2. Scanlon and Smith

So what, on Myers’ interpretation of Davidson, is this alternative account of pro-attitudes that allows Davidson's philosophical approach to accomplish all of these important things? His suggestion is that it is “like the one Thomas Scanlon proposes, according to which pro-attitudes are … states in important respects sensitive to one's judgments about one's reasons for action” (DDTA, p. 145). Here, Myers begins his journey down an interpretive path that finds little support in the text, and that eschews the more Anscombean path that finds robust support in Davidson's writings. The contours of these diverging paths will soon become clear, but the first piece of evidence that this cannot be the right one has already been provided. What Myers appeals to in Scanlon's account, that such pro-attitudes, properly understood, are not mere dispositions to promote some state, but in some sense involve evaluations of their contents as desirable, valuable, or good, clearly does capture an essential feature of Davidson's alternative account of pro-attitudes. But it is equally clear that Scanlon's account can't be Davidson's, because Scanlon's approach violates what Myers has established are the fundamental parameters for any plausible interpretation of Davidson's account of pro-attitudes. In particular, Scanlon clearly does reject the Humean theory of motivation (Myers’ comments notwithstanding, DDTA, p. 145), which Myers himself is clear that Davidson accepts. Indeed, Scanlon sums up his argument concerning desires as establishing that “we should not take ‘desires’ to be a special source of motivation, independent of our seeing things as reasons.”Footnote ⁶ Seeing things as reasons, in turn, is for Scanlon a question of belief, and such beliefs are sufficient to motivate action: “if a rational agent believes that p is a conclusive reason to do a, she generally will do a … for this reason.”Footnote ⁷ In short, Scanlon clearly maintains that beliefs alone are sufficient to motivate actions, hence he clearly rejects both the Humean theory of motivation and the Humean theory of pro-attitudes. Indeed, he takes appeals to desire in accounts of motivation often to function as mere placeholders for the beliefs (judgements about reasons) that in such cases motivate agents to act.

So, although Scanlon's account shares something important with Davidson's, a special relationship between attitudes, whether beliefs or desires, and judgements of the truth and goodness/desirability of their objects, it also rejects something central to Davidson's account: the Humean theory of motivation. Davidson thus cannot be defending a position like Scanlon's because, as Myers rightly points out, Davidson consistently endorses the Humean theory of motivation that Scanlon clearly rejects. A compelling interpretation of Davidson on pro-attitudes should aspire to incorporate what is attractive in each account of desires on his view, their special relationship with evaluative judgements, and their central role in motivation, while avoiding what he takes to be mistaken in each. An account of desires that Scanlon develops in large part to ground his rejection of the Humean theory of motivation must differ in fundamental respects from Davidson's own account, and cannot be a model for it.

This suggests that Davidson's actual path diverges from that advocated by Scanlon (and Smith). The basic contours of these two different interpretive forks are in fact marked out clearly by Myers, and come clearly into view with his Smith-inspired attempt to account for Davidson's commitment to internalism. Interpreting Davidson through a Scanlonian lens invites a picture that is presumptively hostile to Davidson's internalism. We have normative beliefs, e.g., about the reasons that we have for trying to Ψ (DDTA, p. 145), and the pro-attitudes that we typically have are properly understood merely as dispositions to act that depend upon, or are conditioned by, these normative beliefs (DDTA, p. 145). The puzzle is that it is not clear what the internalist story can be that accounts for the dependence of pro-attitudes, thus understood as complex dispositions to act, upon evaluative judgements understood as beliefs.

Myers suggests that there are two possible strategies to undertake in order to vindicate Davidson's internalism while maintaining his commitment to the Humean theory of motivation. These two strategies nicely capture the two different interpretive paths. The first, the individual path, takes individual pro-attitudes to be constituted in part by evaluative commitments to their objects, and accounts for internalism through appeal to these “constitutive aims possessed by individual pro-attitudes” (DDTA, p. 153). Desires are necessary to motivate action, as the Davidsonian theory of motivation requires, but desires are not mere dispositions to act, hence the straightforward tension with a commitment to objectivity does not arise. The second, the systemic path, denies that individual pro-attitudes are constituted in part by evaluative commitments to their objects, and accounts instead for internalism through appeal to the aim of “pro-attitudes considered rather as a system or a whole” (DDTA, p. 153). Myers attributes to Davidson the latter, systemic fork, and does so largely because he takes the former, the Anscombean/Aristotelean fork, to collapse back into the Humean theory of pro-attitudes.

His template for this interpretation of Davidson as undertaking the second, systemic path is Smith, whose claim is “best understood as a claim about the systemic aims possessed by pro-attitudes as a group, not as a claim about the constitutive aims possessed by individual pro-attitudes on their own” (DDTA, p. 153). Pro-attitudes, Myers suggests, are for Davidson more complex dispositions than Humeans recognize, and the internalist claims that Davidson makes are, like Smith's, “systemic in character” (DDTA, p. 154): desires possess, as a group, the “systemic aim to get normative matters right” (DDTA, p. 153). Smith may well take up this systemic path, upon which it is only as a group that pro-attitudes aim at the good. But there is a great deal of textual evidence that Davidson endorses the individual path,Footnote ⁸ and does so in a way that does not collapse back into the Humean theory of pro-attitudes. There is considerable textual evidence, moreover, that it is Anscombe's own account, pursuing the individual path, upon which Davidson draws for insight. This is evidence that for Davidson pro-attitudes are individually constituted in part by evaluative commitments to the goodness of their objects, and that evaluative judgements make explicit these evaluative commitments implicit in such desires. Moreover, it is precisely this aspect of Davidson's view, an aspect that distinguishes it from the Smith/Scanlon hybrid attributed by Myers to Davidson, that is crucial to his defence of the Humean theory of motivation, and to understanding his commitment to intra-dimensional holism. Let me say more.

There is extensive textual evidence that Davidson follows Anscombe in taking individual pro-attitudes to be constituted in part by commitments to their objects as desirable/good,Footnote ⁹ and in taking there to be deep parallels both between theoretical and practical reasoning and between the role that pro-attitudes play in practical thinking and the role that beliefs play in theoretical thinking. He repeatedly characterizes evaluative judgements of desirability, or that an agent ought to act in some way, as making explicit the aim that is implicit in the relevant individual pro-attitude. Desires are holdings of their objects to be desirable, hence implicit in each individual desire is an evaluation of its object as desirable/good. Thus, “I want to eat something sweet, that is, I hold that my eating something sweet is desirable.”Footnote ¹⁰ For Davidson, to want to eat something sweet is, at least in part, to hold that eating something sweet is desirable. It is an “attitude of approval which the agent has toward the … proposition,” and the evaluative judgement is “the natural expression of his desire”: “all pro attitudes may be expressed by value judgements that are at least implicit.”Footnote ¹¹ The pro-attitude itself is an attitude toward its object as desirable, a holding desirable, as a belief is an attitude toward its object as true, a holding true. The expression of an evaluative judgement makes explicit this desirability evaluation that is implicit in the desire. What such judgements express, that is, are “the constitutive aims possessed by individual pro-attitudes” (DDTA, p. 153) that are implicit in such attitudes. This is the other fork in the interpretative road, the individualist fork, not the systemic aim approach attributed to Davidson by Myers.

3. Davidson Through Anscombe

Moreover, such an account is not at all mysterious, nor is its inspiration: it bears striking similarities to Anscombe's account. Davidson repeatedly invokes Anscombe (particularly Intention) in articulating these aspects of his account. Anscombe and Aristotle-as-interpreted-by-Anscombe are two of the most frequently cited philosophers in Davidson's early practical philosophy. It is clear that Davidson's account diverges from their accounts in significant respects,Footnote ¹² but these points of contrast, I suggest, highlight two even more fundamental points of comparison that Myers overlooks. First, Davidson's account, like that of his Aristotelean muses, is one upon which each desire is experienced under the guise of the judgement that its content is desirable/good,Footnote ¹³ as each belief is experienced under the guise of the judgement that its content is true. Goodness plays a parallel and equally fundamental role with respect to pro-attitudes as truth does with respect to beliefs. Anscombe asserts that just as “truth is the object of judgment,” good is “the object of wanting”Footnote ¹⁴; Davidson similarly asserts that we make sense of each person as “a believer of truths, and a lover of the good.”Footnote ¹⁵ He shares with Anscombe the conviction that to believe that p is, in part, to hold that p is true, within the context of other beliefs that we also each hold to be true, and to desire that p is, in part, to hold that p is desirable within the context of other states of affairs that we also each hold to be desirable.Footnote ¹⁶ Each such attitude is constituted in part by a normative commitment — to the truth of the content of the belief and to the goodness or desirability of the desired state of affairs. It is these distinctive holdings as good and as true — a distinctively practical commitment and a distinctively theoretical commitment — that account for the role that such attitude pairs play in jointly rationalizing actions.Footnote ¹⁷

Second, this robust parallel between belief and truth on the one hand, and desire and goodness on the other, suggests that Davidson takes from Anscombe and Aristotle the view that the practical judgement involved in desire is not simply a species of theoretical judgement. Practical and theoretical judgements are, for Anscombe and Aristotle, “distinct employments of reason.”Footnote ¹⁸ For Davidson as well as Anscombe, such a distinct practical judgement is always a necessary element of a rationalizing explanation, and is always an ineliminable element of any valid practical syllogism. We need to know both what an agent takes to be true for reasons, and what she takes to be good to do — to make true — for reasons, to have an adequate rationalizing explanation for her action. At the core of Davidson's endorsement of the belief-desire theory, I suggest, are his convictions that some such practical holding to be good or desirable is an ineliminable element, along with the appropriate theoretical holding true, of any adequate rationalizing explanation of an intentional action, and that it is because desires and beliefs are constituted in part, ceteris paribus, by such holdings good and true, in addition to various dispositions to infer, recall as relevant, act, etc., that they can play the roles that they do in practical and theoretical inference. In short, Davidson joins Anscombe and Aristotle in taking the first fork in the interpretive path, upon which desires, like beliefs, are properly understood as having what Myers characterizes as “constitutive aims” (DDTA, p. 153) possessed by individual attitudes. He does not, pace Myers, take the second, systemic fork. Moreover, I will show that it is precisely because he takes this first interpretive fork that Davidson's commitment to practical as well as theoretical holism becomes intelligible, and that his account can accommodate both internalism and objectivity within the belief-desire theory of motivation.

Of course, it does not follow that, because desires play a parallel role in the practical sphere to that played by beliefs in the theoretical sphere, such desires cannot be expressed by beliefs. Every practical holding that some state of affairs is good to bring about can be expressed as an assertion that the state of affairs is good or desirable.Footnote ¹⁹ If I desire that Ψ, I take Ψ to be desirable for reasons. And, if I take Ψ to be desirable for reasons, I will truthfully assert that Ψ is desirable. Such an assertion is true or false; it makes an objective evaluative claim. But evidence for the truth of the claim is provided by the practical reasons for taking the state of affairs to be good or desirable to bring about in light of the relevant beliefs. That is, the justification for the claim that Ψ is desirable is provided by practical inference from other ends taken to be desirable. It is, on this view, not at all surprising that Davidson's favourite value terms are ‘ought,’ seemingly in the decisive practical reasons sense, and ‘desirable,’ understood as what it is reasonable to desire.Footnote ²⁰

Myers, at points, frames the set of options in a way that elides this alternative from view. His suggestion is that desires are either commitments or they are dispositions to act. If the former, the worry is that the commitments are just beliefs, and Davidson is left with beliefs motivating. If the latter, then some form of the systemic aim approach seems the best way to go. But, as I have demonstrated, Davidson rejects any such exclusive disjunction when it comes to propositional attitudes, whether beliefs or desires. Beliefs, for Davidson, are holdings true constituted in part by commitments to the truth of their contents. But, as Myers himself points out, they are also constituted in part by dispositions to act in accordance with, infer, recall when relevant, etc.Footnote ²¹ So, beliefs are, for Davidson, complex states constituted by both commitments and dispositions. And desires are too. They are constituted in part by commitments to their contents as desirable, and in part by dispositions to bring their contents about through action guided by the reasons they have to take these contents to be desirable. It is this account of desires paralleling beliefs, in particular of both as constituted in part by both evaluative commitments and dispositions, that allows Davidson to account for both ethical objectivity (in virtue of the commitment) and internalism (in virtue of the disposition) within the framework of the belief-desire theory of motivation. The mistake, on the interpretation that I am offering, is to read Davidson as being forced to make a choice between commitment and disposition in the case of desires that he clearly rejects in the case of beliefs.

4. Implications

Note that such an interpretation naturally supports Davidson's commitment to intra-attitudinal holism, the view that “one's pro-attitudes are typically conditioned by other pro-attitudes that one has” (DDTA, p. 146). Myers’ Smith- and Scanlon-inspired interpretation of Davidson leads him to reject such intra-attitudinal holism as being a mistake on Davidson's part. But this Anscombean/Aristotelean interpretation of Davidson, upon which pro-attitudes are constituted in part by commitments to their objects as desirable/good, not only accommodates but even dictates such intra-attitudinal holism of desires paralleling the intra-attitudinal holism of beliefs. If I say that I want Ψ, implicit in this is my commitment to Ψ’s desirability. If you challenge my commitment to the desirability of Ψ, I defend it by appealing to the desirability of φ, in light of which Ψ’s desirability is vindicated. The content of the former is sensitive to — dependent upon — the content of the latter. But this is precisely because a commitment to the desirability of its content is implicit in a pro-attitude, ceteris paribus. This interpretation suggests that the mistake is not Davidson's holism, but taking Davidson's commitment to holism to be mistaken. Both this commitment to practical holism and the grounds for it are also found in Anscombe. She offers parallel accounts of theoretical and practical soundness, and characterizes the practically sound agents as wanting, hence taking to be good, the ends that are in fact good for the reasons that they are in fact good, where this is to take into account the complex ways in which such good ends condition each other, what means are permissible in the pursuit of which ends, etc. To take the object of her want to be good is thus to take it that acting in the pursuit of such an end is conditioned by and consistent with the other relevant good ends that structure “sound practical thinking.”Footnote ²² There is a parallel between practical inference and theoretical inference, and between the roles that beliefs and desires play in such inferences, that invites corresponding commitments to both theoretical and practical holism. Because Smith and Scanlon reject these parallels, they cannot make sense of a practical holism paralleling theoretical holism. But this is also why they are not a useful template for making sense of Davidson's commitment to practical holism, and why Anscombe is.

So, for both Anscombe and Davidson, a pro-attitude like wanting is a holding that some state of affairs is desirable, valuable, good. A belief is also an attitude of holding with respect to some state of affairs, but one of holding that the state of affairs is true. Each propositional attitude is located in rational space, but to desire is to hold to be desirable/good to do/make true for practical reasons, and to believe is to hold true for epistemic reasons.Footnote ²³ They are attitudes toward states of affairs on the one hand as good to make true, and on the other hand as true. Pro-attitudes, that is, are distinguished by Davidson in part by their contrasting direction of fit toward their contents — they involve dispositions to make the state of affairs that is their object true, to alter the world to fit their object.Footnote ²⁴ Beliefs have a contrasting direction of fit. I have already suggested why it is obvious, on such a view, that pro-attitudes condition the content of other pro-attitudes, as Davidson claims that they do. It is also clear on this alternative interpretation why Davidson believes that the belief-desire theory, internalism, and objectivity can be incorporated into his account with no inconsistency. Desires are constituted in part by dispositions to bring about their contents, determine the means to doing so, etc., and in part by evaluative commitments to the desirability of these contents. Thus understood, desires are dispositions to act that can play an ineliminable role in motivating actions, as the Humean theory of motivation insists. But the commitments that implicitly constitute desires can be made explicit in the form of evaluative judgements, assertions that have truth values, and that can be objectively true or false. If a desire is, ceteris paribus, an attitude toward a content as judged to be desirable, as Davidson and Anscombe both claim, then objective evidence that the content is not in fact desirable, provided through appeal to other desirable ends, will extinguish the attitude, just as objective evidence that the content of a belief is not believable, through appeal to other believable states of affairs, will extinguish the attitude. The belief-desire theory is reconciled to objectivity. Moreover, just as to sincerely “make an assertion by uttering a descriptive sentence” is to “represent himself as … having a certain belief,” so too to sincerely “commend an action” is to represent oneself “as holding the action to be desirable,”Footnote ²⁵ i.e., as having a certain desire, and desires dispose us to action. Objective evaluations are reconciled to internalism through the belief-desire theory.

Such an Anscombean interpretation of Davidson's own belief-desire theory can take his commitment to intra-dimensional holism at face value, as Myers’ alternative interpretation cannot. It can take at face value Davidson's claim that evaluative judgements make explicit the commitments to desirability/goodness implicit in individual desires. In contrast with Myers, it can take Davidson's repeated robust parallels between belief and truth on the one hand, and desire and the good on the other, seriously. Moreover, it readily explains why Davidson's focus is upon evaluative objectivity, not, as Myers’ interpretation suggests, upon normative realism. Evaluative judgements are objectively true if there are reasons (in particular, practical reasons) to promote their objects, and the correctness of these evaluations is “independent of even our enlightened personal reasons.”Footnote ²⁶ Davidson cautions that this fundamental question about the objectivity of values “should not be confused with realism,” and we should guard against various forms of “infection” of the objectivity of values question by the realism about value question.Footnote ²⁷ I am suggesting that Myers’ interpretation falls prey to such an infection. Like Anscombe, Davidson's focus is upon justification of the ends that we take to be good through practical inference from other ends that we hold to be good. If some of our ends are thus justified, there is ethical objectivity, and claims that we ought to pursue such ends, and that they are desirable, are true.

I have already suggested why such an interpretation fits better with the commitment to holism that is central to the interpretation argument. But I also suspect that such an interpretation has important implications for triangulation and pro-attitudes. For Davidson, to desire is to take the object of one's desire, the propositional content, to be desirable, such that one is disposed to bring about that content. It is an attitude toward a content as to be made true that disposes us to act in ways that will make it true. Such judgements of reasonableness implicit in pro-attitudes essentially involve triangulation among the world, my desires, and the desires of other agents. My desire to improve the taste of the stew presupposes a great deal of interaction with the world that I share with other agents. The desirability judgement implicit in such a desire, in taking the object of my desire to be desirable, purports to take into account reasons that other agents have to alter our shared world and my own other evaluative commitments. Triangulation, on this understanding, is presupposed in the very interpretation of oneself and others as desirers, and as having the individual desires that we do.

I have presented this alternative interpretation in a way that highlights the differences with Myers’ own interpretation of Davidson's belief-desire theory. But, in closing, I will once again point out how minimal in other respects these differences are. Both interpretations take Davidson to hold that desires are constitutive components of the motivations behind individual actions. And, although only one, mine, interprets Davidson as holding that evaluative commitments to desirability/goodness are constitutive components of individual desires, and that such constitutive commitments supply the agent's practical reason in intentionally acting, the other interpretation does maintain that the “main concern” of desires, understood as “dispositions to act in ways one believes will respond to the reasons for action one believes one has,” “is that one's actions accord with the truth about one's reasons” (DDTA, p. 163). Each, then, takes desires to play an ineliminable role in action explanation for Davidson, and both take desires to play this role, for Davidson, in virtue of their sensitivity to evaluative judgements concerning reasons for action. I look back to Anscombe, and Davidson's repeated invocations of Anscombe/Aristotle, as the key to unlocking the best interpretation of Davidson's account; Myers looks forward to Scanlon and Smith. Each interpretation attempts to incorporate Davidson's application of the interpretation and triangulation arguments to pro-attitudes, and each attempts to demonstrate why Davidson's view rejects the claim that objectivity, internalism, and the belief-desire theory form an inconsistent triad.

But the points of divergence are also important. Interpreting Davidson as taking the individualist fork, thereby adopting Anscombe's parallel between theoretical and practical reasoning and the parallel roles that beliefs and desires play in such reasoning, is, I have argued, not only central to the most plausible interpretation of Davidson's arguments, but also crucial to a full appreciation of his ingenious proposal for reconciling internalism, objectivity and the belief-desire theory. An additional implication is that the normative ethics supported by such an Anscombe-inspired account will not be a Scanlonian-style contractualism, as Myers suggests (DDTA, pp. 184–191), but a more Aristotelean approach emphasizing practical reasoning, the virtues, and other structural features that facilitate excellence in action.Footnote ²⁸