Friday, March 19, 2010

Maitzen on Ordinary Morality and Atheism (Part 1)

(I struggled to find an apposite image to accompany this post. I settled on this simply because I liked it)

I've previously discussed Stephen Maitzen's work on the implications of skeptical theism for morality. Over the next couple of posts I am going to take a look at the following article:
"Ordinary Morality Implies Atheism" (2009) 2 European Journal for the Philosophy of Religion 107-126.
The idea that ordinary morality has disturbing implications for theism has been aired before. For example, Dean Stretton's article "The Moral Argument from Evil" makes this point, as does Raymond Bradley's "A Moral Argument for Atheism", as does James Rachels's "God and Moral Autonomy".

All of these arguments suggest that there is an incompatibility between ordinary moral obligations and standards (or moral autonomy in Rachels's case) and God's existence. Maitzen attempts something similar in his article.

I'm going to break the discussion of Maitzen's argument into two parts. This part presents the basic outline of the argument. Part 2 looks at some detailed objections to the argument along with Maitzen's responses to them.

1. Some Preliminary Issues
Before becoming embroiled in the intricacies of the argument a few preliminary observations are in order.

First, as with my previous discussion of Maitzen's work, I am going to try to illustrate the argument with an argument map. This means that during the substantive discussion I will pause to present a series of numbered premises and conclusions that I can plug into the diagram at the end. Please note that my numbering does not follow the numbering that Maitzen uses in his article. This may make it difficult to cross-compare, my justification for it is that I found it easier to understand the argument by doing it this way.

Second, the argument is really two separate arguments with incompatible conclusions. The first argument suggests that God's existence implies the absence of moral obligations. The second argument suggests that ordinary morality implies the presence of moral obligations. These conclusions are obviously incompatible.

This leads me to a third observation. Does Maitzen's argument (or indeed Bradley's or Stretton's or Rachel's) offer good support for atheism? Maitzen's article does present an argument that reaches the conclusion "God does not exist".

But, as per his comment below, this is intended merely to show that it is possible to go from ordinary morality to the non-existence of God. To be a true argument for the non-existence of God it would need to mount a full defence of ordinary morality, something which it does not try to do.

My own take on it is that the argument really presents a doxastic dilemma for the theist: either they believe in God or they believe in ordinary morality, they can't believe in both.

The idea is that the commitment to ordinary morality is too strong to give up, so belief in God must give way. This might provide support for George Rey's meta-atheist thesis, the argument that no one really believes in God. Based on his interview over at commonsenseatheism, this is something with which Maitzen agrees.

2. Theism Implies No Moral Obligations
The first, and most difficult, part of Maitzen's argument suggests that coupling theism with some standard resolutions to the problem of evil leads to the conclusion that there are no moral obligations.

a. Principle of Theodical Individualism
We begin by simply conceding that God might exist and then draw out some of the difficulties with accepting his existence. The major difficulty comes from what is called the principle of theodical individualism:
Necessarily, God permits undeserved, involuntary human suffering only if such suffering ultimately produces a net benefit for the sufferer.
This principle contains some important qualifiers: "undeserved" -- to placate those who might feel some suffering is deserved; and "involuntary" -- to placate those who think that people could volunteer to endure suffering. By using the modal operator "necessarily" the principle is designed to cover not only cases of actual suffering but also cases of counterfactual suffering (see Maitzen's comment below for more).

This principle of theodical individualism amounts to a constraint on how God can treat humanity. It is an accepted constraint on any resolution of the problem of evil. This may set some alarm bells ringing since the idea of a constrained God might be anathema to some theists. We can say several things in response.

First, this is a standard Kantian principle that highlights the moral impropriety of treating someone as a means to an end. So it suggests that we cannot cause a person to suffer greatly simply to improve the lot of other people. This is the type of thing we would expect of a morally perfect being (Maitzen gives a thought experiment to make this point).

If that seems to presume too much, we can always fall back on the fact that numerous theistic philosophers accept it anyway (e.g. Eleanor Stump, William Alston and Marilyn McCord Adams) and that it is their belief system with which we are concerned in the first place.

Another objection to TI, not explicitly discussed by Maitzen but clearly lurking in the background, would be an Ockhamist one. The Ockhamist would say that it is meaningless to speak of moral constraints on God because he is the one who gets to define those constraints. We can respond by pointing out that if one is an Ockhamist, one has already abandoned ordinary morality. So the argument does not apply.

We can say further that the Ockhamist will still have serious problems knowing if he/she has any moral obligations due to the somewhat cryptic nature of God's attempts to communicate moral standards to us.

b. Theodical Individualism + God = No Obligations
Maitzen argues that if we accept that God exists, and we accept TI, we must accept one of two things: (a) that we have no moral obligation to prevent suffering or (b) that our moral obligations are defined by God's commands.

Why would we have to accept that we have no moral obligations? Think about it: if TI is true, then all actual instances of suffering ultimately benefit the sufferer. If there is an ultimate benefit you have no obligation to prevent that suffering.

Take a simple example: childhood vaccination. The actual act of administering the vaccine might cause the child to suffer, but because vaccination is ultimately beneficial you are under no obligation to intervene to prevent it. The theist (who accepts TI) is in this position whenever they encounter suffering.

Now they might respond by saying there is nothing to stop them from intervening if they see fit. But this is merely to identify a permission to intervene not an obligation. As we shall see in a moment, ordinary morality suggests that we have genuine obligations to intervene in some situations.

We make this point even stronger if we wish. We can say that if all instances of suffering ultimately benefit the sufferer, we are permitted to actively cause suffering. Again, think about the vaccine example: those who administer the vaccine are permitted to cause the child some suffering because of the ultimate benefit that accrues. Is the same not true for the theist who accepts TI?

The other possibility is that a theist who accepts TI still has obligations emanating from God's commands. There are two problems with this. First, as discussed previously, God's commands are difficult to identify and difficult to extrapolate to novel circumstances. So it is not clear that they provide any reassurance to the theist who wishes to prevent suffering.

The second problem with this response is that it forces the theist into an uncomfortable position. On the one hand, they believe that God has commanded them to prevent suffering. And on the other hand, they believe that God has set things up so that all suffering ultimately benefits the sufferer. These beliefs do not a happy pair make.

Let's stop here and summarise the components of the argument to this point.
  • (1) God exists (or, at least, "I believe God exists").
  • (2) Necessarily God only permits undeserved, involuntary suffering if it ultimately benefitted the sufferer (principle of Theodical Individualism).
  • (2a) This is a basic Kantian principle that we might expect of a morally perfect being.
  • (2b) This is a principle that is consistent with orthodox theistic beliefs.
  • (2c) The Ockhamist objection already accepts the incompatibility of theism with ordinary morality.
  • (3) There are no moral obligations to prevent suffering.
  • (3a) Vaccination example shows that obligations do not survive when we know that a benefit ultimately accrues to a sufferer. Theists are in this situation all the time.
  • (3b) Permissions may survive, but they work both ways, i.e. we may be permitted to cause suffering.
  • (4) Obligations emanate from God's commands.
  • (4a) God's commands are difficult to identify and to apply.
  • (4b) Believing both (i) God has commanded you to prevent suffering and (ii) God has set things up so that suffering produces a net benefit to the sufferer, is not easily reconcilable.
These will be put into the argument map at the end.

3. Ordinary Morality Implies Obligations
The second, and far easier, part of Maitzen's argument is that ordinary morality implies the existence of moral obligations. Maitzen does not define ordinary morality, or attempt to provide a concrete ethical theory that supports the existence of obligations. Instead, he takes the existence of moral obligations as a datum that is in need of an explanation, an explanation that cannot be rooted in theism.

Let's consider some cases that seem to support the genuine existence of moral obligations. First, consider the drowning friend example that I mentioned in a previous post. Here, your friend has fallen into the water, is incapable of swimming, and you are the only one in a position to save him. The question is whether you have an obligation to save him (assume there is no serious risk to yourself)? It seems obvious that the answer is "yes".

Even in situations where there is a risk to ourselves or where we are unqualified to prevent the suffering, there is an obligation to seek appropriate help. This is most apparent from criminal law cases where the court imposes liability on failures to act. 

Later in his article, Maitzen suggests that the criminal law offers a reasonable insight into the demands of ordinary morality. As someone who researches the philosophy of criminal responsibility, I think that is about right (plenty of controversy though). Anyway, the criminal law doesn't impose a general obligation to intervene (where "general" means "applies to everybody"), instead it identifies specific situations in which an obligation arises. These include situations where you are the only person who knows about the suffering, or where you are a qualified official.

To give but one example, in the English case of Stone and Dobinson [1977] 2 All England Reports 341, a couple were held guilty of manslaughter for failing to seek professional help for their anorexic housemate. Because the anorexic woman lived with them they were held to have an obligation to help her even though she refused to eat. (The anorexic was the sister of the woman in the couple, but this was irrelevant to the finding of an obligation to seek help -- even if it was relevant it would not defeat the point that ordinary morality holds to genuine moral obligations).

So there are genuine moral obligations arising from ordinary morality. Now, as hinted at above, we might worry about the theoretical foundations of these obligations. But this is basically irrelevant to the argument being made. We are not concerned with which meta-ethical theory can account for these obligations; we are concerned with whether belief in God can account for them.

Let's summarise the components of this part of the argument:
  • (5) Drowning friend thought experiment suggests that we have an obligation to prevent suffering in certain circumstances.
  • (6) Legal cases, such as the anorexic housemate one, suggest that we have obligations to, at least, seek help to prevent suffering.
  • (7) Therefore, ordinary morality gives rise to moral obligations to prevent suffering.

4. Ordinary Morality Implies Atheism
Okay, so now we are in a position to bring both strands of the argument together. The diagram below provides the relevant map. This follows the mapping procedure outlined at the end of this post.

As can be seen, conclusion (3) and conclusion (7) are in direct competition. And since conclusion (3) follows from belief in God, it seems that ordinary moral obligations cannot coexist with God. Thus, we have a doxastic dilemma for the believer.

In Part 2 we will consider more detailed objections to this argument.


  1. John,

    Thanks for another thorough exposition. I knew about Raymond Bradley's article arguing from biblical horrors, as well as James Rachels's article arguing from Kantian moral autonomy. But I'd never seen Dean Stretton's article giving a moral argument for atheism; it looks highly relevant. I'll study it; thanks for finding it.

    I don't think it's quite accurate to say that "The conclusion reached in Maitzen's article is 'God does not exist'." It's true that the final step in the numbered argument I gave is "God does not exist," but that argument was meant only to show that you can get from the existence of ordinary moral obligations to the non-existence of God. The argument can't show the non-existence of God unless it can show that ordinary moral obligations exist, and it never undertook to show that. That's why I titled the article "Ordinary Morality Implies Atheism" rather than "Atheism is True."

    I guess I tried to fend off the Ockhamist objection in my discussion and defense of premise (3). As you point out, the Ockhamist requires us to abandon (or at least "recalibrate") ordinary morality.

    I also tried to word TI so that it's strong enough to ensure not only that all actual undeserved, involuntary human suffering ultimately produces a net benefit for the sufferer but also that all counterfactual such suffering would thus benefit the sufferer were it to occur. I did that in order to forestall the objection mentioned in footnote 4.

    Looking forward to your next post. Cheers. --Steve

  2. Steve,

    I made some changes to the wording to reflect what you said.

    Dean Stretton makes points quite similar to your own but I find his presentation excessively complex. I'd need to re-read it a few times to fully appreciate it.

    I understand what you mean about anticipating the Ockhamist objection. I just wanted to make it more explicit since it is something I can imagine being raised.

    I didn't want to get into the counterfactual issue because, frankly, I'm not that comfortable with modal logic. But I'll set out what the objection is for anyone who might be reading this.

    An anonymous reviewer of Steve's article suggested that his argument involved an illicit jump from an indicative conditional (what is or will be the case given that some antecedent condition is, in fact, true) to a subjunctive conditional (what would be the case, if some antecedent were, hypothetically, true).

    As a general rule, indicative conditionals do not imply subjunctive conditionals.

    In the case of Steve's argument the jump involved would have been from: (a) if we don't prevent undeserved, involuntary suffering, then that suffering will ultimately benefit the sufferer, to (b) if we weren't to prevent suffering, then that suffering would ultimately benefit the sufferer.

    By including "necessary" in the definition of the principle of TI, this objection is defused. If something is necessarily true, it holds irrespective of what we might happen to do.

  3. Maitzen argues that if we accept that God exists, and we accept TI, we must accept one of two things: (a) that we have no moral obligation to prevent suffering or (b) that our moral obligations are defined by God's commands.

    What about a third option? Our moral obligation is to do the best we can given our necessarily limited knowledge (after all, a god could not give us perfect knowledge -- omniscience -- without making us god). It is a "necessary evil" of creating finite creatures that they cannot know that they are preventing evil but they are still expected to try and the god's commands are guideposts in that effort.

    (I'm not a theist, just trying to work through the argument.)

  4. John D,

    I've now read Stretton's article. I'm sorry I'd missed it in my own literature search. Although, as you said, he makes points similar to those I made in my article, I'm relieved to find that he doesn't wholly anticipate my argument. In fact, I think his argument falls short because his stated moral principles are too consequentialist to capture our ordinary moral reasoning. His crucial "Jones Principle" relieves us of a moral obligation to intervene to prevent suffering provided we know that there's an objective justification for the suffering that will obtain regardless of whether we intervene. But his explication of "objective justification" (by way of "Jones's case #2") equates it to any outweighing good. It doesn't require, as I think ordinary morality does, that the outweighing good (in virtue of which we needn't intervene) accrue to the sufferer himself. For suppose I know that my allowing a child to endure excruciating pain of (say) 100,000 pain units will save 1 million other children the mild discomfort of 1 pain unit each. The fact that 1 million outweighs 100,000 hardly justifies my exploitation of the child. Now, I think there's evidence in Stretton's article that he in fact intends a more individualistic moral requirement, but it's never made explicit.

  5. Steve,

    I see that he doesn't make it explicit. That said, in Jones's Case 1 the objective justification for leaving him on the life support machine is that certain benefits will accrue to him. Giving Stretton the benefit of the doubt, you could assume Jones Case 2 envisages the same.

    A related point: one thing that strikes me is the potential similarities between the case of a theist deciding whether or not to intervene to prevent great suffering (knowing there is a God, knowing there is an objective justification for the suffering, but unsure as to whether his intervention is central to that justification) and Newcomb's problem. Is that something you (or others you know of) have ever explored or thought of exploring?

  6. John Pieret,

    I think the response would be (and maybe Steve can comment on this) that your third option at best gives the theist permission to intervene. In other words, it says that they can intervene if they wish to do so but they are under no absolute obligation to do so (i.e. no one will deem them morally accountable for failing to do so).

    The argument is that ordinary morality holds to strong obligations to intervene, not mere permissions. In other words, ordinary morality does hold people accountable for failures to intervene (omissions). It is difficult to see how strong obligations of this sort can remain if we accept TI.

    Your point is somewhat similar to Eleanore Stump's response to this issue, which I will cover in part 2.


  7. John D,

    @ 9:03 PM:

    (1) I agree that a charitable reading of Stretton puts him very close to what I say some ten years later. Oh well. The online Philosopher's Index lists four articles published by Stretton, none of them the article in question. It eluded my lit search.

    (2) As for whether a human agent's intervention could be central to the justification of some divinely permitted human suffering: I read TI as prohibiting God from planning or allowing one person's undeserved, involuntary suffering for the purpose of giving another person the chance to intervene; no perfect being, I argue, could indulge in that mere use of a person. I wouldn't myself try to compare this situation to Newcomb's problem, since I claim in this article that no one can understand the setup of Newcomb's problem to begin with.

    @ 9:11 PM: I'm not sure I understand what John Pieret was suggesting, but I agree with what you say in your response to him.

  8. John D,

    Would you please email me at lukeprog [at] gmail? I have a question.

  9. Steve,

    On reflection what I said about Newcomb's problem is covered by your "necessarily..." qualification. I think this proves my lack of comfort with modal logic!

    I read your article on Newcomb and I must say it's probably the best defence of the regress interpretation that I've come across. I used to think it suffered from a prediction-regress, then I changed my mind because I figured the prediction wasn't necessarily playing a role in the decision. But as you (and co-author) point out it really reflects the kind of deliberational instability that arises in many game theory problems.

    I'm reading through Gary Drescher's arguments on this at the moment (from his book Good and Real) and he's changing some of my previous opinions. He defends the one-box solution on the grounds that it can show how counterfactual reasoning is meaningful in a deterministic universe. Interesting stuff.

  10. I think the response would be (and maybe Steve can comment on this) that your third option at best gives the theist permission to intervene. In other words, it says that they can intervene if they wish to do so but they are under no absolute obligation to do so (i.e. no one will deem them morally accountable for failing to do so).

    Hmmm. Do children have no moral obligation even if we have no expectation that they will succeed in meeting our adult expectations? We can't tell children that they were wrong unless we hold them to an adult understanding of morality?

    What I'm trying to say, Stephen, is that theism may not require an absolute (divine) standard of good human behavior. Rather, a god (especially an omnisient one) could hold us to a standard, not so much being objectively right, but of being right-minded.

  11. This comment has been removed by the author.

  12. I fail to see why (4b) is a problem. Specifically, part of the sufferer's allotment of suffering is the time it takes us to respond and alleviate it. Note that this already includes free will, but is incompatible with open theism - the just allotment of suffering already includes us fulfilling our obligation or not, as the case may be.

    I'd note TI + free will make strange bed fellows - you can choose to inflict involuntary undeserved suffering on another, and this would be to his benefit... I don't think it's an outright contradiction, though.

    Yair Rezek

  13. Yeah I didn't phrase that very well. The point is that moral obligation could not survive knowing both of these things, not that they are logically or otherwise contradictory.