I would like to start by thanking Phil Clark, Fred Schueler, and Christine Tappolet for writing such thoughtful comments on my book. I couldn’t possibly address all the issues raised by these outstanding critics, so I’ll be somewhat selective.
It may be useful to provide a quick general overview of the central ideas of the book. The view that I defend in the book regarding practical reason and intentional explanations, the view I call “the scholastic view,” is grounded on the assumption that the good is the formal end of practical reason, and that various practical attitudes such as desires and intentions should be understood in terms of this formal end. I discuss what I mean by a “formal end” in the book, and I’ll say a couple of things about it below, but for now I just want to say that I understand this claim to imply that “good” plays the same role in practical reason that “true” plays in theoretical reason. In fact, my view is that theoretical and practical reasons are to be distinguished solely in terms of their different formal ends. I mention this at the outset because I think a number of the criticisms raised can be answered if one keeps in mind the suggestion that the scholastic view argues for the existence of such a “structural parallel” between theoretical and practical reason. So I’ll often be relying on an understanding of how theoretical reason operates when answering the objections, and assuming that practical reason operates in a similar fashion. With the exception of some of Schueler’s comments, this assumption is not much challenged, so I’ll leave it standing unargued until the very end of my comments.
Let me start by briefly responding to a concern that Tappolet rises against my treatment of animal desires, a concern that is due entirely to my own careless wording. Tappolet quotes me as saying that animals do not represent the objects of their desires as good. And she correctly points out that this seems to flatly contradict the general scholastic idea that to desire is to conceive something to be good. But the seeming contradiction is the result of an ambiguity in the expression “represent as good.” I do not think that the concept “good,” or any non-conceptual representation of “good,” is part of the content of the animal representation. However, I think it might be that certain animal representations are such that the content of these representations is taken by the animal to be good. Here I’ll only try to explain why this is a coherent possibility without trying to fill out any of the details that the scholastic account of animal desire should provide (I do a bit more in the book).
Many philosophers think that animals have beliefs, and some large subset of these philosophers think that belief has some kind of constitutive relation to the truth, such that believing that p implies holding p to be true, in some sense of “holding.” However, I imagine that very few, if any, of these philosophers would be inclined to say that the brutes have “true” as part of the content of their representation. Yet the combination of these views does not seem to be incoherent. And here is an obvious reason to think not only that the combination of these views is coherent, but also that it is plausible to think that certain beings could have beliefs without having the concept “true” or any related concepts. We could say, roughly, that belief is an attitude such that when one believes that p, one is in a certain relation to the truth of p (say, for instance, the relation of holding p to be true). If anything like this is correct about belief, it would be superfluous to demand that the content of belief would include a representation of “true” so as to guarantee that this relation obtains. I claim that a similar relation holds between desire and good, so that in desiring one conceives the content of the attitude to be good. So it is equally superfluous now to demand that “good” will appear as the content of every desire; having the desire with a certain content guarantees that one is conceiving its content as good. So the scholastic view is not committed to the claim that the brutes must be capable of having any concept of the good in order to have desires. This is all that I meant to say in the passages Tappolet quotes.
Now one might object that without attributing the representation of good to animals, there could be no reason to attribute these desires to animals. After all, on what grounds would we classify any putative mental state of the animal as a case of desire? I do not want to commit to any particular answer to this question here, but one can see there are various candidates for the answer. One can think again about what tempts us to attribute beliefs to animals without attributing the concept of truth as part of the content of their representations. One could agree, for instance, with some dispositionalists and think that if an animal’s behaviour can be best explained as “behaving as if it is the case that p,” then this animal believes that p. In this case, we would similarly say that if an animal behaves as if it would be good that p, then the animal desires that p.
This might also help answering another concern that Tappolet raises. Tappolet asks how the scholastic view distinguishes between desires and judgments of the good, and she’s right to say that it cannot be in terms of the content; after all, the very same thing that I desire I can end up judging to be good. But I don’t find her proposal to distinguish them in terms of the kinds of content they have (conceptual content in the case of judgments and nonconceptual content in the case of desires) plausible. Tappolet claims that the concept of the good is not (typically) part of the content of what is represented by a desire. As I just said, I agree with this point. However, for similar reasons, I do not think that genuinely evaluative judgments, insofar as they are practical judgments, have the concept of good as part of their content either. Moreover, whatever one thinks about the possibility of nonconceptual content, I don’t think one can characterize all desire as having nonconceptual content. I can’t see how my desire to help my graduate students do better than other equally skilled graduate students in the philosophy job market, or my desire to help make sure that my enemy’s book gets remaindered soon, could be an attitude towards a nonconceptual content.Footnote 1
The scholastic view takes the distinction between appearances and evaluative judgments in practical reason to correspond to a similar distinction in theoretical reason.Footnote 2 That is, appearances and judgments are, respectively, prima-facie attitudes and all-out attitudes. Evaluative appearances are attitudes that essentially involve inclinations to take a certain stance in the practical realm, and evaluative judgments essentially involve taking a certain stance in the practical realm (more on this momentarily). This is why we need both attitudes; the fact that something appears to the agent good in a certain way does not entail that the agent judges it to be good, and one could hardly judge something to be good if some things did not appear to be good in some way. I hope this also helps in answering Schueler’s complaint that the notion of appearance is dark and cannot be properly distinguished from a notion of believing that something is good. I take it that Fred finds it dark in part because he takes “appearance” to be a metaphor from the visual realm. However, I use the notion of appearance to capture instances of mental states that are often picked out by phrases such as “It appears to X that . . ..” Visual appearances here are merely one species of a larger genus; “perceptual appearances” or “perception” cover only one possible class of cases in which things may appear to someone to be true in the theoretical realm.
Now it’s pretty straightforward how to draw the distinction between appearances and beliefs about the good in my view. As Clark points out in his paper, strictly speaking, believing that something is good is on my view not even a practical attitude, but rather a theoretical stance regarding practical matters. But ignoring this complication, the scholastic view distinguishes between beliefs about what is good (or, in my view, intentions or genuinely evaluative judgments) and appearances about the good in terms of what we have been calling “prima-facie” and “all-out” attitudes, in terms of being (cognitively) inclined to form a certain judgment and making the judgment itself. It can also allow that, as Davidson says, “A man may his whole life have a yen to drink a can of paint without ever believing [or judging] it would be worth doing.” This case is in fact not unlike Gary Watson’s case of having the urge to smash the racket on an opponent’s head. These are cases that in the book I classify as illusory appearances, appearances that one knows to be illusory even if one is still tempted by them (as opposed to cases in which a certain appearance is overridden; when, for instance, it is no illusion that, at least in certain circumstances, it would be good to eat chocolate but right now what’s good — or at least better — is to give the chocolate to my friend). In the book, I argue that these are the practical counterparts to theoretical appearances that range from the Müller-Lyon illusion to the way in which the gambler’s fallacy appears to be valid (as Clark points out, “jinxing” would also be illusory in this way).
I must say, however, that the yen to drink a can of paint is a bit more complicated, and I am not sure how to classify at least the ordinary case of having a yen to drink paint (something that I believe that many of us have experienced). My first reaction is to think of it as the kind of thing that could never generate action; at least, it’s hard to imagine that the ordinary yen would lead someone to say to herself, “Oh, what the hell!” and actually drink the paint. If this is the case, then the scholastic view will not classify the yen to drink a can of paint as a desire, since it is not a potential item in an intentional explanation and its content has no role to play in our practical reasoning. One might think that this does violence to our ordinary usage of “desire,” but, I must say that I am not sure that this is true, and I certainly do not think that it would matter; I am happy to concede that my view implies that our ordinary usage of the term “desire” is ambiguous. Another possibility is that the yen to drink a can of paint is like an obsessive compulsion, something that could interfere in the process of practical reasoning without thereby being conceived to be good. In this case, again, the yen would not be considered a desire, and the treatment I would give would be similar to the treatment I give of obsessive-compulsive disorders in the book. In a nutshell, I argue there that although the end of activity generated by such desires is not one judged to be good, the instrumental structure of the pursuit of the end allows us to say that the agent pursues this end as if it were good, and provides us with intentional explanations of the activities in pursuit of such a goal. Finally, it is possible that the ordinary yen is such that one actually conceives drinking the paint to be good but not good enough to override other considerations. This would be a case in which one weighs the benefits of having the experience of the smooth texture in her mouth against its unpleasant taste and the nearly certain death that the paint would cause. I don’t find this a particularly plausible description of the ordinary yen, but I mostly want to argue that however one is tempted to think of our yen to drink a can of paint, I don’t think it should cause any problems for my view.
Before I move to what I take to be Schueler’s central criticism of my view, I would like to say a few words on my notion of “perspective.” No doubt the notion is introduced as a technical notion; all that I want to take from the perceptual notion of “perspective” is the relations of compatibility and incompatibility implied by this notion. The notion is supposed to cluster together desires by relations of coherence that would not obtain across separate perspectives. So someone can both desire and be averse to eating chocolate, but not from the same perspective. One generally desires chocolate from a gastronomic perspective, but is averse to eating it for different reasons. Moreover, if one desires chocolate from a gastronomic perspective, one should not fail to desire to eat chocolate in virtue of properties or relations that are irrelevant when considering this perspective. There is something odd, to say the least, in someone whose otherwise ordinary desire for chocolate excludes chocolate that it is exactly fifty miles away from Little Rock, Arkansas. Moreover, as Tappolet points out, the notion of an evaluative perspective rules out the possibility of a being with just one desire, and she finds this somewhat troublesome. But I do find it hard to wrap my mind around this idea, and I think Tappolet herself points out the reasons why it is difficult to conceive of such a being. I don’t know, for instance, what to make of a being who wants to be in Tegucigalpa at some point, but has no preferences about how and when to get there, no desires to be at a particular place in Tegucigalpa, no desires to do anything once there, or to be anywhere else in the world, etc.
At any rate, this is how I would put what I take to be Schueler’s central criticism. Suppose we agree both that the good is the formal end of practical reason and that intentional explanations aim to mirror the agent’s practical reasoning. We can even accept a principle of charity and embrace the view that we need to understand the agent as succeeding in her pursuit of the formal end of practical reason as much as possible. But of course accepting all that still will not, and certainly should not, grant us the conclusion that we are all perfectly rational. So we need to make room for irrationality of some kind or another. However, the scholastic view rules out one particular form of irrationality; namely, pursuing that which we do not judge to be good at all. But why should we have such scruples about this kind of irrationality but not about others? This is an excellent question and I am not sure I have a conclusive argument in favour of my view. And I must confess that the question is all the more pressing in my case, since I have misgivings about the Principle of Charity, especially the strong version of the Principle of Charity that Davidson puts forward.
I’ll begin again by looking at the analogous issue in theoretical reason and try to see how we can apply the lessons from the theoretical to the practical case. So we should first note that we would like to say the same thing about beliefs that Schueler has said about actions. In ascribing beliefs to agents, we’re often guided by some version of the Principle of Charity (perhaps a very weak one). However, this should not preclude us from ascribing irrationality to agents and to various irrationally formed beliefs; some believe that there is no evidence for evolutionary theory, some accept as valid patterns of inferences that are guided by principles of what jinxes what, and some will gladly affirm the consequent. One could then go on to ask, why should we not accept that some people believe irrationally by simply believing what they judge to be false; why should we have any scruples about this kind of irrationality? Yet most of us, if not all, do have such scruples. I think we have such scruples because the notion of belief picks out our all-out attitude in the realm of theoretical reasoning. Now if one thinks that the formal end of theoretical reasoning is the true, then one holds that insofar as one is engaged in theoretical reasoning one is aiming at the truth, whatever else one might be aiming at. But from these two points it would follow that belief is our all-out attitude in the pursuit of truth, our final stance about where the truth lies. A similar point goes then for intentional action. Since intentional action is the outcome of deliberation, that is, of practical reasoning, and the good is the formal end of practical reason, the intention with which one acts will similarly be one’s final stance with relation to the good. Of course one needs to do more work to show that now desires, as prima facie attitudes, should also “aim” at the good in the same way, but I’ll just here briefly mention that in my view we get this consequence from the fact that these attitudes count as desires insofar as they incline us to adopt, and provide us with putative grounds of, all-out attitudes.
Now one might deny that intentional action is best seen as the outcome of practical reasoning, or that the good is the formal end of practical reason. But I took Schueler’s objection not to be denying these claims but rather to be challenging their capacity to support something as strong as the scholastic view I defend in the book. But I hope this helps showing that the implication does hold at least for the case of intention. However, Schueler does take issue also with this understanding of the good as the formal end of practical reason. I’ll briefly address this issue momentarily, but we might accept the argument I just provided and think of it as a reductio. It is often said that people pursue what is bad, and that ordinary experience teaches us that people sometimes pursue something just because it is bad. I’m not sure that Schueler had this objection in mind, but it is suggested by some discussions of this topic. Velleman, for instance, claims that those who subscribe to strong versions of the guise of the good view will never do full justice to perverse desire; when describing Satan, for instance, they’ll come up only with a Satan who tries to be a do-gooder, but who is simply mistaken about what the good is; what Velleman calls a rather “sappy Satan.” But I must say these intuitions seem to me to be predicated on slipping from a formal notion to a more substantive moral notion of good and evil. It seems false to say that Satan just desires the bad as such; it seems hard to believe that Satan would want to consume Spam sandwiches because they are so bad or even watch Plan 9 from Outer Space without the relevant campy attitude. Of course, there is some way of saying “Satan is just trying to be good” that will evoke the idea of a sappy Satan. But I think the idea is being evoked by thinking that such expressions imply a certain view of Satan as someone who is trying to help old ladies cross the street, but accidentally keeps sending people to hell instead.
In any event, let us examine the following descriptions of a more familiar sort of Satan:
(a) Hitler thought that killing Jews was morally good.
(b) Hitler thought that killing Jews was immoral, but he also thought that pursuing what morality commands was an activity suitable for vermin; he thought that the true good for an Aryan was to annihilate the inferior races.
I don’t find either description, and certainly not the second, to be describing a sappy Hitler. Moreover, I think that although ordinarily we slip from the formal notion to the moral notion, I think the formal notion also has a firm footing in ordinary language (“It’s not so good to be thinking about others so much;” “What’s so good about spending your life helping others and doing all these nice things?;” “There’s nothing better than listening to the cries of your enemy;” “It was so good to hear that my enemy’s book had been remaindered”).
But now I think we can reject Clark’s suggestion of how the scholastic view should understand Satan. Although Clark is right that the scholastic view is not committed to thinking that Satan believes that corrupting mortal souls is good, I don’t think that any plausible understanding of Satan’s motivation needs to (nor should) deny that he has this belief. I find it hard to think that this would be the correct description of the sort of evil character a devil is supposed to be. For in my view, beliefs about the good are theoretical stances on the question of what the outcome of correct practical reasoning would be in the specified circumstances. And I don’t think that Satan would see his saintly peers as better reasoners than he is; the classic character would more likely conceive of them as “fools” or otherwise inferior in their practical thinking. But I also think that, just as in the case of Hitler, we should not think that we’re missing anything about Satan’s evil nature when describing him as aiming at the good.
Schueler thinks that the analogy between theoretical and practical reason breaks down when one thinks in terms of the relations of consistency that hold in the theoretical realm but not in the practical realm. He is right to point out that I can desire incompatible things, but this I think is fully compatible with the theoretical realm in which things can appear in incompatible ways. It is at the level of all-out attitudes that the relations of consistency obtain both in the theoretical and practical realm. And here too, at least at first, the parallel seems perfectly fine between theoretical and practical reason since incompatible intentions are also irrational. I do think, however, that Schueler is right that there are a number of cases in which there seem to be a lack of parallel in the way that there could be many goods and only one truth, but I think (or at least hope) that these are like the cases that I discuss in chapter 4 of the book, when I consider why there could be personal goods but not personal worlds, or many goods but not many worlds. These all boil down, I think, to the question of why there are genuinely permissive inferences in practical reason but not in theoretical reason. So think about cases in theoretical reason in which the evidence does not warrant a very high confidence on either p or not p. Arguably, in some such case it might be permissible for me to believe either p or not p. However, whatever one’s view is about this permissibility, it is clearly not true that in any such cases (with the possible exception of some bizarre self-referential cases), if I form the belief that p on the basis of such permissions, then p is true. But at least in some cases in practical reason this is correct mutatis mutandis, or so I argue in Appearances of the Good. That is, in some cases, if I am permitted to infer that it is good to do X, then my inferring that it is good to do X makes doing X good. However, the scholastic view does not rule out such failures of parallel. What the scholastic view is committed to is that these absent parallels be explained in terms of the nature of the different formal ends of theoretical and practical reason. A hopefully promising way to look for this explanation goes roughly as follows: The idea of the world as it is, or the true description of the world, is of a shared world in which we’re only its observer, not its creator. On the other hand, the idea of a life as it ought to be, or a good life, is the idea of an individual life that one creates (or, more idiomatically, that one leads). So there’s nothing in the nature of this idea that will rule out more than one appropriate way of leading a life, but this possibility is ruled out by the idea that we’re all observers of the same world.
I want now to move on to Tappolet’s criticisms of the scholastic treatment of akrasia and accidie. I would like first to clarify one aspect of my view regarding akrasia. Tappolet describes the scholastic view as denying the possibility of “strict akrasia.” I have nothing against this characterization if the expression “strict akrasia” refers to a view that can only be described in technical philosophical vocabulary and does not imply the denial of the existence of any ordinary phenomena, or even the appropriateness of any ordinary description of such phenomena. After all, the scholastic view does not deny that we can act against what we believe to be all-things-considered best. It denies that we can act against our all-out or sans phrase judgment of what is best. Since the distinction between all-things-considered and all-out judgments does not appear anywhere in common parlance, I can’t see how pre-theoretical judgments could force us to accept an interpretation of the phenomenon as one in which one acts against one’s all-out, rather than against one’s all-things-considered, judgment. But, of course, Tappolet is not relying just on pre-theoretic judgments when arguing that the scholastic view makes clear-eyed akrasia impossible. She argues that insofar as my account sees the agent as persuaded by certain direct cognitions that conflict with her oblique reflective judgment, we can’t make sense of the possibility that the agent knows full well that A is better than B when choosing B over A. First, let me immediately grant that I do think that there must be a cognitive difference between the incontinent agent and what Aristotle would consider to be a virtuous agent; that is, an agent who is not even tempted to pursue the lesser option. So it is true that on my view the incontinent agent is necessarily in a less-than-ideal cognitive state. But it’s also important to note that the virtuous agent and the incontinent agent make the exact same all-things-considered judgment. The difference lies in their understanding of the grounds of the all-things-considered judgment rather than in its actual content. So I am not sure why the scholastic view needs to reject that, in the relevant sense, the akratic agent may know full well that A is better than B when choosing B over A. I know full well that general relativity is better confirmed than classical mechanics, that Fermat’s last theorem is true, and that temperature is mean molecular kinetic energy. Although I can be said to know all these things, doubtless my grasp of why any of these is true is considerably inferior to the understanding held by those working in the relevant areas.
Tappolet also objects to my treatment of accidie, and this is in part due to the somewhat ill-advised way in which I presented my views on value. Tappolet correctly points out that “it does not seem that the agent suffering from accidie would consider her options as good if she did not consider something else as good, or if she did not believe that it was not in her power to bring about the options she considers,” and she quotes the definition in the book which on its own would imply that the scholastic view is committed to interpreting accidie in this manner. However, I do later (p. 61) say that this definition is just an approximation, and I explicitly point out both that conditions (a) and (b) are not supposed to be exhaustive and also that the chapter on accidie would introduce at least one more way in which something that is not judged to be good can be judged to be valuable. The basic idea is that the valuable is that which in certain privileged counterfactual conditions would be judged to be good. Conditions (a) and (b) only spell out a minimal set of relatively uncontroversial privileged conditions. That is, for instance, it seems relatively uncontroversial that I find valuable those things that I would bring about if it were in my power to bring them about. The chapter on accidie introduces the relation of conditionality to the scholastic framework; that is, according to the scholastic view I defend, the following relation between a state-of-affairs C and an evaluative perspective E sometimes obtains or is judged to obtain:
Strong Conditionality: C strongly conditions an evaluative perspective E for an agent A if and only if, for every O conceived to be good from E, A should judge O to be good only if C obtains.
The claim, which admittedly I should have made clearer, is that at least if C is a desirable condition, O will be judged to be valuable when C does not obtain (i.e., the absence of C is a privileged counterfactual condition in cases of strong conditionality in which C is itself a valuable state-of-affairs).
I find Clark’s description of my view so compelling and clear that it pains me to say that there’s one part of his description that I think should be revised once the full explanation of the scholastic view is on the table (I’m not sure Clark would disagree with this). He introduces a problem in characterizing practical reasoning that comes from what Clark nicely describes as practical reasoning having two masters to serve: knowledge and action. And this might seem like an important disanalogy between practical and theoretical reason if we start from the point of view that knowledge must be aimed at the true. But I think once we see that true is not the aim of practical knowledge, theoretical and practical reasoning stand on the same footing. One could similarly say that theoretical reasoning has two masters to serve: knowledge and belief.
I must confess that I always thought of my disagreement with Davidson as narrower than Clark describes. But at any rate, much of what Davidson has to say in this area I want to preserve, including the distinction between an all-out and an all-things-considered judgment. Clark is right, however, that I do not think that our all-out judgment simply happens to always issue in action, but I also would be reluctant to attribute this view to Davidson. More plausibly, Davidson thinks that the mark of intentional action is that, at least under a certain description, it is the expression of an evaluative judgment. So far Davidson and I are in agreement. Davidson might have a different conception of what an evaluative judgment is, but I find this a minor disagreement. The important thing is that as an evaluative judgment, it is the kind of thing that expresses the agent’s final stance on evaluative matters. This is what I think the kind of functionalism Clark describes would not be capturing, and this is also the reason that, in my view, it would not provide an adequate picture of intentional explanations.Footnote 3 In particular, this kind of functionalism would allow explanations of compulsive radio turners to be just as legitimate as (in fact, identical with, given a few twist and turns) explanations of the actions of radio aficionados.
The functionalist as described by Clark still needs to introduce evaluative judgments somehow into her picture of the mental world. But this kind of functionalist will have to think of those judgments as an independent kind of realm to which our minds might turn. In my view, it is indeed possible that one’s theoretical attitudes about what good practical reasoning is and the practical attitudes one forms in reasoning practically might part company. However, according to the scholastic view, when this happens the agent is conflicted. On the other hand, if we simply superadded beliefs about values to this functionalist view, there would be no warrant to say there could be conflict between these the practical and theoretical attitudes of the agent. For this version of functionalism, these would be simply two independent attitudes. So while the scholastic view regards valuing, desiring, intending, as all different aspects of a single unified rational activity, no such door is open to this kind of functionalist.
