the extent that I can map my otherwise incommensurate and non-transferrable religious convictions onto metrics and ethics, then mapping easily to Bayesian metrics and virtue ethics would likely do for me at about P .90+. I think it’s more an aspiration than a norm implied in Bayesianism to say that, “believe because you ought to believe that everyone ought to believe the truth” (Anonymous). We don’t yet have a differential equation for global conditions nor logical omniscience.

    And even if we did, the divide from "is to ought" is still naturalistic.

    When Michael says he hasn’t thought through the large scale questions of “globalism and holism” as influenced by his “entire set of .. beliefs as a whole,” then Michael touches the exposed nerve of any naive Bayesianism. Michael may feel this large scale question important enough to wrestle out, but Gaia may be one mean bitch enough not to care or respond in kind.

    I say bring every metric to bear: virtue ethicists are divided on whether goods are necessary for net eudaimonia, so bring every utilitarian metric and deontological consideration to bear on ecological questions!

    Just don’t be surprised if the ecological effects of your/my beliefs are about as significant as the Newtonian effects of a single ice crystal in the epsilon ice ring in Uranus on the earth’s solar orbit. Life’s an open system. Dynamics change. Rene Thom, better than Bayes, metricized that.