II. LANG'S SOLUTION: EXPOSITION AND CRITIQUE

The WKR problem has been extensively discussed in recent literature and as Lang notes the problem can be illustrated in a variety of ways.Footnote ⁵ Here it suffices to give just one example. Suppose that a malicious and powerful demon threatens to inflict severe pain on us unless we favour him. That seems to provide strong reason for us to favour the demon, but it is obviously not the case that such a demon is thereby positively valuable. In other words, the reason to favour the demon in this scenario is of the wrong kind from the point of view of value analysis. But as recent debate has shown, it is notoriously difficult to formulate a satisfactory definition of the right kind of reasons. After rejecting a number of recent attempts Lang offers his own proposal, which he calls BPV6:

BPV6 is intuitively plausible as it seems to take care of a good deal of the examples of the WKR problem that have been suggested.Footnote ⁶ But BPV6 faces a number of challenges. First, it seems that BPV6 needs to be interpreted liberally if it is to take care of all examples that have been suggested. For instance, in order for BPV6 to take care of the example mentioned above we must assume that something that saves us from pain that we would otherwise have suffered benefits us. But let us grant that such a liberal interpretation of what it is for something to benefit us is tolerable in the present context. Second, it seems that BPV6 needs to be reformulated slightly in order to take care of other examples that easily spring to mind. Suppose for instance that an evil demon threatens to inflict severe pain on people on the other side of Earth, who are all strangers to us, unless we favour him. That seems to provide reason for us to favour the demon even though we would not benefit from favouring him. According to BPV6 this means that the demon is good. This is obviously unacceptable so BPV6 would have to be reformulated along the following lines:

I take it that such a slight reformulation is still in the spirit of the basic idea behind BPV6.

But third, we can formulate counterexamples to BPV7 that make that basic idea look much more questionable. Suppose that an evil demon threatens to create a number of individuals who suffer greatly unless we favour him. If we favour him he will not bring these suffering individuals into existence. That seems to provide reason for us to favour the demon even though no one would benefit from our favouring him. This is because one cannot benefit individuals who do not and will not exist. According to BPV7 this means that the demon is good. This is obviously unacceptable so BPV7 would have to be reformulated along something like the following lines:

BPV8 is more complex than its precursors, but perhaps not objectionably so. But further problems lie ahead.

Fourth, suppose that an evil demon tells us that unless we favour him he will erase completely something that many people cherish deeply – such as the works of Shakespeare – from collective memory. That seems to provide reason to favour him. But since the demon threatens to erase completely all works of Shakespeare from collective memory our favourable attitude towards him will benefit no one and it will not prevent suffering from occurring. According to BPV8 this means that the demon is good. This is obviously unacceptable so BPV8 would have to be reformulated. How? Well, Lang could suggest that eradicating all works of Shakespeare from collective memory would make our lives less rich and on this basis he could offer the following reformulation:

BPV9 is more complex than its precursors but a more striking problem is that the claim that something would make our lives more or less rich is clearly an evaluative claim. Since BPV9 is an analysis of value that involves an evaluative notion in the analysans it is circular.Footnote ⁷ But perhaps there is another option. The reformulation of BPV9 could mention the supervenience base of the evaluative property appealed to in the analysans, rather than the evaluative property itself. This would give us BPV10:

But it is obvious that only lack of imagination stops us from conjuring up demons that threaten to eradicate from collective memory the works of Plato, Aristotle, Dostoevsky, and what have you. In order to account for this possibly infinite series of counterexamples the reformulation of BPV10 would have to take the form of a hopelessly complex conjunction.

III. A BRENTANO-STYLE APPROACH: OUTLINE AND DEFENCE

The gist of the Brentano-style approach suggested by Sven Danielsson and me is that the buck-passer should distinguish between two types of reasons: reasons for the correctness of an attitude (what we call content-reasons) and reasons to have an attitude (what we call holding-reasons). The former type of reasons trades on Brentano's notion of correctness which applies to conative attitudes. Correctness is a normative notion but crucially, from the fact that some attitude would be correct it does not follow that we ought to have that attitude (i.e. that there are holding-reasons for that attitude). It is however a plausible normative assumption that we normally ought to have correct attitudes. Hence we suggested that content-reasons for an attitude normatively imply holding-reasons for that attitude.Footnote ⁸ Furthermore, we suggested that holding-reasons are analysable in terms of content-reasons: to say that there is holding-reason for some attitude, a, is to say that there is content-reason to favour the occurrence of a, or possibly that there is content-reason to disfavour the non-occurrence of a.Footnote ⁹

Given the Brentano-style approach, the buck-passer can maintain that an object, X, is valuable iff there is content-reason to favour X. This means that in the example given in the previous section there is holding-reason but no content-reason to favour the demon. Favouring a demon that threatens to inflict severe pain on us if we do not favour him would not be correct, but if the threat is credible we ought nevertheless to do so. That there is holding-reason to favour the demon means that there is content-reason to favour favouring the demon, which is the right result since favouring the demon is instrumentally valuable in that it shields us from pain.

Lang criticizes the Brentano-style approach on two counts. First, he argues that the notion of correctness is too evasive and, second, that the approach is only ‘nominally different’ from a suggestion by John Skorupski that Danielsson and I aim to reject.Footnote ¹⁰ I begin with the second criticism.

Skorupski's basic move is to distinguish not between types of reasons for attitudes but between reasons for actions (what he calls practical reasons) and reasons for attitudes (what he calls evaluative reasons).Footnote ¹¹ Utilizing this distinction, the buck-passer can say that X is valuable iff there is evaluative reason to favour X.Footnote ¹² The next move is to say that there can be reason to bring it about that one has some attitude without there being reason to have that attitude. Reasons to bring it about that one has some attitude are reasons for actions, or series of actions. Reasons to bring it about that one has some attitude are thus practical reasons while reasons to have some attitude are evaluative reasons. In the demon scenario in the previous section, there is (practical) reason for us to bring it about that we favour the demon but no (evaluative) reason for us to favour the demon. Hence, the buck-passer need not conclude that the malicious demon is good.

Danielsson and I reject this approach. Lang takes our ground for rejection to be that: ‘if there is a reason to bring it about that one admires the evil demon, then it seems difficult to deny that the demon has properties that make it appropriate to bring it about that one admires him’.Footnote ¹³ But this is not in fact our ground for rejecting Skorupski's approach. (I don't think that Skorupski would be concerned to deny that the demon has properties which make it appropriate to bring it about that one admires him.) Our complaint was that it is difficult to find an independent rationale for making the required distinction between reasons to bring it about that one has an attitude and reasons to have an attitude. In the demon scenario, there is a strong intuition to the effect that there is reason to favour the demon (that is part of the explanation why the WKR problem is so troubling); after all, the reward (of being spared from punishment) is tied to actually favouring the demon and not merely to bringing it about that one favours him.Footnote ¹⁴ So in so far as there is reason to bring it about that we favour the demon this would seem to be so because there is reason for us to favour the demon. It is odd to say that we have reason to bring it about that we be in a state (e.g. of favouring the demon) that we have no reason to be in, and possibly even that we have reason not to be in (on the assumption that there is evaluative reason not to favour, or to disfavour, the demon and not merely no evaluative reason to favour him).

Lang acknowledges that he accepts our criticism of Skorupski (although he misconstrues it). He goes on to say, however, that the Brentano-style approach collapses into something equivalent to Skorupski's approach. This is allegedly because

But here Lang simply fails to note that we state clearly that ‘[the fact] that there is a holding-reason for some attitude does not imply that there is a content-reason for that attitude’.Footnote ¹⁵ The fact that there is a holding-reason for some attitude, a, entails merely that there is a content-reason to favour the occurrence of a, or possibly to disfavour the non-occurrence of a, since, as has already been noted, this is what it means to say that there is a holding-reason for a. One merit of this view is that it accounts for the intuition that there is reason to favour the evil demon who threatens to inflict severe pain on us if we do not favour him. But since this is a holding-reason it does not imply that he is good. We are not committed to Skorupski's dubious claim that there is reason to bring it about that we favour the demon although we have no reason to favour him (and presumably also reason to disfavour him!). Again, while Skorupski's approach distinguishes between reasons for attitudes and reasons for action, the Brentano-style approach distinguishes between types of reasons for attitudes.Footnote ¹⁶

I now turn to Lang's criticism that the notion of correctness is too evasive and that Danielsson and I fail to say anything illuminating about it. As mentioned, the notion of correctness is adapted from Brentano and when we introduced it in our article we said that according to Brentano correctness is analogous to the notion of truth. We noted that according to Brentano, correctness and truth are fundamentally epistemic notions that must be understood with reference to the notion of experiences of self-evident knowledge. We then went on to say that much as we may find it both acceptable and useful to talk about true beliefs without being precise about the nature of the truth predicate, we may also find it acceptable and useful to talk about correct conative attitudes without being precise about the nature of the notion of correctness. We remained neutral between rival metaphysical and epistemological accounts of correctness.Footnote ¹⁷

Lang finds this unacceptable since being non-committal regarding the nature of the truth predicate ‘presuppose[s] a secure grasp of the extension of the predicate “true”’ but – he seems to suggest – we don't have such a grasp of the extension of ‘correct’.Footnote ¹⁸ This last claim is contestable. I am inclined to believe that many people have a fairly secure grasp of the extension of ‘correct’. For instance, I am sure that many people would agree that it would not be correct to favour demons who threaten to inflict severe pain on us if we do not favour them.Footnote ¹⁹

I suspect, however, that Lang's real worry here is one that he develops in more detail towards the end of his article. This is what he calls the directionality problem. To put it briefly: how do we know that ‘it is [not in fact] our intuitions about value that are driving the formulation of the right reasons/wrong reasons distinction’?Footnote ²⁰ If it is, it seems that we have not succeeded in reducing values to reasons. This is a fair worry. Applied to the Brentano-style approach the worry becomes the following twofold one: first, can we really distinguish between holding-reasons and correctness-reasons for an attitude; second, can we do this in a way that does not appeal to more fundamental intuitions about value?Footnote ²¹

In response to the first question, it should be acknowledged that in the demon scenarios most of us have intuitions to the effect that there is reason to favour the demon but there is also reason not to favour him. The way to account for and regiment these intuitions is to say that in such cases there are holding-reasons to favour the demon (we ought to favour him) but there is no content-reason to favour him (it would not be correct to favour him), where holding-reasons to favour the demon do not imply content-reasons to favour the demon. In response to the second question, let me point out that Danielsson and I close our article by posing this very question. We concede that ‘what provides the content-reasons to favour x will be roughly coincidental with . . . the good-making properties of x’.Footnote ²² One might then ask why an analysis of value in terms of content-reasons for attitudes is at all preferable. One attractive feature of such an account that we stress is that it establishes the following plausible link between values and attitudes: necessarily, to claim that an object is valuable is to claim that a favourable attitude towards that object is (would be) correct.Footnote ²³