Can we outsource all the reasons?

Abstract

Where does normativity come from? Or alternatively, in virtue of what do facts about what an agent has reason to do obtain? On one class of views, reason facts obtain in virtue of agents’ motivations. It might seem like a truism that at least some of our reasons depend on what we desire or care about. However, some philosophers, notably Derek Parfit, have convincingly argued that no reasons are grounded in this way. Typically, this latter, externalist view of reasons has been thought to enjoy the advantage of extensional adequacy—that is, the ability to account for all the reasons we intuitively think people have. This paper provides a novel argument against this assumption by considering a type of case wherein the relative strengths of the agent’s reasons can only be adequately explained by reference to what she cares about. Adding some further assumptions yields that there are at least some internally sourced reasons.

1 Introduction: internalism, externalism, and extensional adequacy

Where does normativity come from? If we think of normative claims as ultimately grounded in or reducible to claims about reasons, we can ask: in virtue of what do we have the reasons we do?¹ Following Ruth Chang’s (2013b) usage, I will call source internalism the view that an agent’s reasons for action obtain in virtue of some appropriate sort of motivational state that they have, or would have, if suitably idealized, along with facts about what would promote the goal of that motivational state. Different source internalists have emphasized different kinds of motivational state as being central. On the other hand, call source externalism the view that agents have their reasons solely in virtue of the objective features of the world—all reasons are thus stance-independent in some sense.² Speaking about morality in particular, Russ Shafer-Landau puts the view thus:

The way I would prefer to characterize the realist position is by reference to its endorsement of the stance-independence of moral reality…the moral standards that fix the moral facts are not made true by virtue of their ratification from within any given actual or hypothetical perspective. That a person takes a particular attitude toward a putative moral standard is not what makes that standard correct. (Shafer-Landau 2003, 15)

Both source internalism and externalism have garnered several defenses over the past few decades.³ Recently, source hybridism, according to which there are both types of grounds for our reasons, has received some support.⁴

The debate is one about the fundamental grounds, or sources, as Christine Korsgaard (1996) puts it, of normative facts.⁵ There is a separate but related debate one can have, about whether normative facts can reduce to natural facts. Typically, source internalists tend to endorse such a reduction, while externalists tend to embrace non-naturalism.⁶ However, the latter debate is not the focus here.

Now, source externalists need not deny that some of our reasons can obtain in virtue of our desires in a derivative way. Suppose you want to concentrate on your work, but have this nagging desire to reply to some inconsequential email. This desire is distracting you from fully focusing on the work. The source externalist is happy to say that you might have decisive reason to reply to the email, due to the presence of this desire. However, for the externalist, the dependence will be derivative—the fundamental source of the reason will consist in some objective good that your answering the email is a means to realizing. In this case, plausibly, concentrating on and completing your work is the objective good to be achieved, and that is the more fundamental ground for your reason to reply to the email.

Source externalism and internalism both seem to enjoy distinct advantages. Externalism has been taken to possess the advantage of extensional adequacy—that is, the resources to account for reasons in a way that comports with our intuitions, particularly when it comes to prudence and morality.⁷ The problem for source internalism with regards to extensional adequacy is the following: there can be agents who simply lack those motivational states which could, given suitable idealization as needed, ground certain kinds of moral and prudential reasons. Thus, even if such agents are idealized in some way—taken to know all the relevant non-normative facts, be coherent with respect to their evaluative attitudes, avoid any errors in deductive or inductive reasoning, and so on—they would still lack certain kinds of reasons if source internalism is true. Yet, nevertheless, it’s plausible that such agents have reasons to act in the ways prescribed by morality or prudence. Here is Chang’s way of putting the problem: “Source internalism fails to guarantee the right substantive results about what reasons we have because the constraints it puts on desires are purely formal in nature. No formal constraint, however intricate, can guarantee the intuitively right answer as to what reasons we have” (Chang 2013b, 173).

Consider, for example, Parfit’s (2011) case of Future Tuesday Indifference. Here, we are to imagine an agent who wants to avoid pain in the future, except on Tuesdays. Such an agent would thus prefer undergoing a severely painful surgery on Tuesday, in order to avoid a mild injection on Monday. Of course, on Tuesday, they may wish they hadn’t avoided that injection, but no matter—let’s focus on what reasons this person has ex ante. Since this agent has no motivation to avoid pain on future Tuesdays, it follows on source internalism that they have no reason ex ante to avoid pain on future Tuesdays. Importantly, we can imagine an agent with such motivations even if they are coherent and fully informed about the relevant non-normative facts (Street 2009).

To many, however, this will seem implausible—the agent described above does have a reason to avoid pain on future Tuesdays. In such cases, at least, motivations and reasons seem to come apart. Similar points apply to agents without the requisite moral motivations. We can imagine a well-informed, coherent agent who simply doesn’t care about causing harm to others. Such an agent, source internalists have to say, has no reason to avoid causing harm to others, except indirectly insofar as harming others may frustrate other aims or goals he has, say, by attracting their retribution. We, of course, might have all sorts of reasons—to stay away, for example—with respect to those with such motivations.⁸ This result, again, seems like a cost of the view—many have the intuition that such agents have reasons not to harm others regardless of their actual motivations.

Unlike source internalists, externalists can account for these reasons because they take reasons to be stance-independent. Thus, it is the nature of pain that grounds the reasons that agents have to avoid inflicting it on others or on their own future selves. The fact that someone doesn’t care about not inflicting pain is neither here nor there. More generally, since source externalists don’t have to explain the existence of particular reasons by reference to an agent’s motivational states, they can feel free to posit reasons wherever they like, by appealing directly to the objectively desirable or undesirable features of the acts or consequences in question.

Why be a source internalist, then? Let me quickly sketch three chief strands of argument emphasized in the literature. The first is a consideration of ontological and epistemological simplicity. If normative facts obtain in virtue of, or better yet, reduce to, certain motivational states, then they can be accounted for in naturalistic terms. On the other hand, some have thought, objective normative facts—which putatively obtain independently of our motivations and yet are meant to have an authority over our deliberations—seem to be epistemically and metaphysically very strange objects that would be difficult to account for naturalistically (Mackie 1977). Second, source internalism is able to show how an agent’s reasons are considerations that she could (in some appropriately fleshed out sense of ‘could’) be motivated by. Some consider this to be an important consideration in favor of the view (Williams 1981; Korsgaard 1986). Third, according to some, to the extent that evolutionary theory can give etiological explanations of the evaluative attitudes we have, source externalism is either scientifically suspect or it threatens global moral skepticism (Street 2006).

In this paper, I argue that source externalism fails to secure the advantage of extensional adequacy, its supposed strong point. While the view may well have the upper hand in Future Tuesday Indifference and similar scenarios, I construct a distinct type of case and argue that externalists cannot account for the weights of the reasons that agents plausibly have in such cases. This means that in some sense, normativity originates at least partly from us as agents. Because source externalism does not have the clear advantage when it comes to extensional adequacy, we may have better reasons to accept either internalist or hybridist views of the grounds of normativity than have been hitherto appreciated in the literature.

The central case, which I will flesh out in greater detail later on, is the following. Imagine that Carla has some money to donate, and she can give to either the opera or the theater. Stipulate that the objective good that will be secured by her giving to either institution—the pleasure/enjoyment accrued to Carla and others, the aesthetic goods themselves, the community-building facilitated by the institutions, etc.—is equal. However, Carla values the opera more. In such a scenario, I will argue, Carla has more reason to give to the opera, and therefore her values play a role in determining the relative weights of her reasons. In other words, the fact that she has more reason to give to the opera than the theater obtains in virtue of the fact that she values the opera more than the theater.

What this means, at the very least, is that the weights of our reasons are sometimes determined by our values. In this sense, normativity at least partly originates from us as valuing agents. Moreover, if facts about weights are themselves to be understood in terms of reasons, then at least some of our reasons are internally sourced.

2 Obvious cases?

It might be thought, at the outset, that surely some of our reasons obtain in virtue of our motivations, in all kinds of mundane ways. Consider Mark Schroeder’s (2007) case of Ronnie and Bradley. There is supposed to be a party tonight that involves dancing. While Ronnie loves dancing, Bradley does not. In addition, Bradley, perceiving himself to be an awkward dancer, prefers not to be around situations where dancing is going on, due to fear of embarrassment. Thus, the fact that there will be dancing at the party is a reason for Ronnie to go, yet not a reason for Bradley to go. In fact, it’s a reason for him not to go. But what could explain this difference in their reasons other than what they desire or want to do?

Despite this initial worry, the source externalist who thinks that none of our reasons are ultimately grounded in our motivational states has a compelling response here. She may contend that what fundamentally explains the difference in Ronnie’s and Bradley’s reasons in this case is not simply the difference in the desires they have. Rather, there are objective goods that Ronnie’s going to the party would achieve, which Bradley’s going to the party would not. Those objective goods are what ultimately explain the differences in these two people’s reasons (cf. Parfit 2011).

Let’s flesh out the case a bit more. If Bradley wants to avoid dancing, we can imagine that dancing is not a pleasurable action for him. It would cause discomfort, and discomfort is an objectively bad state to be in. The fact that dancing, or being surrounded by dancing, would elicit discomfort in Bradley is the explanation for why the fact that party involves dancing is a reason for Bradley not to go to the party. The presence of the relevant desire does no fundamental explanatory work—at best it indicates the presence of other facts that explain why Bradley has a reason to avoid the party. Likewise, for Ronnie, the fact that he wants to go to the party is neither here nor there; rather, the positive hedonic states that Ronnie will find himself in at the party are the grounds of his reason to go there. On this picture then, desires can do explanatory work only in an indirect way, for example by (either causally or non-causally) ensuring that other facts obtain, which do the fundamental grounding-explanatory work.

Parfit (2011, 54–56) notably distinguishes between “hedonic likings” on the one hand, and what he calls “meta-hedonic” desires on the other. He argues that in the typical sorts of cases where it can be tempting to think that our reasons obtain in virtue of our desires, it is rather these likings (or dislikings) that are doing the more fundamental explanatory work. Pleasurable states—e.g. the feeling of enjoying oneself while dancing at a party—are objectively good for people to be in (at least in ordinary cases) and therefore give rise to reasons for agents to act so as to get themselves in those states.

This style of source externalist response seems very plausible in the above case. For, imagine a situation where Bradley desires not to go to the party, but in fact, Bradley would enjoy himself there, would meet interesting people, and so on, and nothing bad would occur. Suppose also that Bradley knows these things. In this case, despite Bradley’s desire not to go to the party, he has good reasons to go. If nothing bad and only good would come of him going to the party, then we might also plausibly say that his desire to avoid the party gives him no reason to avoid the party. The desire, one might think, is an irrational one for Bradley to have, in the same way the kleptomaniac’s desire to can be irrational. Analogously, to use another example from Parfit (2011, 73–82), the fact that Φ-ing would cause some agent to be in agony gives that agent no reason to Φ even if they desire to be in future agony.

However, noting this objectivist response suggests a potential strategy for the friend of internally sourced normativity. That is the following: consider the prospect of cases where A has more reason to Φ rather than Ψ, where A has the relevant motivational pro-attitude towards the object of her Φ-ing, whereas she lacks it with respect to the object of her Ψ-ing. Nonetheless, the objective goods that would be secured by her doing either action are equal in importance. If there are such cases, then A’s greater reason to Φ rather than Ψ can only be plausibly explained by the presence of her motivational pro-attitude towards the end which her Φ-ing promotes. And if so, externalism cannot account for the relative weights of the reasons we have. In that sense, source externalism is extensionally inadequate.

Furthermore, consider the fact that A’s reason to Φ has greater weight than her reason to Ψ. The notion of ‘weight’ here is itself a normative notion. What this means, in effect, is that A ought to deliberate, or it is correct for her to deliberate, in such a way as to give more weight to her reason to Φ. It is plausible that this entails that A has a reason to give more weight to her reason to Φ, as opposed to her reason to Ψ. Suppose the buck stops here in the sense that A does not have any reason to give more weight to her reason to Ψ.⁹ Now, A’s reason to give more weight to her reason to Φ, I contend, is presumably grounded in the fact that A has the relevant motivational pro-attitude towards the object of her Φ-ing. Consequently, if it is right to characterize the weights of reasons themselves in terms of further reasons in this way, then source externalism cannot account for all our reasons. That is, it’s not merely that some of our pro-attitudes affect the weights of the reasons we have, but moreover (since facts about those relative weights themselves are, or entail, facts about further reasons), they ground the very existence of at least some of our reasons.

Before proceeding, I would like to flag one worry the reader might have. It might be thought that if “hedonic likings” in fact involve a desire-like attitude, and if such likings ground some of our reasons, then there are stance-dependent, internalist reasons after all. The idea would be the following. Surely, facts about our current welfare (sometimes) give us reason for action. The fact that Φ-ing would increase my current welfare, compared to the other available actions, is typically a defeasible reason for me to Φ. Moreover, even if it is not the only thing that contributes to welfare, pleasure is plausibly at least one of the ingredients of welfare (and pain is one of things that detracts from welfare). Thus, if pleasures have their status as pleasures in virtue of their relation to a desire-like attitude then there are some internalist reasons for action.

This is one of the strategies pursued recently by David Sobel (2019) as a response to Parfit’s argument. Sobel contends that “hedonic likings” are the sort of thing that, if they ground reasons, then those reasons are internally sourced (or stance-dependent, in his usage) in the sense that is relevant to the present debate. The key point is that hedonic likings are not simply felt sensations, but also involve a pro-attitude. Now, there is an ongoing debate in the well-being literature about whether pleasures are simply felt sensations or essentially involve such attitudes.¹⁰ Some readers may be convinced by Sobel’s point, insofar as they agree that a felt sensation theory of pleasure is inadequate. However, even if the felt sensation theory of pleasure is true, the argument in this paper, if successful, establishes some internalist reasons. My argument thus moves the debate forward by not assuming a disputed view of the nature of pleasure.

Another response the source externalist might offer here is that likings or dislikings, even if they involve attraction or aversion, are not the right kind of mental state that can genuinely ground reasons while still retaining the theoretically attractive features of internalist reasons. Indeed, not just any mental state can ground internalist reasons. Consider a toy theory of reasons on which our reasons obtain in virtue of particular experiences of color. Thus, when we experience seeing something green, we have a reason to Φ, when we see red, we have reason to Ψ and so on. This theory appeals to our mental states in explaining the sources of our reasons, but it’s not internalist in an important way. For, it doesn’t tie our reasons to our evaluative perspective. And it would not satisfy the desiderata that have typically appealed to internalists like Korsgaard or Williams—for example, it would not bear the appropriate connection to our agency. In a similar vein, then, the externalist might claim that just as experiences of color are not the appropriate kind of mental state to ground internalist reasons, neither are likings or dislikings.

While this move might initially appear ad hoc, there seem to be good reasons to think hedonic likings are importantly different from desires or values. Parfit (2011, 54) notes for example that hedonic likings cannot be satisfied, in the way desires can be. My desire for vanilla ice-cream can be satisfied by my purchasing and consuming some. But it makes no sense to say that my liking of vanilla ice-cream can be satisfied. Second, likings or dislikings cannot be aimed at the future. You can desire to have the pleasurable experience of drinking a piña colada in the future, but you cannot like this future pleasurable experience now. Parfit further stresses that desires can be reasons-responsive, whereas likings cannot. If I desire to do something that’s counterproductive, you can give me reasons for why I shouldn’t want to do that thing. In other words, desires are typically subject to rational evaluation and revision. However, liking vanilla more than chocolate doesn’t seem to be subject to rational evaluation in the same way, according to Parfit.

I do not mean these considerations to be decisive. I just want to note that it is at least controversial whether hedonic likings are the right kind of mental state for grounding internalist reasons. However, if we can identify cases where agents have normative reasons whose strengths (or existence) cannot be explained either by reference to objective reason-giving facts or hedonic likings/dislikings, then we will have identified internally weighted (or sourced) reasons which even someone with the theoretical commitments of Parfit vis-à-vis likings can accept.

3 Valuing

In the next section, I will discuss in detail a case where the agent in question has more reason to choose one option rather than another, in virtue of their motivations. The key motivational state will be that of valuing. In this section, I describe the notion of valuing and provide some reasons for focusing on this pro-attitude rather than others.

Defenders of source internalism have often focused on desires as the central motivational state that gives rise to reasons (Schroeder 2007; Smith 1994). However, one familiar problem with desires is that we can often be alienated from them (Frankfurt 1971; Watson 1975). Not so with values. As David Lewis puts the point:

[W]e’d better not say that valuing something is just the same as desiring it. That may do for some of us: those who manage, by strength of will or by good luck, to desire exactly as they desire to desire. But not all of us are so fortunate. The thoughtful addict may desire his euphoric daze, but not value it. Even apart from all the costs and risks, he may hate himself for desiring something he values not at all. It is a desire he wants very much to be rid of. (Lewis 1989, 115)

I may have a desire to smoke but would like to be rid of that desire. The things I value—health, longevity, and the other goods that these are preconditions for—are at odds with my smoking. It is natural therefore to think that insofar as our motivational states can give rise to reasons, our values should take precedence over our desires.

Valuing is not equivalent to desiring, but neither is it a belief. Valuing something can come apart from believing that thing to be valuable. For instance, one might believe that the traditional cultures native to Siberia are valuable without in any serious sense valuing them. Or one might acknowledge how a particular instrument is valuable within the world of music, without oneself valuing the kinds of music produced with it. This difference will be important in the argument to follow.

Here is Samuel Scheffler on the distinction: “There are, for example, many activities that I regard as valuable but which I myself do not value, including, say, folk dancing, bird-watching, and studying Bulgarian history. Indeed, I value only a tiny fraction of the activities that I take to be valuable.” Scheffler clarifies that this distinction isn’t merely one between activities that one participates in as opposed to those that one does not. You can participate in an activity without properly valuing it, insofar as it leaves you cold. Furthermore, he says, “there are many things besides activities that I regard as valuable, and many of these are also things that I myself do not value. There are, for example, many paintings, historical artifacts, and literary genres that I believe to be valuable but do not value myself. With respect to these things, the suggestion that to value them is to believe they are valuable and to participate in them oneself would obviously make no sense” (Scheffler 2010, 21).

Valuing centrally involves two components, one emotional and one deliberative. The emotional component involves a certain vulnerability with regards to the object of the value and its relationship to us. If you value a particular friendship, then you will be prone to feelings of disappointment, frustration, melancholy, and the like, if the friendship comes under strain. If you value a particular painting, you will feel sorrow if you learn that it was destroyed in a fire.

The deliberative component of valuing something involves taking certain considerations as reasons for action (Frankfurt 2004; Scheffler 2010). Thus, valuing—or alternatively, caring about—a person involves taking their welfare as a reason for action. If you care about someone, you take the fact that they need help as a reason for you to provide them with help (insofar as you can). Valuing different sorts of objects involves treating different kinds of considerations to be reasons for action. Valuing a painting or artform might involve taking steps to try and preserve it. It might also involve a disposition to share the virtues of that painting or artform with others.

The notion of taking certain considerations to be reasons for action can be understood in terms of certain non-doxastic dispositions, so that it need not explicitly involve a belief about what reasons one has. At any rate, if taking something to be a reason were just a matter of believing that thing to be a reason, then the creation of reasons from our values would involve a paradoxical situation wherein believing something to be a reason makes it so. Fortunately, however, we can understand taking something to be a practical reason as a disposition to engage in certain forms of practical reasoning. As a very rough gloss, I take some consideration R to be a pro tanto reason to Φ just in case when deliberating about whether to Φ, the perceived fact that R is weighed in favor of Φ-ing. Thus, if we think of deliberation regarding pro tanto reasons on the model of putting weights on a scale so that the side on which the scale tips is analogous to the course of action that is chosen, then taking something to be a pro tanto reason, of a particular weight, in favor of Φ-ing, is akin to being disposed to place an object of a particular mass on one side of the scale.

4 Reasons in virtue of values: a case

In this section, I present and discuss a case which suggests that an agent’s valuing something can ground the fact that they have more reason to take some action rather than another. Moreover, this dependence is non-derivative: the ultimate ground of the fact that they have more reason to take that action must include the fact that they have a particular value.

The strategy pursues an isolation test. In particular, we can consider cases where an agent can promote, to the same extent, two ends which are equally objectively valuable and are recognized by the agent as such, but nonetheless the agent values one of the ends more. I will claim that in such cases, the agent has more reason to pursue the end which they value more, and thus the relative weights of these reasons are internally determined. This is because, since the reasons that could plausibly be objectively given are in equipoise, the difference has to be explained in some way by the agent’s values.

Some clarifications are in order before proceeding. Philosophers have recently distinguished between the reasons there are for an agent to do something and the reasons an agent has to do that thing. For an agent to have a reason R, the fact that R must be in her ken in some way—she must have the appropriate epistemic access to R. To illustrate, suppose that unbeknownst to you, there is a button nearby such that if you were to press it, climate change would be halted (without any dire side-effects, let’s assume). Now, there is a very strong reason for you to press the button, but you do not have, or possess, that reason. The notion of having reasons is important because these are the considerations that are normatively relevant to one’s deliberation. Good deliberation, from your perspective, would not recommend the action of looking for that button, because you would lack the requisite epistemic access.

In addition, philosophers have recently emphasized the distinction between reasons and rationality. Some think there is a tight connection between the reasons one has and what it’s rational for one to do, so that being rational is simply a matter of properly responding to one’s possessed reasons. Others deny this.¹¹ The question I’m interested in here is not, in the first instance, the question of what it’s rational for the agent to do. Rather, the notion that is key for our purposes is that of having a reason.

There is also the familiar distinction between normative and motivating reasons. Motivating reasons are facts which figure in the explanation of why some agent chooses a particular action (for example: the agent desired gin, and believed the bar offered gin drinks).¹² Normative reasons, on the other hand, are facts that count in favor of an agent choosing some course of action, or choosing some action as contrasted with some other action.¹³ What’s operative in the argument below is the notion of a normative, as opposed to motivating, reason.

With these clarifications in mind, consider the following case.

OPERA: Carla values the opera as an artform. This consists of her, among other things, taking there to be reasons to admire good operas, taking there to be reasons to read up about the opera more, and taking there to be reasons to go and see the opera when she gets a chance. She also has an emotional vulnerability with respect to the opera—for example, she’d feel irked if she overheard someone say that operas are boring. Furthermore, while she believes that the opera is valuable, she does not think that it is more valuable than the theater. Suppose, as a matter of fact, that each artform is equally (objectively) valuable. Now suppose she can donate a thousand dollars to either the opera or the theater. Where does she have more reason to donate and why?

Carla manifestly has more reason to donate to the opera than the theater. And plausibly, this is simply because she cares about the opera more. It can’t be because the opera is objectively more valuable (and thereby more reason giving) than the theater—the case stipulates that either artform is equally valuable, and moreover, that Carla recognizes this to be the case.

It is plain here to me that Carla’s having more reason to give to the opera rather than the theater obtains primarily in virtue of her valuing the opera more, rather than any objective goods that might be realized by her donating to the opera that wouldn’t be realized if she were to give to the theater instead. Here is one potential response that it might be tempting to give on behalf of the externalist, however. One of the things that is objectively reason-giving, if anything is, is the prospect of pleasure or satisfaction. Carla would presumably feel more satisfaction if she gave to the opera, given that she values the opera more than the theater. This prospect of satisfaction generates an objective, stance-independent reason for Carla to give to the opera.

However, the case can be modified so as to block the prospect of satisfaction. If Carla still has greater reason to give to the opera rather than the theater, then this fact cannot obtain entirely in virtue of the prospect of satisfaction. The only plausible fact that can explain the greater weight of this reason, then, would have to be that Carla values the opera more than the theater. And if that is right, then at least in some cases, the relative weights of our reasons are determined by our values.

Here is the modified case I have in mind.

OPERA+: Carla has been given a pill which will make her forget where she donated her thousand dollars as soon as she clicks the “submit payment” button on the online form. She will also forget that she had the thousand dollars lying around in the first place: let’s say her memory will be reset to t-24 h, when she hadn’t yet discovered that she had a thousand dollars in an old bank account. Suppose now that Carla is told about this memory-erasing pill and its effects. Where does she have more reason to give?

Carla plausibly still has more reason to give to the opera. However, this fact cannot obtain in virtue of the prospect of satisfaction. The only plausible explanation for the greater strength of the reason is the fact that Carla values the opera more than the theater.

Now, one might think that even if she will forget where she donated once she submits the payment, nonetheless there is still that modicum of satisfaction that she will experience in the process of filling the form and clicking the button. Perhaps this (small) amount of satisfaction breaks the tie, making it the case that Carla has more reason to give to the opera. However, it’s plausible that Carla’s reasons to give to the opera substantially outweigh her reasons to give to the theater. In other words, her reasons to give to the opera are a fair bit stronger than her reasons to give to the theater. And it’s hard to see how a few seconds of satisfaction can bear that amount of normative weight.

One way to further press the point is to assume for the sake of argument that the theater is just a tiny bit more objectively valuable than the opera so that even with the inclusion of those few seconds of satisfaction, the externally sourced reasons are in equipoise. Even assuming this though, it’s plausible that Carla still has more reason to give to the opera.

A more difficult objection concerns the reasons individuals have to live coherent lives. Given a realistic fleshing out of the case, Carla’s valuing the opera presumably occurs against a backdrop of various activities and projects she has undertaken over time vis-à-vis the opera. These could include things like having substantial knowledge of various operas and singers, having attended or seen a variety of performances, having built relationships with fellow opera enthusiasts, etc. In light of this history, Carla’s life would form a more coherent whole if she donates to the opera, as opposed to donating to the theater. The latter action just does not fit within her life, viewed as a whole, as compared to the former. Donating to the opera would make for a life with a better story, as it were. The externalist might point to the coherence of a life and its constituent projects as providing the ground for Carla’s stronger reason to give to the opera.

However, this objection can be successfully addressed, I think, if we delete the important ingredient from the case—namely, Carla’s valuing the opera. Suppose that even though Carla has spent a lot of time enjoying and engaging with the opera over her life so far, she just doesn’t care about it anymore. It’s now as dry to her as the theater is. Here, it seems plausible that her reasons to give to the opera would be greatly attenuated. If she really doesn’t care about it anymore, she might as well flip a coin. This is not to say that we can’t have external reasons given to us by the consideration of how coherent our life would be, but to simply notice that our values can strengthen reasons with an independent force.

At this point the externalist might concede that our values affect the relative weights of our reasons but nonetheless insist that all our reasons are externally sourced. What results then is an internalism (partial or full) about weight but full externalism about source. However, one challenge for defenders of this view, as discussed earlier, is to cash out what it means for the weights of our reasons to be internally sourced without ultimately having to appeal to the existence of some internally sourced reason.

Thus, suppose Carla’s reasons to give to the opera are indeed weightier than her reasons to give to the theater. What this means is that it is correct for her to deliberate in such a way so as to place more weight on her reasons to give to the opera. Reasons fundamentalists (like Parfit, among others) will want to analyze correctness, qua normative notion, itself in terms of reasons. As a result, it will have to be the case that Carla has reason to place greater weight on her reasons to give to the opera. However, what the case suggests, given its stipulations, is that the existence of this further reason is grounded in the fact that Carla values the opera more. If that’s right, then we cannot outsource all our reasons.

5 Chang’s hybridism

Before concluding, I want to contrast the arguments and the resulting view I have presented thus far with those developed in Chang (2013b), which are in some respects motivated by similar concerns. Chang argues for a hybrid view of the source of normativity; on her picture, some of our reasons are given to us by the world, while others we create ourselves. The argument of this paper is not committed to a hybrid view of source, but rather simply the claim that the weights of at least some of our reasons are internally sourced. However, Chang does argue against the source externalist account of reasons by noting that in certain cases of tough choices, the account lacks the resources to adequately explain why the agents have the reasons they do.

Chang centers her critique of source externalism on what she calls “hard cases.” A hard case is one where people can reasonably disagree about how the balance of reasons comes out. Note that my arguments in this paper do not rely on such hard cases. In fact, it seems to me that the case I have presented is easy—the agent clearly has more reason to act in a way that promotes her values.

Chang’s primary example of a hard case involves Jane, who is deciding between becoming an artist or a banker. Let’s suppose the case is hard in the sense that it elicits reasonable disagreement. The source externalist tells Jane that the balance of reasons weighs in favor of her being a banker. If Jane asks for further explanation, Chang contends, the externalist is forced to simply say, “that’s just how things are” (Chang 2013b, 171). This is the problem of “explanatory shortfall,” where the externalist runs out of explanation just where it’s needed.

It’s not obvious that the externalist is forced into this position, however. For, she can in this case point to all the goods that Jane might achieve by being a banker. It seems as though the source externalist is simply being asked to do some first-order normative theorizing. If that’s the demand, she might claim (depending on her first-order moral views), for example, that Jane would able to give lots of money to charity if she became a banker, given the typical salaries, and thereby improve many lives. Of course, there may be reasonable disagreement about how demanding morality can be or uncertainty about how effective various charities are, but such disagreement and uncertainty are a mundane fact of life.

Furthermore, the hybrid view Chang ends up defending holds that some of our reasons are given by objective features of the world, whereas others are created by us through acts of will. These reasons are voluntarist. In contrast, on my view, the non-external reasons we have are grounded in facts about what we value, or care about. The latter picture enjoys some important advantages over the former. The key issue is that it seems plausible that sometimes, our internal reasons are not transparent to us (Arpaly 2000; 2002).

Consider a college student, Salim, who thinks he has most reason to go into law. Growing up, he has been told about the importance of having a high-paying career. His parents are lawyers and many of the family’s friends are in professional occupations like medicine or engineering. When he reflects on it consciously, it seems to him that the values realized in a potential legal career—prestige, fulfilling family expectations, contesting important cases to make a long-term impact on society, etc.—are the values that he has most reason to promote. However, what Salim is really passionate about and good at is teaching. When he’s volunteered to visit high-school classrooms, he’s always found the experience to be fulfilling. The fact that he will find more meaning within a teaching career, and as a result has decisive reason to pursue this career, might well take a process of self-discovery and sustained experimentation. Thus, while Salim might think that a legal career will most fit his values, in reality, his values will most be realized if he pursues a career in teaching.

If valuing is a matter of emotional vulnerability coupled with certain deliberative dispositions, it’s easy to see how we might be mistaken about what we truly value. Our dispositions need not be transparent to us, and discovering them will often involve putting ourselves in different circumstances. To use a simple example—to find out whether one likes vegemite, one will have to actually try it. Similarly, for Salim to reasonably become confident about what career to pursue, he might have to try his hand at different things—perhaps by working in a year-long post-graduation teaching program and/or working as a paralegal. If internalist reasons are grounded in one’s values, then, it’s not mysterious how we could be mistaken about them. On the other hand, if such reasons are grounded in willing, then it’s harder to see how we could be mistaken with respect to them. Our willing, presumably, is transparent to us.

Moreover, it seems plausible that Salim could will certain considerations—prestige, money, etc.—to be reasons to embark on a legal career, while still having more reason to go into teaching, given what he actually cares about. That he has more reason to go into teaching might have to be something that he discovers over time. Insofar as this type of situation can occur, it’s plausible that the willing some considerations to be reasons by itself is not the fundamental ground of one’s internally sourced reasons. Or more weakly, our values (which are not always transparent to us) have to enter the explanatory story somehow, independent of the willing.

Indeed, it seems most natural to think of willing as the output of deliberation, and reasons as the proper inputs. Normative reasons are the proper bases on which a rational agent wills, or decides, some course of action. If this is right, then a picture according to which willing generates reasons (in more than a derivative way) seems to put the cart before the horse. However, this is not a worry on the view presented here, whereby our values—which are emotional and deliberative dispositions we have—can make certain considerations normatively weightier for us.

6 Conclusion

In this paper, I have argued that our reasons sometimes have the force they do because of what we care about, or value. This is best brought out by looking at cases where two actions will each secure something of equal objective value, but nonetheless, the agent differentially cares about the goods in either case. Thus, for example, if someone cares about the opera more than the theater, she has more reason to donate to the opera, even supposing that the theater is just as valuable as the opera. The greater weight of this person’s reason to give to the opera can only be explained by appealing to her values. Moreover, if the normatively loaded notion of weight here is itself to be analyzed in terms of further reasons, then there are at least some internally sourced reasons.

This last point has particular dialectical import because source externalists like Parfit have also accepted reasons fundamentalism, on which reasons are the basic building blocks, as it were, of normativity. All normative facts, according to this view, ultimately boil down to facts about reasons. Note though that even if weight is not further analyzable in terms of reasons, there are two resultant options. Either weights are the (or at least among the) fundamental building blocks of normativity or they are not. If they are, then the OPERA case suggests that at least some normative facts obtain, non-derivatively, in virtue of what we care about. On the other hand, if weights are not analyzable in terms of reasons, and further they are to be understood in terms of some other notion, say fittingness, then the challenge for the externalist will be to cash out what fittingness-facts are grounded in without appealing to our values. In other words, why is it fitting for Carla to decide to give to the opera rather than the theater? Plausibly, it’s because she values the opera more. If that’s right, then our values are (at least sometimes) fundamental sources of normativity.

Furthermore, my argument does not depend on an attitudinal view of pleasure and pain, which is a matter of dispute. Even if pleasures and pains do not involve desire-like attitudes, but rather only felt sensations, the cases discussed in this paper suggest that the weights of some of our reasons are stance-dependent.

If all this is correct, then the source externalist or stance-independent picture of normativity, defended notably in Parfit (2011), cannot account for all the reasons people have. It has been commonly thought that the problem of extensional adequacy is a challenge facing the internalist, not the externalist. What the internalist gives up in extensional adequacy, she tries to make up with respect to other virtues such as simplicity or methodological naturalism. However, if this paper’s argument succeeds, then source externalists do not have the clear upper hand in terms of extensional adequacy. While they may well be the only ones with resources to render plausible verdicts in cases like Future Tuesday Indifference, they lack the resources to render plausible verdicts in cases like OPERA. Overall, this lends more plausibility to source internalism, as well as source hybridism, according to which there are both internally generated and external given reasons. You just can’t outsource all the reasons.