Here is the second of the two threads, on Jonathan Way's "Transmission and the Wrong Kind of Reason." Critical precis begins below the fold.
Jonathan Way’s paper offers a novel and fascinating argument defending the skeptical solution of the Wrong Kind of Reasons Problem (WKR problem) for the Fitting-Attitudes analysis of value (FA-analysis). According to the skeptics, such as Gibbard, Parfit and Skorupski, there really are no wrong-kind reasons. The incentives for attitudes, which one might be inclined to consider as wrong-kind reasons for the attitudes in question, are in fact right-kind reasons for wanting these attitudes or for bringing them about. Thus, the “WKR skeptics” claim that the WKR-problem is spurious. The “WKR defenders”, who assume the existence of wrong-kind reasons, cannot dismiss the WKR problem in this way.
The standard argument used by the skeptics appeals to one of the two “common marks” of the wrong-kind reasons: to the difficulty of responding to those considerations in a direct way, by an immediate adoption of the relevant attitude. (Cf. Mark Schroeder’s first “earmark” in “The Ubiquity of State-Given Reasons”, in the same issue of Ethics. The second “common mark” Way mentions corresponds to Schroeder’s third earmark: The wrong-kind reasons might make the relevant attitudes required, but they do not bear on their correctness; they do not make them “fitting”.) According to the standard skeptical argument, the fact that we cannot respond to these consideration by a direct adoption of the attitudes they provide incentives for shows that they aren’t reasons for the attitudes in question.
Way finds this argument appealing but also problematic. As he points out, “it is difficult to specify a notion of ‘responding to a reason’ that will vindicate its … premises”. He therefore sets this line of reasoning aside and instead proposes a new argument that turns on the pattern of reason transmission.
Way’s argument takes off from the plausible idea that reasons transmit: “there is often reason for one action or attitude because there is reason for another.” (p. 493) He begins with describing a transmission pattern for actions (Action Pattern): “If there is a reason to A, then the fact that B-ing facilitates A-ing is a reason to B.” (p. 494) B can facilitate A in at least three ways: by being a means, an enabling condition, or a “constitutive means”, i.e. a way of A-ing. There are qualifications that need to be made if we are to accept Action Pattern, such as the requirement that B is not a superfluous facilitator, but Way takes it that given the appropriate qualifications the Action Pattern can be assumed to hold. The transmissions patterns for right-kind reasons for different attitudes are similar to the Action Pattern. But for attitudes the right-kind reasons are transmitted across the facilitative relations between their objects. Thus, given the analogous qualifications as in the case of the Action Pattern, the “Intention Pattern” looks like this: “If there is a reason of the right kind to intend to A, then the fact that B-ing facilitates A-ing is a reason of the right kind to intend to B.” (p. 496) And the “Desire Pattern” (p. 497) has the same format: “If there is a reason of the right kind for you to desire x to F, then the fact that y’s G-ing facilitates x’s F-ing is a reason of the right kind for you to desire y to G.” (It is somewhat unclear to us why Way needs this complex formulation. The idea, after all, seems to be that if you have a reason of the right kind to desire A and B facilitates A, then you have a reason of the right kind to desire B.) Transmission patterns for other attitudes concerning right-kind reasons are not described in any detail, even though Way mentions that for beliefs right-kind reasons transmit across logical implications. Given that the focus is on the WKR-problem and thus on the Fitting-Attitudes account of value, it is a pity that there is no intimation concerning transmission of right-kind reasons for such pro-attitudes as admiration, awe, love, etc. But perhaps the explanation is that, unlike as in the case of desire, the right-kind reasons for these other kinds of pro-attitudes do not exhibit any general patterns?
As for the (purported) reasons of the wrong kind, the pattern of their transmission is different. The wrong-kind reasons are transmitted across the facilitative relations between attitudes themselves and not between the corresponding relations between their objects. Subject to appropriate qualifications, we thus get the following “Wrong Reason Pattern: “If there is a reason of the wrong kind for attitude A, then the fact that attitude B facilitates attitude A is a reason of the wrong kind for attitude B.” (p. 498)
Now, Way presents a challenge to the WKR defenders: What is the explanation of the Wrong Reason Pattern? Why do wrong-kind reasons transmit in this way?
He considers this question to be easily answerable by the WKR skeptics. On their view, the alleged wrong-kind reasons for attitudes are in reality right-kind reasons for wanting those attitudes (or reasons for bringing them about). Thus, the transmission pattern for wrong-kind reasons reduces to the transmission pattern for right-kind reasons: Transmission of reasons goes in this case through the facilitative connections between objects of attitudes of higher order. If there is a right-kind reason to want an attitude A and an attitude B facilitates A, then there is a right-kind reason to want B.
Way takes it that from the point of view of the skeptic this provides “a straightforward explanation of why the Wrong Reason Pattern appears to hold.” (p. 502) Is it so? We are not fully convinced. First, Way doesn’t explain why wrong-kind reasons for attitudes appear to be reasons for those attitudes and not just reasons for wanting them or bringing them about. Some psychological explanation of this alleged misapprehension would need to be a part of the WKR skeptic’s account. Second, Way doesn’t explain why the right-kind reason patterns are valid. Thus, his reduction of the wrong-kind pattern to a right-kind pattern leaves us with some questions unanswered.
But what about the WKR defenders? Way considers two avenues the defenders might try to explore in order to explain the Wrong Reason Pattern. The first one takes its departure from the WKR defenders’ own account of the distinction between the right and the wrong kinds of reasons. There are several such accounts but Way convincingly shows that most of them aren’t helpful in this respect: They focus on the right-kind reasons, since it is in terms of those that the FA-analysis is supposed to be framed, and thus they have little to say about reasons of the wrong kind. The account proposed by Sven Danielsson and Jonas Olson is the most informative in this respect, but even that one ultimately fails to meet the challenge. Since Way’s discussion of the Danielsson-Olson proposal to a large extent follows the same thread as his discussion of the second way to meet the explanatory challenge, we will only focus on the latter approach. The idea here is to account for the Wrong Reason Pattern in terms of the value of the attitudes. This is, in fact, how wrong-kind reasons often are intuitively explicated: The suggestion is that such reasons for an attitude have to do with the value of the attitude itself rather than with the value of its object. Thus, to take the well-known example, the reason to desire a saucer of mud has nothing to do with the value of the object of this desire, but everything to do with the value of the desire itself—with its being a means to avoid punishment by the evil demon. This makes it a reason of the wrong kind.[1]
On the "Value-Based Theory," "[f]or p to be a reason of the wrong kind for attitude A is for p to be a respect in which it is good to have attitude A" (p. 506). On this theory, the Wrong Reason Pattern is very easy to explain: If reasons of the wrong kind have to do with the value of an attitude itself, then they obviously transfer from an attitude A to an attitude B that facilitates A. If A has value, then B thereby has a derived value and thus there are reasons of the wrong kind to have it.
However, the problem with this “Value-Based Theory” of the wrong-kind reasons is that, as it stands, it makes the FA-analysis circular: “It looks circular to hold both that the valuable is what there is reason to value [as the FA-analysis would have it], and that reasons to value are to be understood in terms of what is valuable.” (p. 506) In our view, this is perhaps not as worrying as it looks: To begin with, it is by no means obvious that a proponent of the value-based approach to wrong-kind reasons would want to analyze such reasons for an attitude in terms of the value of the attitude in question. It is more plausible to see the claim that the value-making features of an attitude provide reasons (of the wrong kind) for this attitude as a substantive view about (wrong-kind) reasons and not an analysis of this notion. In that case, there is no circularity in the analysis. But even if we waive aside this response, it is possible to argue that the FA-analysis can be informative even if it is circular. As we suggest elsewhere, “[c]ircular analyses still allow us to exhibit structural connections between central concepts (value, reason, pro-attitude). Thereby, they can provide relevant information to those who have the concepts but are not clear about their mutual relationships” (Rabinowicz & Ronnow-Rasmussen, "Buck-Passing and the The Right Kind of Reasons," Philosophical Quarterly 56 (2006): 114-120, p. 120).
Anyway, Way thinks that the WKR defenders who are wedded to the FA-analysis can break out of the circle if they manage to define right-kind reasons without appealing to the notion of value. For it is only such reasons, and not reasons of the wrong kind, that the FA-analysts appeal to. However, in Way’s view, combining the Value-Based Theory with the FA-analysis implies Reductionism—the claim that “[f]or p to be a reason of the wrong kind for attitude A is for p to be a reason of the right kind to want attitude A.” (p. 507) If we understand Way correctly, Reductionism makes an identity claim: It identifies the property of being a reason of the wrong kind for A with the property of being a reason of the right kind to want A. The argument for this implication from the Value-Based Theory to Reductionism is that, on the Value-Based Theory, for p to be a wrong kind of reason for an attitude is for p to be a feature in virtue of which this attitude is valuable, and thus, on the FA-analysis, a feature that is a right kind of reason for favouring the attitude in question. And wanting is a paradigmatic form of favouring. Is this a good argument? We doubt it. What it proves is only that the following equivalence holds: p is a reason of the wrong kind for attitude A if and only if p is a reason of the right kind to want attitude A. But an equivalence is not an identity. At least prima facie, being a reason for an attitude does not seem to be the same as being a reason for another attitude, of the higher order. Thus, the WKR defenders who appeal to the Value-Based Theory and accept the FA-analysis might well refuse to accept Reductionism.
But let us round up our presentation. Reductionism, if accepted, goes some way towards meeting the explanatory challenge, since it allows us to reduce the Wrong Reason Pattern to the Right Reason Pattern for higher-order attitudes.[2] However, as Way points out, what distinguishes Reductionism from WKR skepticism is the reductionist’s claim that wrong-kind reasons for attitudes are reasons for those attitudes. This is the central claim of the WKR defenders and, in Way’s view, the reductionists have not presented any satisfactory argument on its behalf. Which means that this attempt to meet the explanatory challenge also fails, after all.
Here are two problems that we would like to put forward for discussion.
(1) It is interesting to note that Way does not draw the distinction between right- and wrong-kind reasons when it comes to actions. This is surprising, because one would think that the distinction is applicable in this case as well. If we think of an action as bringing about a state- of-affairs, then the reasons to perform the action that have to do with the value of the state-of-affairs in question are right-kind reasons, while the reasons that have to do with the value of bringing that state about rather than with the value of the state itself are wrong-kind reasons. This is analogous to the case, of, say, intentions. The reasons that have to do with the value of what is intended are right-kind reasons, while the reasons that instead have to do with value of the intention itself are reasons of the wrong kind. Or, if we take the case of, say, desire: The reasons that have to do with the value of the object of desire are right-kind reasons, while the reasons that instead have to do with value of the desire itself are wrong-kind reasons.
Now, the thing to note is that in the case of actions it seems very convincing to say that wrong-kind reasons are bona fide reasons for action, and the skeptic’s position in this case would appear to be highly counter-intuitive. Why is it so? What’s the relevant difference between actions and attitudes? Well, the difference seems to do with the fact that actions can be done at will, while attitudes are not under direct voluntary control. This means that what makes the skeptic’s position strong in the case of attitudes is precisely the argument that we are not responsive to wrong-kind reasons for attitudes in the direct way, in contrast with the right-kind reasons. However, as pointed out above, Way in his paper puts this argument aside and endeavors to present a different argument against wrong-kind reasons—an argument that turns on their transmission pattern. The problem is that the latter argument would be just as applicable to the case of wrong-kind reasons for actions. This makes one wonder how good Way’s argument really is, in view of the fact that the WKR skeptic’s position is so counterintuitive in the case of actions.
One might object against our reasoning above by an appeal to the common marks of the wrong-kind reasons. These marks are absent in the case of reasons for actions that have to do with the value of bringing about a state-of-affairs rather than with the value of that state itself. As we have seen, it is not difficult to perform the action for those reasons and they also bear on the correctness of the action. However, it seems to us that this only shows that the so-called “common marks” do not characterize wrong-kind reasons in general. Some such reasons do not exhibit them at all. At least, this is so if we relate the distinction between the right and the wrong reasons for an action or an attitude A to the distinction between the value of the object of A and the value of A itself.
(2) In his paper, Way assumes that the transmission pattern for wrong-kind reasons requires explanation. But he does not consider the similar challenge concerning the transmission pattern for the right-kind reasons. Why this difference? It seems to us that the answer might have to do with the FA-analysis of value. If we accept that format of analysis, then we take it that the value of an object consists in the existence of right-kind reasons for various positive atitudinal and behavioral responses towards the object in question. Now, if x is valuable and y facilitates x, then—one might argue—y thereby has a derived value, at least ceteris paribus, which means that there normally are right-kind reasons to positively respond to y as well. This explains the transmission pattern for right-kind reasons that follows facilitating relations between objects. The mystery about wrong-kind reasons is that there doesn’t seem to be a comparable explanation for the transmission pattern across facilitating connections between attitudes towards objects.
Now, we think that the solution to this mystery might be found in the value-based approach to wrong-kind reasons. Let’s take the case of actions first, before we consider the case of attitudes. Consider an action A. If this action is valuable, then, obviously, there is a reason to perform it. The reason in question is the feature that makes the action valuable. If that feature makes the action valuable rather than its object (i.e. the state-of-affairs that it bring about), then the reason it provides is of the wrong kind. Since it is A itself that is valuable, then an action B that facilitates A has a derived value. So there is a reason to perform B as well. This reason, the fact that B facilitates a valuable action, is of the wrong kind, since it makes B itself valuable, independently of whether B’s object, the state B brings about, has any value.[3]
Now, if we move from actions to attitudes and like Way set aside the non-voluntariness issue (see above), we can similarly argue that a value-making feature of an attitude A provides a reason for the attitude in question. If this feature makes the attitude valuable rather than its object, then the reason it provides is of the wrong kind. Since it is A itself that is valuable, then an attitude B that facilitates A has a derived value. So there is a reason for B as well. This reason, the fact that B facilitates a valuable attitude, is of the wrong kind, since it makes B itself valuable, independently of whether the object of attitude B has any value.[4] The value-based approach thus gives us an explanation of the Wrong Reason Pattern that we guess many WKR-defenders would be willing to accept. It is not clear to us that Way has provided any convincing argument against this proposal.
_________________
[1] In footnote 29, Way seems to suggest that the alleged reasons of the wrong kind always have to do with instrumental, or at least non-final, value of the attitude. This cannot be correct. To take an example, a proponent of the axiological view that locates all final value in certain kinds of mental states might think that some such finally valuable mental states consist in favouring, for their own sake, certain non-mental objects, which – by hypothesis – lack any final value. If a proponent of such a view takes it that we have reasons to have the attitudes in question, then these reasons are of the wrong kind, even though they are provided by the features that make these attitudes finally valuable. For this example, see Rabinowicz & Rønnow-Rasmussen, “The Strike of the Demon: On Fitting Pro-attitudes and Value”, Ethics 114 (2004), 391-423. [2] This route from the Value Based Theory via Reductionism to the explanation of the Wrong Reason Pattern is, in fact, unnecessary. As we have seen, the Value Based Theory itself provides an immediate explanation of the pattern in question provided that the circularity issue mentioned above can be dealt with. And if it cannot be dealt with, then what’s the point of appealing to that theory in conjunction with the FA analysis in order to derive Reductionism? [3] It would be different if the facilitator of A were B's object–the state B brings about. Then that state would have a derived value. Therefore, the reason to perform B would then be of the right kind, if we assume that bringing about a state-of-affairs is a positive response to that state of affairs. [4] But in the special case in which this facilitating attitude B is a proattitude towards A, there is also a reason of the right kind for B. The reason of the right kind has to do with the fact that B is a positive response to a valuable object (attitude A), the reason of the wrong kind is that B facilitates this object.
Dear Wlodek and Toni,
Thanks for the extremely thorough and insightful critical precis. Regarding the first problem that you put forward for discussion, why can’t Way just deny that there are reasons for actions over and above reasons for intending to perform those action? What would it mean to have a reason to do x over and above a reason to intend to do x? (It makes no sense to me to say that I have a reason not only to intend to do x, but also to do x.)
Also could you perhaps give a concrete example where it is intuitive to think that there is a wrong-kind reason to do x, because I didn’t quite follow this: “If we think of an action as bringing about a state- of-affairs, then the reasons to perform the action that have to do with the value of the state-of-affairs in question are right-kind reasons, while the reasons that have to do with the value of bringing that state about rather than with the value of the state itself are wrong-kind reasons.” So the state of affairs that is brought about by the action is the object of the action, and is the action the state? But doesn’t the state of affairs that is brought about by the action include the action itself? I’m not clear how to parse state-given and object-given reasons when we’re not talking about intentional attitudes that have intentional objects.
Dear Doug,
Regarding your first question, I am not sure what to say. It seems to me obvious that we do have reasons to act, and not just to intend to act, but when you question this prejudice ofmine, then I find it difficult to provide a good argument. Anyway, isn’t it rather you who have the burden o proof in this case, given that the position I take is commonsensical?
If the burden of proof were to fall on me, after all, I would probably point to two things: (i) In some cases, we have reasons for an action, but intending this action would not be such a good idea, perhaps because it is of value if we perform it without intending. So we might lack reasons for intending it. (ii) There could be reasons for intending an action that aren’t reasons for the action itself. Think of cases when intending is needed for coordination with other agents. Mark has provided such examples in his paper.
Regarding your second question, we were thinking of an analysis of actions according to which an actin, A, consists in bringing about some state of affairs, p. This state of affairs cannot involve the performance of A, on pain of circularity.So when you ask “But doesn’t the state of affairs that is brought about by the action include the action itself?”, my answer would be; No, it doesn’t. Then it is another matter that bringing about p implies the state of affairs that consists in the performace of A. But the latter state is not a part of p.
Now, sometimes it might be that the good consequence of bringing about that p are not caused by p, but by the fact that p is brought about by the agent. (Suppose his bringing about that p sets a good example for other agents.) This is the case when the agent’s reasons for bringing about p are of the wrong kind: they have to do not with the value of the object of the action but with the value of the action itself.
Many thanks to Jonathan Way, for his paper, and to Wlodek Rabinowicz and Toni Ronnow-Rasmussen, for their précis.
I would like to ask for some clarification about the ambition of Way’s paper, which I may be misunderstanding.
It seems clear to me that (and how) Way is arguing for what he calls “wrong kind of reason skepticism,” that is, for the view that reasons of the wrong kind for a given attitude are actually not reasons for that attitude, at all, but are, rather, reasons to want or to bring about those attitudes. He makes the argument by considering how reasons “transmit”—how it is that reasons for one action or attitude become reasons for another action or attitude, and claiming that skeptics can, while non-skeptics cannot, account for the transmission. I find this an interesting set of issues.
It also seems to me that, in offering his argument for wrong kind of reason skepticism, Way means to be offering a solution to the wrong kind of reasons problem.
But I’m very doubtful that the claims Way makes, in arguing for his skepticism, can provide a solution—at least, given my understanding of what would count as a solution. So wonder if I am misunderstanding Way’s ambition.
A solution to the wrong kind of reasons problem, as I understand it, would offer a principled way to distinguish reasons of the right kind from reasons of the wrong kind. (It would be useful for the fitting-attitudes account of value if it did so without presupposing an understanding of when things are valuable.)
If one had a solution the the wrong kind of reasons problem, one might then go on to argue that the features (whatever they turn out to be) which show a reason to be of the wrong kind (for a given attitude) also, and thereby, disqualify it to be a reason (for that attitude) at all. That is to say, one might use the guts of one’s solution to the WKR problem to argue for WKR skepticism.
Way seems to want to argue in the other direction—he seems to want to argue that, because reasons of the wrong kind are not reasons at all, we can stop worrying about the wrong kind of reasons problem. The problem, he says, “disappears” when we stop treating reasons of the wrong kind as reasons. (p. 492)
I am not seeing how one can argue in this direction. To make the problem disappear in this way, you need to know, of all the things that are apparently reasons, of all the things that seem to count in favor of an attitude, which are the reasons of the wrong kind (and so not reasons at all)—you need some way to distinguish the wrong kind of reason (or, those that aren’t really reasons) from the right kind (or, the genuine reasons). But that is just to say, you must presuppose a solution to the wrong kind of reason problem.
This wouldn’t be a serious worry (it might only be a matter of rhetoric or presentation or interpretation) if, in arguing for WKR skepticism, Way in effect offered a solution to the WKR problem. If so, one could simply restate his account in the form I have suggested, moving in the other direction.
But I don’t see how we can glean, from the considerations Way advances, a solution to the WKR problem. His claims are about how reasons of the right kind, or reasons of the wrong kind, do and do not transmit—that is, his claims are about how a reason of the right kind gives you other reasons of the right kind, or a reason of the wrong kind gives you other reasons of the wrong kind. They are of the form, “Given that there is a reason of the right kind, this or that follows,” and “Given a reason of the wrong kind, the other follows.” But, to take that as given, we must already know how to distinguish reasons of the right and wrong kind.
(One might try to use facts about how reasons transmit to solve the wrong kind of reason problem, by describing a pattern of transmission that qualifies any apparent reason as being of the wrong (or the right) kind. One would have to say something like, “If a consideration, C, seems to provide a reason for an attitude, and if C together with the fact that a particular action would facilitate the attitude provides you reason to so to act, then C is the wrong kind of reason for the attitude.” But this principle, at least, is false. Suppose you find it difficult to intend to do something you believe you should do. Perhaps you find it difficult to intend to confront your friend. In such a case, your reasons for doing it (that it will clear the air, say), together with the fact that, by promising your spouse, you will facilitate your intention, gives you reason to promise. But that does not show that your original reasons were of the wrong kind.)
WKR skepticism is an interesting issue, in its own right, and the matters of “transmission” are interesting, as well. (I may post about them, separately. I believe I disagree with Way, though the disagreement may turn out to be merely verbal.) But I am not seeing, in a very general way, how the argument for skepticism can provide a solution to the WRK problem. But perhaps I am misunderstanding the ambition.
Dear Wlodek (and Toni):
Like Doug, I am having problems understanding how you are conceiving the possibility of acting for the wrong kind of reasons, (but this might be because I don’t fully understand your analysis of action). The case you gave in answer to Doug, the case of setting an example, the fact that p seems to be a constituent of the state of affair that is valuable (that p and p is brought about by me), and, at least arguably, a state-of-affair that is a constituent of a valuable state has at least derivative value, and the reasons we have to bring about valuable states, even when the value is derivate are of the right kind. Perhaps there can be cases in which we separate the event of bringing about that p and p itself, and it is really only p that is valuable. But these cases would have, I think, the characteristic earmarks. So, for instance, think about a modified toxin puzzle case in which the subject is strapped to a machine that has a green button that you press to deliver the toxin and a red button that stops the machine. Once you press the green button, the toxin will be delivered to you unless you press the red button in the next ten minutes. The amazingly insightful psychologist that strapped to the machine pays the usual million dollars once it is true that you brought about that you will have taken the toxin, and since he can tell whether you will press the red button or not just by looking into your mind, he makes his payment exactly a minute after you pressed the green button (unless he predicts you will press the red button in the next nine minutes, since this would make it false that you brought about that you will have taken the toxin). This would be a case of wrong kind of reasons for action, but it would have the traditional earmarks (at least the first earmark). Again, I am not sure that your analysis for action would allow for this kind of case, or if you have some reason to think that p does not have derivative value , or that I did not miss something else completely different.
I, too, am interested in Wlodek and Toni’s characterization of what wrong-kind reasons for action would be like. In my original paper on the wrong kind of reason problem, I discussed cases of actions that looked like they admitted of a right-kind/wrong-kind distinction, because they were actions that were governed by a special standard of correctness – like that governing how to tie a knot, or how to set a table for a state dinner.
But from their remarks, these aren’t the kinds of cases that Wlodek and Toni have in mind. They seem to have in mind a distinction between reasons that have to do with the value of a state of affairs brought about by an action, versus reasons that have to do with the value of the agent’s bringing about that state of affairs.
This sounds a little bit like the object-given/state-given distinction, but I don’t know why we should think that it is a case of a right/wrong distinction. On the contrary, on many ways of trying to make sense of distinctively deontological reasons, they are precisely reasons that connect to the value of the agent’s bringing something about, and not just to what the agent brings about. Yet I wouldn’t think that distinctively deontological reasons should all be classified as wrong-kind.
I would like to register my agreement with Mark here. However, I do want to add to Mark’s point that it is not just deontological reasons that get classified as WKR if I understand the proposal correctly (not that Mark suggests otherwise). Many people think they have reason to fish without finding any value in any of the independently specifiable states-of-affairs brought about by fishing (they ‘catch and release’, try to leave no traces, etc.). Similar things can be said of people going for walks, gardening, etc. (of course, deontological reasons have the advantage of not allowing for the usual hedonistic replies, but I think that your theory of reasons for actions should not depend on the availability of such replies). It seems to me that a more intuitive view would say that the object of a reason for action is the action itself. Sometimes there is reason to A because A is instrumental in bringing about a state-of-affairs S. One might even think that all (or most) reasons to A are of this kind but this would be a substantive view about reasons for actions (though, again, I am not sure I fully understand Wlodek’s and Toni’s proposal).
This is a reply to Mark’s question, which also touches upon the issue raised by Pamela. In his paper, in note 1, Mark points out that “there are two different ways in which the distinction between the ‘right’ and ‘wrong’ kind of reasons entered the literature; one way, which will not be a primary focus of interest in this article, is through the dialectic surrounding ‘fitting attitudes’ and ‘buck-passing’ accounts of value.” This is perfectly fair, but it is this latter way that is in the focus of Jonathan’s paper. Now, from the point of view of the FA-analysis, the ‘pre-analytic’ distinction between right-and wrong-kind reasons is perfectly straightforward: Right-kind reasons are those that are appealed to in the FA-analysis of value: they are the features that make the object of a proattitude, or – more generally – the object of a pro-response, valuable. Wrong-kind reasons are instead the features that make the attitude (response) itself valuable rather than its object. This explains the way we interpret this distinction in our comment to Jonathan. This pre-analytic distinction gives then rise to the WKR-problem for FA-analysis: the problem being how to explicate the contrast between right- and wrong-kind reasons *without circularity* – without assuming the very notion of value that the FA-analysts are trying to define. The object-given/state-given theory (OST), on which right-kind reasons are reasons that are the features of the object of a proattitude and wrong-kind reasons are reasons that are the features of the attitude itself, is one way of attempting to solve this problem, a way which Toni and I consider to be unsatisfactory. But the pre-analytical distinction we rely on is of course not the same as OST.
Now, Mark wonders about the way we apply the distinction between right- and wrong-kind reasons to *actions*. Well, we take it that actions can be pro-responses: Bringing about that p is a pro-response to p. So we apply the pre-analytical distinction in this case as well: The features that make the object of the action (that is, p) valuable are right-kind reasons for the action, the features that make the action itself valuable, rather than its object, are wrong-kind reasons for the action in question.
How does this relate to the case of deontological reasons for actions? As Mark quite rightly points out, “on many ways of trying to make sense of distinctively deontological reasons, they are precisely reasons that connect to the value of the agent’s bringing something about, and not just to what the agent brings about.” This, I would say, makes them reasons of the wrong kind *from the point of view of the FA-analysis”. It would be incorrect to say that p’s value partly consists in the deontological reasons to bring p about. But, on the other hand, Mark might well be right that deontological reasons can be seen as right-kind reasons if our perspective on the distinction between right- and wrong-kind reasons is not one of FA-analysis. Thus, to refer to Mark’s own account of this distinction, deontological reasons do often seem to relate to the point of actions they are reasons for.
This relates to the issue Pamela has raised. She wonders how Jonathan can assume that the WKR scepticism makes the WKR-problem disappear. Even if the sceptic is right and wrong-kind reasons really aren’t reasons, don’t we still have the WKR-problem? Don’t we still need to provide a way to distinguish between right-kind reasons (the real reasons) and wrong-kind reasons (the spurious ones)? In fact, Toni and I, in our reply to the WKR-sceptics, pushed exactly the same line as Pamela in our old paper on “The Strike of the Demon” (2004) However, I am not longer sure we were right. As John Skorupski has argued, if wrong-kind reasons aren’t reasons, then FA-analysis no longer confronts the WKR-problem: The FA-analyst can then simply say that the value of a object consists in the existence of reasons for favouring it. There is no longer any need for him to appeal to the special notion of right-kind reasons in in this account of value, It is in this sense, then, that the WKR-problem disappears *for FA-analysis*, if the WKR scepticism is correct.
But, of course, things are different if one like Mark and Pamela approaches the distinction between right- and wrong-kind reasons from a different point of view than the one of the FA-analysts’s. If the issue is how to draw distinction between two kinds of reason that applies quite generally, not just to proattitudes and pro-respnses, but also to beliefs, etc., then the WKR-problem by no means disappears just because the WKR-sceptic is right. Pamela is right on the spot on this matter. So, I think that the lesson to be learned is that there really are at least two distinct WKR-problems, one for the value theorists and the other for the reason theorists, and that these problems should not be confused with each other.
Thanks to everyone for their comments, and especially to Wlodek and Toni for their extremely helpful and challenging discussion. Sorry to have been a bit slow to get involved – I was away from the computer all day yesterday. Here are a few comments on some of the points Wlodek and Toni raise in their initial post – I’ll try to say something about further points that have been raised later (hopefully later today).
(1) Wlodek and Toni wonder whether WKR skepticism really does explain the way in which the wrong kind of reasons transmit. They have two concerns – that I don’t explain why someone might mistake reasons to want and bring about an attitude for a reason for that attitude, and that I appeal to transmission patterns for the right kind of reasons which I leave unexplained.
I agree that in both respects there is something further to be explained here. But I’m not sure to what extent this counts against my argument. With respect to the other transmission patterns, my thought is that these will be common ground between WKR skeptics and defenders, and thus that it’s fair game to appeal to them in an argument for WKR skepticism, without first providing an explanation of them. With respect to the first point, my idea was that WKR skeptic’s have a certain hypothesis – that reasons to want and bring about an attitude are mistaken for reasons for that attitude – and that this hypothesis would explain the way in which putative reasons of the wrong kind behave. It is of course a further interesting and important question why some people might make this mistake. However, it’s not clear to me that we need that explanation on board before we can argue that the WKR skeptic’s hypothesis has some explanatory value.
(2) As I say though, it’s certainly a very interesting question why the other transmission patterns hold, and at the end of their comments Wlodek and Toni suggest an explanation of transmission for reasons for action, and for the right kind of reasons to desire. Both of these explanations appeal to a kind of value transmission – roughly, that if A is valuable, and B facilitates A, then B is valuable too. However, I worry that these explanations are in tension with an FA account. If reasons are prior to value, then it seems like our explanation should go the other way round – we should explain value transmission in terms of reasons transmission. And the FA account can do that – after all, value transmission is a consequence of transmission for the right kind of reasons to desire. So I’d be a little wary of the explanation they suggest here. (Elsewhere in their comments, Wlodek and Toni note that circularity may not be a conclusive objection to an analysis. This seems right, but it nonetheless strikes me as a cost, and one that many proponents of FA accounts would want to avoid if possible).
(3) Wlodek and Toni suggest that FA accounts might accept a Value-Based Theory of the wrong kind of reasons without accepting the view I call Reductionism about the wrong kind of reasons. I’m not so sure. On the value-based theory, for something to be a wrong kind of reason for an attitude A is for it to be a respect in which A is good. And on the FA account, for something to be a respect in which A is good is for it to be a right kind of reason to favour A. It seems to follow from these claims that for something to be a wrong kind of reason for A is for it to be a right kind of reason to favour A. But that’s just the Reductionist thesis.
Now Wlodek and Toni make another suggestion which would block this inference – they suggest that we might think of the Value-Based theory as a substantive account of the wrong kind of reasons, rather than an analysis. That’s an interesting suggestion and I need to think about it some more – hopefully I’ll be able to comment further on this later.
(4) Finally, I’ll just add that my initial reaction to Wlodek and Toni’s other main comment – concerning whether the right/wrong distinction applies to actions – was the same as Mark’s. Since the distinction they draw doesn’t seem to bear any of the standard marks of the right/wrong distinction, I’m not sure why we should think of it as such. But I’ll think more about the other comments that have been made on this, and chip in if I have anything to add.
Sergio’s example with fishing is interesting and thought-provoking. If we consider the action of landing a fish, then we can see it as bringing about a state of affairs consisting in the fish being caught. If the reason for landing a fish has to do with the value of that very activity rather than with the value of the state that is being brought about, then we do want to say – just as in the case of deontological reasons for action – that it is a reason of the wrong kind. So Sergio is perfectly right that this is another example of WKR. But in his comment Sergio gives example sof actions that seem to be very difficult to analyze as cases of bringing about a state of affairs. Such examples are walking, gardening and fishing. For such case it may well eb that the activity-object format of analysis is inappropriate. If so, then we would say that the distinction between right- and wrong-kind reasons cannot be made for such actions *from the point of view of FA-analysis*. (From Mark’s angle, on the other hand, the distinction can be made, given that such activities as walking or fishing have characteristic ‘points’ or function. But this not the distinction that the FA-analysts are trying to make.)
Now, Sergio suggests that in the case of such actions it is the action itself that is its object. Many people would say something like this, but we wonder whether this does not involve a confusion between the object of an action and the object of an *intention*, or a *desire*, to act. That walking or fishing can be the object of my intention or desire is unproblematic, of course.
Thanks to Pamela for her very helpful comment. It raises lots of tricky issues and I need to think more about it. Here is an initial thought in response, which I take to be along the same lines as that which Wlodek suggests in his comment above.
I agree that if the WKR problem is the problem of sorting the right from the wrong kind of reasons, then WKR skepticism is no solution. But we don’t have to think about the problem in this way. Rather, we can just think of it as the problem that FA accounts (and other views – e.g. evidentialism about reasons for belief, the view that rationality is a matter of responding to reasons) face a family of alleged counter-examples. WKR skepticism gives us a way to handle those examples – we say that the apparent counter-examples are no counter-examples at all, since the considerations adduced in the examples are not reasons for the relevant attitudes, but only reasons to want or bring about those attitudes. And my thought is that we can defend this line by observing that the considerations adduced in the examples are guaranteed to behave in a certain way and noting that this would be explained by the hypothesis that all we have here are reasons to want and bring about attitudes – since those are things which are also guaranteed to behave in this way.
This way of putting the point may perhaps help to allay another worry suggested by Pamela’s comment – roughly, that since the transmission principles I appeal to are couched in terms of the right and wrong kind of reasons, we’ll need an account of that distinction before we can make sense of these principles. I don’t think that this is quite right. It’s agreed on all sides that there are considerations in the WKR examples which have a certain kind of normative significance. We can think of the transmission principles as saying that considerations of that sort are guaranteed to behave in a certain way. It seems like that’s an observation we can make without committing to whether or not these things are really reasons for the relevant attitudes, or only reasons to want or bring about those attitudes.
Pamela also raises the question of whether we might use the sorts of observations I make to solve the WKR problem in the form that she poses it. This is a really interesting question which I’m not sure of the answer to. I think there might be some possibility of proceeding in this way though – I might try and say more about this later.
I also have a comment about the issue of whether the right/wrong distinction applies to reasons to act. Above I expressed some skepticism about whether the distinction Wlodek and Toni draw between the two kinds of reasons to act is well-thought of as a distinction between right and wrong kind of reasons. But Wlodek and Toni’s use of this distinction to object to my argument may be independent of this question. Wlodek and Toni’s worry was that my argument over-generalises – that if it works, it implies that there are no WKRs for action. But even if reasons to act which depend on the value of so acting are not best thought of as WKRs, it would be bad news if my argument implied that there are no reasons to act of this sort.
Fortunately, I don’t think my argument would generalise in this way. That’s because these “act-based reasons” don’t seem to obey any distinctive transmission principle. It’s not the case that if there’s act-based reason to A and B-ing facilitates A-ing, then there’s act-based reason to B – it might be that there’s reason to B only because it helps you to A. What does look plausible is that if there’s act-based reason to A and B-ing facilitates A-ing, then there’s reason to B. But that is just an instance of the general transmission of reasons to act. So I don’t see that my argument raises any special problems about whether there are act-based reasons to act (as opposed to any other kind).
Thanks to Woldek, Toni, and Jonathan for their clarifications. The idea that there are two problems here might be very helpful to me (because I do not come to the problem via the fitting attitudes analysis) — but I don’t quite understand it, yet. I am going to walk through things very slowly (because that’s what helps me).
So, suppose I want to say that an object is valuable in some way, if, only if, and because a certain attitude toward it would be fitting. I’m trying to reduce unruly notions of value to something better behaved. I then encounter a problem: “fitting” is too promiscuous. Attitudes can be shown “fitting” even when the object is not valuable (thanks to the demon). Counter-example.
Woldek and Toni suggest that proponents of a fitting-attitudes account can provide, pre-analytically, an account of the difference between the right and the wrong kind of reason for valuing attitudes: The right kind are those features of an object that make it valuable. The wrong kind are those that do not.
We can now constrain “fitting.” We can say that a valuing attitude is fitting when there are reasons of the right kind for it, and otherwise not. We avoid counter-example.
But we do so at the cost of failing in the original ambition—at the cost of taming the unruly notion of “value.” Because now, when I say that an object is valuable in some way if, only if, and because a certain attitude would be fitting, I have to add that such an attitude is fitting only if there are reasons of the right kind for it, where the reasons of the right kind are those provided by the valuable feature of the object. The analysis has become circular.
The skeptic wants to come to the rescue. He says, “Reasons of the wrong kind are not reasons at all!”
How does this help?
Maybe like this: I say, an object is valuable in some way if, only if, and because a certain attitude toward it is fitting. The demon points out that “fitting” is too promiscuous. I say, no, an attitude is fitting only if there are reasons for it. The demon now tries to point out that there can be reasons that are not given by valuable features of the object. The skeptic says, no, those are not reasons at all!
It seems there are two issues. The less pressing is, why are those not reasons at all? (Jonathan says, because, by taking the position that they are not reasons, we can explain facts about transmission.)
The more pressing is, unless I have an independent way to distinguish the only apparent reasons from the real reasons, how am I avoiding circularity?
I want to say that an object is valuable if, only if, and because certain attitudes towards it are fitting, and that an attitude is fitting only if there are reasons for it. The demon points out that there seem to be reasons that are not given by valuable features of the object. The skeptic says, no, there are not. Those are not really reasons. Now, without an independent way to identify the real reasons, aren’t we just saying that the only real reasons are those given by the valuable features of the object? It seems I am now saying that an object is valuable if, only if, and because there are reasons to value it, where that means, if, only if, and because there are features of the object that show it valuable. Aren’t we back in circularity?
If I had an independent way to say which of the apparent reasons aren’t really reasons, then I would avoid circularity.
So, there’s something I am not understanding.
Jonathan seems to suggest (though I may be misunderstanding this, as well) that he has provided a way for the FA analyst to confront the counter-examples one by one. The demon says, but here’s a case in which you have a reason to have the attitude, even though the object is valuable! And Jonathan says, I can presume that’s a wrong kind of reason, feed it into my transmission principles, and conclude it is no reason at all. (Is that right?)
But, why can I presume it’s a wrong kind of reason? If I do that, aren’t I again helping myself to just the pre-theoretical notion of value that I am trying to tame?
Yes, jonathan. I think I agree with you. The explanation for this difference in patterns of reasons transmission between attitudes and actions seems to be the fact that actions *cause* their objects while attitudes don’t. Which means that, if B-ing is bringing about that p, B-ing can facilitate A-ing in two ways: either (i) it is p that facilitates A-ing (and B-ing causes p) or (ii)it is bringing about that p, but not p itself, that facilitates A-ing. In both cases there is a reason for B-ing if there is a wrong-kind reason for A-ing, but it is a right-kind reason for B-ing in case (i) and a wrong-kind reason for B-ing(an act-based reason, as you call it)in case(ii).
Thanks, Pamela. That’s a very clear and helpful way of setting out the worry, and I think there probably is a serious challenge to my view here. I’ll try to say a bit more about how I’ve been thinking of the WKR skeptic’s move (which I think is pretty close to how you describe it) although I doubt that what I say will fully answer the concern.
So the FA theorist begins with the idea that for something to be good is for there to be reasons to favour it. Then the evil demon case comes up. There are two possible responses. The first is to concede the counter-example and revise the theory – we now say that for something to be good is for there to be the right kind of reasons to favour it. The second is to reject the counter-example – we deny that the demon’s threat is a reason to favour him, and we stick to the initial theory.
Now the first of these responses clearly commits the FA theorist to a further task – distinguishing the right and wrong kinds of reasons. (This is the WKR problem, as you describe it). But the second response (the WKR skeptic’s) doesn’t commit the FA theorist to this task. That looks like an advantage of this second reply.
However, you might think that there’s no real advantage here. After all, FA theorists will still need to distinguish reasons to favour from things which aren’t reasons to favour (without relying on notions of value). But while this is true, it is also something that FA views will need to do anyway – regardless of their stance on WKR skepticism. So my thought was that there is no further commitment for the FA theorist here.
(How do we distinguish reasons for pro-attitudes? Well, I take it we engage in first-order normative theorizing – we try and figure out what sorts of things are good, or (as the FA theorist will understand it) what sorts of things there are reasons to favour).
Still, you might well think that this doesn’t really get rid of the problem. For how do we know that the correct account of reasons to favour will tell us that incentives for attitudes, of the sort we see in the WKR cases, aren’t reasons for those attitudes?
It thus looks like WKR skeptics will still need two things. First, they’ll need some way of picking out the candidates to be WKRs – the kinds of considerations we see in the WKR examples. Second, they’ll need a way of arguing that those sorts of considerations are not reasons for the relevant attitudes, but only reasons to want and bring about those attitudes. Now in my paper I attempt to do the second of these things. But I think your point is that this still leaves the first task. And the worry here is that picking out these candidate WKRs is either going to be a matter of providing a full account of the right/wrong distinction – in which case we’ve not saved ourselves any work – or is going to rely on the pre-analytic sense of this distinction, with its explicit appeal to value.
Perhaps though there is another way to pre-analytically pick out the candidate WKRs – we appeal to the earmarks. Thus we might say that candidate WKRs are considerations which clearly have some kind of significance with respect to an attitude, but which don’t bear on the rationality or correctness of that attitude, and which it is difficult or impossible to respond to in the way in which we respond to paradigmatic reasons for that attitude. This way of picking out these considerations doesn’t so obviously appeal to value, and it also leaves open whether these considerations are reasons for the relevant attitudes. It thus allows for the FA theorist to argue that these sorts of considerations aren’t reasons of this sort.
Of course, since this way of setting things up relies on the earmarks, it is subject to the worries you raise in your very interesting reply to Mark’s paper. So I’ll certainly have to think more about whether this approach is legitimate.
Thanks to all for these rich and interesting papers and discussions! I hope it is not too late to throw something in–these discussions on your Interweb are speedy indeed, and Dan Jacobson and I have been processing them more slowly.
We agree with many of those here that it is important to distinguish some different problems lurking in the WKR neighborhood. One problem concerns the prospects for FA theories of value, posed by apparent counterexamples. Another problem is, as Pamela puts it, to “offer a principled way to distinguish reasons of the right kind from reasons of the wrong kind.” Pamela and Mark are more confident than we are that there is a common notion of right-kindedness of reasons that applies across discrete “activities”: believing, intending, having various evaluative attitudes, and (for Mark) even tying knots. If there is, then a principled way to distinguish reasons of the right kind across all these domains will be very general–as in fact Pamela’s method in her original WKR paper is. Mark’s original method is too, in a sense, though it also yields more specific accounts for each discrete type of activity. And one might hope to extract a method from Jonathan’s observations about transmission, though as Pamela points out this will not be easy to do.
But any such general method of distinguishing right from wrong kinds of reason will leave another problem that is also important and has not been clearly enough distinguished to be fully appreciated. This is what we call the Opacity of Normative Force, and it is the topic of the paper we will be giving at the Metaethics Workshop in Madison this year: Sometimes is actually quite difficult to determine, of various specific considerations, what kind of reasons they are, even when it seems clear that they count in favor of or against having an evaluative attitude. Even if one had a general account of what right kinds of reasons are like, a “principled distinction” between R and WKRs, it would not be a device for deciding in these hard cases what sort of a reason we were looking at.
Consider the fact that you are not responsible for some trait you have, as a reason not to be ashamed of it. That seems like some sort of reason not to be ashamed, but what sort? Does your lack of responsibility for a withered arm show that it is not shameful, or is it a WKR—a reason not to be ashamed that does not speak to the question of shamefulness? This is controversial, actually. Someone who advances such a consideration may be confident of its importance while uncertain, even confused, about just what kind of normative force to assign it. If so, then its normative force is opaque to him: he is confident that it is a reason, but unsure what kind of reason it is.Moreover someone may be able, by reminding himself of this fact, to reduce his propensity to shame at his arm–on good days, maybe to eliminate it altogether. He may think that he is responding to an RKR. But he may be mistaken about this, perhaps even by his own lights. You might think that if he can indeed respond to it then it must be an RKR at least by his lights. Not so, we’ll argue.
This is just one example of a general problem: once one looks beyond exogenous incentives, which are obvious WKRs, there are lots of reasons whose relevance to the question of what is shameful, admirable, pitiable, prideworthy and so on is opaque. These lead to confused evaluative judgments and specious disagreements, to which even good philosophers are not immune.
This is a different problem, we think, because it seems not amenable to solution simply by finding some general account of the difference between R and WKRs. At any rate, none of the general criteria of right-kindedness of which we are aware seem to help at all with determining the normative force in the hard cases. Once we decide that a consideration is a WKR we can then decide that it is best understood as state-given, or as only a reason to do something else, like try to have the attitude, or as not directly followable, or as not bearing on the question of whether the trait is shameful, or as not shared necessarily by anyone engaging in the activity of being ashamed, or as transmitting like a WKR, etc. But no such general fact about WKRs can be deployed as a useable test of normative force in the real-life cases in which that is in doubt.
That’s not an objection to any of the above suggestions as general accounts of right-kindedness, of course. It’s an attempt to draw attention to another important problem in the WKR vicinity, and a shameless advertisement for the paper in which we try to show this problem is real and endemic in some actual philosophical discussions.