Saturday, February 5, 2011

Ethical Intuitions (Part I)

Consider two theses the moral skeptic wants to defend:
ST1: Moral knowledge is impossible.
ST2: Justified moral judgment is impossible.
The intuitionist denies both. In this post, I want to look at some of Sinnott-Armstrong’s arguments for moral skepticism.

The first of these arguments draws upon recent empirical research into moral intuition and judgment:
Until the last decade of the twentieth century, philosophers and psychologists usually engaged in their enterprises separately. This was unfortunate, because it is hard to see how to determine whether certain moral intuitions are justified without any understanding of the processes that produce those intuitions. We are not claiming that psychological findings alone entail philosophical or moral conclusions. That would move us too quickly from ‘‘is’’ to ‘‘ought.’’ Our point is different: moral intuitions are unreliable to the extent that morally irrelevant factors affect moral intuitions. When they are distorted by irrelevant factors, moral intuitions can be likened to mirages or seeing pink elephants while one is on LSD. Only when beliefs arise in more reputable ways do they have a fighting chance of being justified. Hence we need to know about the processes that produce moral intuitions before we can determine whether moral intuitions are justified.

Let’s suppose that the available evidence does strongly suggest that moral intuitions can be influenced by factors that we take to be morally irrelevant. Specifically, let’s suppose that the evidence does support the hypothesis that our intuitive moral judgments are subject to framing effects. When this happens, we might classify equivalent options as morally different as a result of how these options are presented. Because of this, Sinnott-Armstrong thinks that the intuitionist view is sunk. The intuitionist says that our intuitions can justify our moral beliefs without needing any independent support from non-moral beliefs. If they’re subject to these effects, he thinks that the justification of judgment depends upon whether we can provide independent support that’s not dependent upon our unreliable intuitions.

It seems that Sinnott-Armstrong has this sort of argument in mind:
(1) The empirical evidence suggests that your moral intuitions are unreliable.
(C) Thus, your moral intuitions can neither justify your moral judgments nor serve as an adequate basis for moral knowledge.

One way to take the argument is as resting on the implicit assumption that intuitions must be reliable if they are to justify moral our moral beliefs and a further assumption that knowledge requires justification. It’s not clear whether the argument rests on (i) the claim that intuition is unreliable or (ii) the claim that the evidence suggests that (i) is true. Let’s bracket (ii) for now and suppose that (i) is true.

One problem with this argument if taken as an argument against ST2 is that it assumes that justification requires reliability. This is a matter of some controversy, but many epistemologists are moved to say that there is not on the basis of a rather simple thought experiment:
I think the evil demon hypothesis (or its contemporary neurophysiologist version) uncovers a defect in the Reliabilist position. We can see this by supposing the hypothesis to be true. Imagine that unbeknown to us, our cognitive processes (e.g., perception memory, inference) are not reliable owing to the machinations of the malevolent demon. It follows on a Reliabilist view that the beliefs generated by those processes are never justified.

Cohen offers these remarks concerning his example:
It strikes me as clearly false to deny that under these circumstances our beliefs could be justified. If we have every reason to believe e.g., perception, is a reliable process, the mere fact that unbeknown to us it is not reliable should not affect it’s justification … My argument hinges on viewing justification as a normative notion. Intuitively, if S’s belief is appropriate to the available evidence, he is not to be held responsible for circumstances beyond his ken.

If Cohen is right (more on this later), the argument I’m attributing to Sinnott-Armstrong fails because the justification of a belief doesn’t depend upon whether the methods that led to its adoption are reliable.

Suppose, however, that reliability is necessary for justification and for knowledge. The hypothesis that intuition is unreliable might be true, but the empirical evidence gives us no reason to think it’s necessarily true. If so, even if we assume that (i) is true, it couldn’t support an argument for ST1 and ST2 since these theses imply that moral knowledge and moral justification are impossible, not simply unattainable as a matter of contingent fact.

Does the argument do any better if we read it as relying on (ii) rather than (i)? Suppose there is relatively strong evidence that intuition is an unreliable guide. Evidence of unreliability (which is misleading evidence, since we’re assuming (ii) is true, not (i)) can defeat justification, but only if someone has it or ought to have it in her possession. There are certainly some moral philosophers who have this evidence in their possession and it might be that some of them ought to know better than to trust their intuitions. The folk, however, don’t have this evidence and it is counterintuitive to say that they should have known better than to take their intuitions at face value before consulting the literature. So, I don’t think the argument for ST2 is persuasive if we assume (ii).

The argument doesn’t succeed as an argument for ST1. Evidence of unreliability (which isn’t the same thing as unreliability) doesn’t defeat someone’s knowledge unless that evidence is in someone’s possession or ought to be in her possession. If this is right, the best that the argument could do, assuming that (i) is correct, is show that we don’t have moral knowledge as a matter of contingent fact. It doesn’t tell us anything about the epistemic status of moral judgments made by those who don’t have the empirical evidence that figures in the argument against intuitionism.

The Threat of Moral Nihilism

In this section, I want to look at another of Sinnott-Armstrong’s arguments against intuitionism. He endorses moral nihilism, the view that there are no moral properties or moral facts. If this view is correct, our moral judgments are systematically false. Even if this view isn’t correct, it still poses a significant epistemological problem. The moral realists might think that moral facts or properties provide the best explanation of our moral judgments, but the nihilists have their own non-moral explanations as to why we make the moral judgments that we do. These rival explanatory hypotheses are skeptical hypotheses that threaten our moral judgments much in the way that, say, the hypothesis that we’re deceived by a Cartesian demon threatens our perceptual judgments. Sinnott-Armstrong’s second skeptical argument can be stated as follows:

(1) You’re not justified in believing that moral nihilism is false.
(2) You know that if it is wrong to torture babies just for fun, moral nihilism is false.
(3) If you know moral nihilism would be false if it is wrong to torture babies just for fun and you justifiably believed that it’s wrong to torture babies just for fun, you would be justified in believing that moral nihilism is false.
(C) Thus, you’re not justified in believing that it’s wrong to torture babies just for fun.

How does this skeptical argument fare?

In his discussion of the arguments for external world skepticism, some have argued that the claim that we’re not justified in believing skeptical hypotheses are false is the sort of claim that the skeptic should have to argue for. It’s not the sort of thing they should be allowed to argue from until they earn the right to do so. I agree. A similar point applies here. Sinnott-Armstrong does offer this justification for (1):
To be justified … the believer must have some way to rule out moral nihilism. To try to do that, moral intuitionists might simply cite a moral belief that is contrary to moral nihilism … Here’s one example:
(T) It is morally wrong to torture innocent children just for fun.
… But suppose that a moral nihilist appears on the scene and denies (T). What could the moral intuitionist say against such a moral nihilist? Not much. Moral intuitionists can point out that (T) seems obvious to them … However, to appeal to such a moral belief in an argument against moral nihilism clearly begs the question … Such moral beliefs appear obvious to almost everyone who is not a moral nihilist, but that appearance is just what would be predicted by the moral nihilist’s hypothesis that all moral beliefs are evolutionary or cultural illusions … When both of two hypotheses would predict an observation, that observation cannot be used as evidence for one as opposed to the other.

Can we “rule out” moral nihilism? Should the intuitionist say that we could rule out nihilism? It depends upon what it takes to rule out the nihilist’s view.

There are two natural ways of reading this talk of ruling out a view. First, we might say that you can rule out views that you know are incompatible with propositions you know are true. Second, we might say that you can rule out views that you know are incompatible with your evidence. On the first reading, it looks like the justification offered for (1) comes to this: you don’t know that (T) is true. Doesn’t this beg the very question at issue? The intuitionist view is that we have some non-inferential moral knowledge. The selection of (T) was arbitrary. It was chosen because it was thought that if we know anything non-inferentially about morality, it would be (T). Not only does this justification for (1) beg the question, it misses its mark. On this reading, you can’t justifiably believe something unless you can rule out being mistaken. In other words, you can’t justifiably believe something unless you know that you’re not mistaken. You shouldn’t think that unless you think justifiably believing something requires that you know that it’s true.

On the second reading, it looks like the justification offered for (1) comes to this: your evidence is consistent with (~T). I think the argument is question begging on this second reading, too. To see this, let’s think about what the intuitionist should say about the evidence we have for our non-inferentially justified moral beliefs. Suppose, for example, that you think your evidence consists of all and only those propositions that you know. If E=K is correct, the second reading of Sinnott-Armstrong’s defense of (1) begs the question for the same reason the first did and it also misses its mark. Now, I don’t think that E=K is correct, but I also don’t think the difference between E and K is great enough to save Sinnott-Armstrong’s argument. Suppose you know p non-inferentially. Now, suppose you know some other proposition, q, is a logical consequence of p. Does your evidence for q include p? It’s clear that p stands in the right support relation to q to be potential evidence for q, but the question that interests me is whether p is part of your evidence for q. According to Conee and Feldman, your “ultimate” evidence consists of those things that provide evidential support where you don’t need independent evidence to treat these things as evidence. Given this gloss on evidence, we ought to say that anything you know non-inferentially belongs to your stock of evidence. Thus, it looks like Sinnott-Armstrong would have to say that your evidence is consistent with (~T), in which case he’d have to say that your evidence includes nothing that entails (T). Among other things, (T) entails (T). So, it looks like Sinnott-Armstrong can only say that your evidence doesn’t rule out (~T) if your evidence doesn’t include (T). So, his skeptical argument gets off of the ground only after establishing that you don’t know (T) non-inferentially. It’s still grounded.

I suspect that the root of the problem is that he’s assuming a dialectical conception of evidence, a conception of evidence on which your evidence against a theory is limited to facts that would be acceptable to proponents of this theory. I think this is a defective conception of evidence. It’s one thing to say that it’s not sporting to appeal to obvious mental facts in a public debate with eliminativists and another to say that you oughtn’t treat such obvious facts as the basis for judgments about whether to head to the medicine cabinet for aspirin. If ruling out moral nihilism required citing evidence acceptable to the nihilist, I don’t think the intuitionist should claim that she could rule out this view since it’s not an obvious consequence of intuitionism that the nihilist’s view is contradicted by her own evidence. There might be a fourth way of understanding this talk of ruling out, but I’ll wager that the fourth way will suffer from one of two defects. Either it would beg the question against the intuitionist to say that she cannot rule out the nihilist’s view (as with the first two readings) or the fact that the intuitionist cannot rule out the nihilist’s view tells us nothing about whether the intuitionist’s moral beliefs are justified (as with the third reading).


Anonymous said...

This seems right, except one point. You say at one point that SA is relying on reliability being a necessary condition on justification. That's not right. He is only relying on this weaker claim: S's belief that p is justified only if S is not justified in believing that the method by which S came to believe p is unreliable. That sounds plausible to me, and I'm no reliabilist. Track records can matter for evidentialists.

Jon Matheson

Daniel said...

I really like this analysis, but I have trouble buying it. Let's simplify the evil demon case. Suppose Bob's perception works perfectly well. Everything he senses in the external world is veridical. What the evil demon did was put a switch on the back of his head. Once the switch is flipped to 'off', Bob's perception is no longer veritical and his senses are systematically deceiving him. I take it that what Cohen is saying is that Bobs senses are unreliable in some sense, but the switch is beyond his ken. I take his argument to be the following:

(1) Bob's perception switch may be on or off
(2) The perception switch is beyond Bob's ken
(3) By 2, Bob has no reason to believe his perception switch is off
(4) Therefore, Bob's perception is justified

If this is a proper reading of Cohen's argument, I'm not comfortable with it. It seems to suggest that the veridicity of my perception is neither here nor there for justification. I'd think that it is. After all, if it turned out that his switch was off, I'd want to say that his experiences are not justified. I don't think that's an unreasonably skeptical claim. Maybe the term 'justified' is just too vague, here.