1. Grounding normativity: internalism, externalism, and hybridism

Consider the commonly accepted view that practical reasons are facts that count in favor of an agent performing some act or other. We can understand this view as one that represents practical normativity as a relation – the reasons relation – which we can express in the following form: R(f, A, Φ), where f is the argument place for some fact, A for some agent, and Φ for some action. One meta-normative issue to be sorted out concerning that relation is determining the correct answer to the following question: when some particular fact stands in the reasons relation, in virtue of what does it do so? Answering that question is to identify the way in which practical normativity is grounded.Footnote ¹

Competing accounts of normative grounding have traditionally belonged to one of two camps. In the Source Internalist camp, it is maintained that all facts that constitute reasons do so in virtue of a relation in which they stand to the actual or hypothetical attitudes of agents. In the opposing Source Externalist camp, it is maintained that all facts that constitute reasons do so either brutely, or in virtue of a relationship in which they stand to facts about value.Footnote ² In this debate, both internalists and externalists have typically taken a hardline stance, insisting that no reasons are grounded in the way that the opposing camp insists that all are. It is only very recently that there has been serious attention paid to the possibility of a more inclusive camp: according to Source Hybridism, some, but not all, normative grounding is as the internalists understand it, and some, but not all, normative grounding is as the externalists understand it. This hybrid position has most forcefully been argued for by Ruth Chang, but its potential merits have also lately drawn the attention of Sarah Paul and Jennifer Morton (Chang Reference Chang, Sobel and Wall2009, Reference Chang2013; Paul and Morton Reference Paul and Morton2014).

I am also inclined to think that a hybrid account is correct, and so see practical normativity as ultimately deriving from two distinct sources. However, this paper is not an attempt to provide a positive argument to the conclusion that Source Hybridism is true. Rather, this paper is about rescuing hybridism from possible objections. So, in what follows, I will nearly always be on the defensive. This kind of defensive work is obviously important in making the strongest overall case for hybridism; even if its competitors suffer from serious flaws, it may be that hybridism is even worse off in other respects, and so ought to be rejected. My aim is to help show that this is not so.

Although I will not be directly interested here in providing positive reasons to accept hybridism, it is important that I provide at least some general comments about why I find the view attractive, so as to provide needed context once I turn to the paper’s primary defensive purpose.

In short, and perhaps unsurprisingly, I find Source Hybridism appealing for the simple reason that I think that the problems that it faces are not as serious as those that beset either of its ‘pure’ competitors. Both internalism and externalism face challenges to which I do not now see adequate solutions, and they are challenges that hybridism easily avoids. Take internalism first. My complaint against internalism is a familiar one: that the theory cannot account for the existence of all of the reasons that there are. I am one of those philosophers who takes very seriously cases involving the following sorts of characters: ideally coherent agents who want to cause themselves to be in agony for its own sake, who want to starve themselves for the sake of a body shape incapable of sustaining life, or who lack any pro-attitudes that could ground reasons for them to act in ways that morality appears to demand. Of course, many internalists can take scenarios of this sort seriously as well, and then go on to argue that the view simply yields the correct verdict in these cases, either because the view can actually ground a reason for the agent to avoid causing herself agony (or whatever), or because, contrary to the intuitions of people like me, the subjects in these cases really do not have any such reasons.

Street’s (Reference Street2009) ‘In Defense of Future Tuesday Indifference’ is an excellent example of just such an internalist response to these cases, in my view. Although I ultimately think that her arguments therein in defense of internalism are unsuccessful, I think that she is exactly right about the import of the odd characters that populate them for understanding exactly how externalist and internalist approaches differ, and for forcing those theories to render different first-order verdicts about the existence of reasons across cases.Footnote ³ Again, I have provided no arguments here that the cases in question pose a genuine challenge for Source Internalism, and nor do I intend to; I offer those arguments elsewhere, and will assume throughout that they represent a substantial challenge to Source Internalism (Behrends, Reference Behrendsunpublished). Understanding this motivation of mine for resisting internalism will prove helpful once I begin considering objections to Source Hybridism.

Trouble arises for Source Externalists, I think, when attempting to accommodate the apparent truth that our pro-attitudes are in some way normatively relevant. When choosing between two otherwise equally choice-worthy options, it really does seem right that I have more reason to choose the first if I happen to desire it, prefer it, be inclined toward it, or have toward it whatever pro-attitude we ultimately think is most normatively important.Footnote ⁴ There are various ways in which externalists might attempt to make sense of this: by arguing that desire-promotion is itself good, that desire-frustration is itself bad, that there are important costs to acting against inclinations of the relevant sort, or perhaps even by arguing that the fact that Φ-ing would promote A’s desire is itself a reason for A to Φ, the normativity of which is not derived relationally via A’s pro-attitudes, but which is itself a brute, external normative truth. Undoubtedly there are other options, as well. All I mean to be doing here is acknowledging that externalists have live options available to them in attempting to account for the normative significance of pro-attitudes. I think that successful arguments can be given against these various externalist options, but I will not attempt to characterize those arguments here, as doing so would distract us from the current paper’s project.

To briefly summarize, I am attracted to hybridism because internalists cannot account for all of the reasons that agents have across cases, and because externalists cannot account for the normative significance of pro-attitudes in a plausible way. But, as I have said above, I do not take any of the foregoing remarks to constitute arguments in hybridism’s favor, and this paper is not directly aimed at making the case that there are good positive reasons to be a hybridist. The remarks so far are intended only to situate the coming discussion in a broader context, and to make clear with which assumptions I am beginning.

In what follows, I articulate and respond to three potential objections to Source Hybridism. In the case of the first two, I try to show that the objections are completely answerable, and that they therefore do not show the view to be at any disadvantage whatsoever with respect to its competitors. As for the third objection, I argue that even if it does reveal a theoretical disadvantage of the view, it is not a very serious one, and certainly not one that could be decisive all on its own. I then conclude with a more speculative line of reasoning that aims to show that the third objection, like the first two, fails to establish even a weak theoretical drawback for Source Hybridism. If the speculative line of reasoning holds up, then I will have shown that none of the three objections canvassed here demonstrate hybridism to be beset by problems that its competitors avoid.

2. Source Hybridism and the weight of reasons

The first objection to Source Hybridism that I wish to consider is best understood in light of a different sort of objection that is commonly leveled against Source Internalism. I said above that one advantage that hybridism has over internalism is that internalists cannot account for all of the reasons that agents have in particular cases. It will be helpful now to have one such case in mind. Here is a case that Derek Parfit uses in his own attempt to undermine internalist views:

It seems that internalism entails that in Case Two I have decisive reason to cause myself to be in agony for its own sake, but, the complaint goes, no agent could have decisive reason to do such a thing, so internalism must be false.

If hybridism really does hold an extensional advantage over internalism with respect to scenarios like Case Two, then hybridism must avoid the possibility that agents have decisive reasons to do terrible things for their own sakes. It is easy to see how the first step in avoiding that possibility is to go. Hybridists, unlike internalists, can point to the intrinsic badness of the agony as grounding a reason against inflicting it on oneself. But notice that this alone is not sufficient to get hybridists all that we need. For the view is one that does recognize an internal source of normativity, and so is committed to the result that in Case Two I do have some reason to cause myself to be in agony for its own sake. There is an externally sourced reason for me to do just the opposite, but the mere existence of that latter reason does not entail that the internally sourced reason cannot be decisive.Footnote ⁵

What is needed in order for the hybridist to comfortably maintain that the internally sourced reason in Case Two is not decisive is an account of weighting that guarantees that the externally sourced reason will win out. And it is here that an objection looms, or at least a kind of challenge: the hybridist must have some way of showing that there is a general account of weighting that can make sense of differently sourced reasons competing with one another, and that will render the correct results in the range of cases that she takes to undermine Source Internalism.

In dealing with normative weight, the Source Hybridist can take either of two very broad approaches: she can tie her theory of normative weight directly to her theory of normative source, or she can let the theory of normative weight stand on its own, free to be applied independently of any particular approach to normative grounding. I begin here by exploring the first route, taking as my starting point a hybrid approach to weighting favored by Ruth Chang. After motivating the kind of approach that she endorses as best I can, I will argue that it is nevertheless not adequately supported. Luckily for hybridists, though, a more general approach to weighting is available, and that approach is certain to yield the right results in cases like Parfit’s Case Two.

Chang endorses a particular version of Source Hybridism, one that recognizes both externally sourced reasons and internally sourced reasons that are grounded in willings. Although Chang identifies a particular kind of mental action in which the internally sourced reasons are grounded, and I am trying to remain mostly neutral on that issue here, we can attempt to abstract away from that detail of her view when considering whether her approach to weighting is one that can serve a Source Hybridist of any stripe.

According to Chang, there is a kind of lexical priority of externally sourced reasons over those grounded in the will; external reasons, on her view, always outweigh will-based reasons (Chang Reference Chang2013). More precisely, she maintains that whenever external reasons

So, according to Chang, will-based reasons could affect what an agent has all-things-considered reason to do only when that agent’s alternatively sourced reasons do not on their own decisively favor one option over any of the others. And, according to her, that restriction does not render will-based reasons practically insignificant, as she argues elsewhere that external reasons frequently fail to favor an option in that way (Chang Reference Chang2002).Footnote ⁶

Chang’s approach to normative weighting is a somewhat natural route for a hybridist to take in attempting to handle the problematic cases that motivate objections to Source Internalism. If the hybridist can somehow use the multiple sources of normativity available to her in motivating a plausible theory of weighting, her solution to the problem may appear more viable, while at the same time reinforce the significance of a hybrid approach. So we should do our best here to generalize Chang’s approach, and consider whether it can be given motivation beyond the support that Chang herself provides.

The generalization is simple. Chang’s idea is that when external reasons favor some option Φ over any other, will-based reasons cannot alter the all-things-considered reasonableness of Φ; they cannot make Φ no longer be the most strongly favored option either by making some other option most favored, or by causing Φ to be exactly as favored as a competing option. This is what Chang means when she says that will-based reasons cannot alter the valence of reasons provided by external reasons. All that is required to make this thesis applicable to any version of Source Hybridism is to drop mention of specifically will-based reasons, and replace it with internally sourced reasons more generally:

So far as I can tell, External Priority is compatible with any way of developing Source Hybridism more fully on the first-order level. It is compatible with recognizing many external sources of normativity, as well as with an approach that grounds internal reasons in several distinct mental states or events, for example. It is simply the thesis that externally sourced reasons enjoy a kind of lexical priority over internally sourced reasons; if the externally sourced reasons tell in one direction in some case, then, no matter what they or the internally sourced reasons happen to be, the latter cannot alter what is all-things-considered best supported by reasons in that case. Of course, on virtually any fully developed first-order account, much more would have to be said about normative weighting; External Priority does not tell us, for example, how externally sourced reasons interact just among themselves in order to favor or disfavor some course of action, but we can ignore issues of that kind for our purposes here.

If correct, External Priority is of great benefit to the hybridist vis-à-vis accommodating intuitions about scenarios like Case Two. That is so because, if it is true, then, given a very natural assumption about what externally sourced reasons look like, we could never have decisive reason to cause ourselves to do something terribly disvaluable for its own sake because, even if there were internally sourced reasons to do so, it would nevertheless be the case that the externally sourced reasons to the contrary would always carry the day.Footnote ⁷ The natural assumption at work here is that, on the first-order level on the external side of things, there will always be reason to avoid agony, starvation, causing gratuitous pain to innocent people, and so on.

So, if true, External Priority does the job that hybridists require of it. But what can be said in its favor? I can think of only one line of thought that might be offered as independent support for the view. The line of thought begins with the contention that, despite some popular thoughts to the contrary, the source of internal reasons should not be limited to only a fairly narrowly restricted class of desires. That contention might be bolstered by the kinds of observations offered by David Enoch, who argues that versions of Source Internalism that involve some kind of idealization are objectionably ad hoc, and Mark Schroeder, who argues that at least some idealizations are simply not necessary to yield extensional adequacy, or at least not in many cases (see Enoch Reference Enoch2005; Schroeder Reference Schroeder2007, 84–102).Footnote ⁸ The spirit of Schroeder’s response is that resisting an idealization view simply yields the right results, an idea to which I am sympathetic; when all else is equal, an option that is favored by some agent is more choice-worthy, even if the desire in question is not idealized in the putatively proper way.

But if it can be successfully argued that even nonidealized desires can ground reasons, then some of our reasons might be grounded in desires that are fickle, unstable, and responsive to odd sorts of external influences. For example, we may find ourselves with reasons grounded in desires that have been affected by things like what television commercials we have seen in the last hour. Compared with a reason grounded in the external value of well-being, for instance, we might think that the internally sourced reason rightly deserves a less prominent place in our practical deliberations. This line of thought appears motivated first by the idea that our desires are at least sometimes given to us by the external world (to borrow another of Chang’s locutions), rather than being directly under our control. In addition, it might be argued that externally sourced reasons ought to enjoy special normative significance because of the metaphysical necessity of their source. Unlike facts about my particular desires, facts about morality, for example, that might ground externally sourced reasons are presumably metaphysically necessary.

But this route of justifying External Priority is unsatisfying. First, it is not clear that there are no stable desires. Even if some of my pro-attitudes are easily manipulated, it does not follow that all of them are, and reasons grounded in these other desires would presumably enjoy the same degree of stability had by externally sourced reasons. In addition, the appeal to metaphysical necessity is misleading. Although my individual desires are not metaphysically necessary, if desires are indeed a normative source, then presumably whatever principle that accurately captures the relationship between desires and normativity is itself necessary. Similarly, the particular circumstances that ground externally sourced reasons, if there are any, are presumably not themselves metaphysically necessary; rather, the principles that accurately describe the relationship between certain nonnormative properties and the reasons that they give rise to are what are necessary.

Perhaps most importantly, though, it is not clear that a positive justification for External Priority could be given that explains why externally sourced reasons should carry the day in every possible scenario. Suppose, for example, an agent is forced to choose between two options. Her externally sourced reasons barely favor the first over the second – the first option will provide her a nearly imperceptible additional amount of fleeting pleasure, say. However, the second option satisfies, whereas the first option does not, some important pro-attitude of the relevant sort. Why should it be impossible in this case that the internally sourced reasons outweigh the very minor difference (in terms of externally sourced reasons) between the two options? Without argumentation that is much more compelling than what has been canvassed here, it is not clear that there is a promising answer to that question. If internally sourced reasons are genuinely normatively significant, then it is prima facie plausible to think that, at least in certain circumstances, their normative significance could be weightier than that of alternatively sourced reasons. If the hybridist can adopt an approach to weighting that leaves open that possibility, then she can avoid the unintuitive consequences of the External Priority approach.

As I mentioned above, the hybridist need not approach the issue of normative weight directly through the issue of normative source. Several approaches to weighting, including those favored by Schroeder, John Horty, and normative particularists, such as Jonathan Dancy, all operate independently of any particular theory about how reasons are grounded, so far as I can see (Dancy Reference Dancy2004; Horty Reference Horty2012, 2; Schroeder Reference Schroeder2007, 123–145, Reference Schroeder2011).Footnote ⁹ A general lesson from these approaches is that the normative weight or significance of any given reason can be significantly influenced by the presence or absence of additional reasons that are relevant to the agent’s situation. As such, we need to take an approach to weighting that is, in a way, holistic; we cannot, as it were, isolate each individual reason, assign it some weight that is invariant across contexts, and then perform something resembling simple arithmetic to determine which reasons carry the day. Rather, we must understand the normative significance of a particular consideration as being in part determined by the surrounding normative context.Footnote ¹⁰

So as to avoid confusion, it is worth emphasizing here that a hybridist can endorse a thesis according to which the weight of reasons varies according to context, without endorsing the thesis that the direction in which reasons favor varies according to context. According to that latter position, and to quote Margaret Little in its formulation, ‘A consideration that in one context counts for an action, can in another count against it or be irrelevant’ (Little Reference Little2001, 34). If this thesis is understood to be a claim about all considerations, then I think that it is certainly false. However, strictly speaking, I do not think that Source Hybridists need take a stand here one way or another. What is important to note is that one can reject the thesis that Little describes, and so think that there are some reasons that invariably count for or against certain kinds of actions, but nevertheless think that the strength of those reasons is at least partially determined by details of context.

As I see things, then, the kind of holism about the strength of reasons with which I am concerned is compatible with a wide variety of more detailed approaches, including (1) a Rossian approach, according to which certain facts always count for or against certain kinds of actions, but without a codifiable weighting schema (Ross Reference Ross1930); (2) an extreme particularist approach, according to which there are no facts that always count for or against certain types of actions, and is also absent a codifiable weighting schema; and (3) some mixed approach that does recognize a codifiable weighting schema, so long as that schema has built into it complex rules (or, perhaps, one infinitely long rule) explicitly governing how reasons relate to one another differently across varying contexts. I take it that it is compatible with (1), (2), and (3) that externally sourced reasons might typically outweigh internally sourced reasons; unlike External Priority, they merely allow that externally sourced reasons need notdo so across all cases.

Return now to the thought with which we are primarily concerned, that having to do with the observation that reasons are in at least one sense to be understood holistically: that the weight of a given reason is not fixed outside of context. How does this observation help the Source Hybridist? It does so because it helps to make clear the idea that any plausible approach to normative weight, when combined with hybridism, is going to yield the correct result in cases like Case Two. This is because the very same intuitions that show these cases to be problematic for Source Internalism will also show them to be problematic for any theory of weighting that does not give greater weight to the reasons to avoid causing oneself to be in agony for its own sake (or whatever).

What makes these cases forceful as objections to internalism is that we at least implicitly endorse the minimal point made above about the holism of reasons, and recognize that even if the agent in Case Two has some reason to cause herself agony for its own sake, given the surrounding normative context, it cannot be a decisive reason. In contexts like Case Two, the reasons for the agents involved not to carry through with their goals are weightier than their reasons to the contrary. Any approach to weighting that fails to respect that restriction will be rejected for the same reason that Source Internalism should be rejected. Notice that the restriction is silent on the issue of normative grounding; it does not insist that one set of reasons is weightier because of its source, or that reasons grounded in a certain way will always be weightier than reasons grounded in another. Again, it is left open here whether externally sourced reasons will sometimes be outweighed by internally sourced reasons. What is important for the Source Hybridist’s position is that that will not happen in cases like Case Two, a point that can be established simply by reflection on the cases, and without adverting to any specialized theory of weighting that appeals to normative grounding.Footnote ¹¹

Developing an account of grounding that is not beholden to a particular account of weighting and, vice versa, is not only beneficial to hybridists, but is also independently motivated. What the grounding accounts provide us are metaphysical theories about how it is that normativity arrives on the scene; they tell us what gives rise to normativity. But weighting accounts provide us with normative theories about how, once they are on the scene, practical reasons interact with each other to justify our behavior. Of course there are very difficult questions about how exactly that works, and which kinds of considerations will be the strongest across different contexts, but we can endeavor to address those issues once we know what the considerations are, and it is another matter entirely how the considerations came to have their normative force.

Consider a rough physical analogy, one that I hope is illustrative, but that I do not mean to carry argumentative force. Imagine that we are scientists concerned with figuring out how electrically charged particles interact with each other. For our purposes, it is irrelevant what it is in virtue of which some particles come to be charged whereas others do not; all we want to know is what happens once they are so charged and they start banging into one another. Similarly for reasons and their weight, I am suggesting. It is irrelevant what it is in virtue of which certain facts come to be normatively charged whereas others do not; all we want to know is what happens once they are so charged and start banging into one another.

3. Source Hybridism and practical deliberation

Unlike its competitors, Source Hybridism is dis-unified; hybridists maintain that there is a single kind of thing – practical normative force – that can be grounded in two distinct, apparently unrelated sources. To put it crudely, there is something messy, or untidy about this picture. Although the possibility of a hybrid view is not frequently discussed, this feature of Source Hybridism has not gone unnoticed. T. M. Scanlon worries that such a hybrid approach to grounding might give rise to a ‘puzzling duality’ about reasons: ‘If the normative status of counting in favour of acting a certain way is something that certain considerations can just have, how can it also be something that we can confer on certain considerations by our choice of ends?’ (Scanlon Reference Scanlon, Jay Wallace, Pettit, Scheffler and Smith2004, 231). Although Scanlon is thinking in particular about our ends here, it is easy to see how his puzzlement might arise about the grounding of normativity in any pro-attitude at all, while simultaneously coupled with external grounds.

As noted by Niko Kolodny, Scanlon’s expression of puzzlement should be made more precise if it is meant to indicate the presence of a serious objection to hybridism (Kolodny Reference Kolodny, Jay Wallace, Kumar and Freeman2011, 46–48). Presumably, Scanlon is not concerned merely about the number of sources on a hybrid view, as Scanlon himself recognizes a wide variety of facts or properties in which normativity could be grounded, as do many externalists. Rather, his concern with hybridism is more likely to do with the duality of the kinds of sources in play.

Kolodny is sympathetic to Scanlon’s worry about hybridism’s puzzling duality, but develops it in what I think is a somewhat surprising way. I think it is worth looking at the key passage in its entirety. Kolodny writes that externalism

Kolodny’s complaint is that Source Hybridism would commit us to some implausible result concerning practical deliberation. The passage above suggests at least two interpretations of what that result is supposed to be.Footnote ¹²

On the first reading of Kolodny’s objection, he is concerned with practical deliberation from the point of view of the deliberating agent. According to this reading, Kolodny sees deliberation as explicitly involving an attempt on the part of the reasoner to identify the grounds of her reasons. What is supposed to be objectionable about hybridism is that it would require an agent to do this is in a dis-unified way, by identifying both internal and external grounds.

What is most problematic about this version of Kolodny’s complaint is that no theory of the grounds of normativity should be understood as requiring an accompanying theory of practical deliberation that explicitly requires deliberating agents to think about metaphysics when deciding what to do. On this reading, the objection makes a mistake not only about hybridism, but also about normative grounding generally: the mistake of thinking that deliberation needs to involve thinking about how one’s reasons are grounded. From the first-person point of view, what matters in deciding what to do via deliberation is identifying one’s reasons and then attempting to determine their relative weight; it is irrelevant in virtue of what those things have their status as reasons.

On another interpretation of his objection, Kolodny is concerned not with a unified process of deliberation from an agent’s point of view, but rather with a unified account of practical deliberation from the point of view of normative theorizers. On this reading, what is supposed to be objectionable about hybridism is that it can countenance no unified account of what practical deliberation is or consists in. That is because it appears that different kinds of steps would be required in deliberating over reasons that are externally sourced than would be required in deliberating over reasons that are internally sourced. For example, it might be that, if a certain kind of internalist view were true, then practical deliberation would be nothing other than means-end reasoning: deciding what to do just is determining which action would best promote my ends, or desires, or whatever. But deliberating over externally sourced reasons, reasons grounded in value, perhaps, is not like this; if there are value-based reasons in play, merely determining the best instrumental course of action may not lead me to act in accordance with my reasons. So, if both sorts of reasons are in play, what kind of deliberative stance could one assume in order to come to act? Since there is no single answer that hybridists could give in response to this challenge, the objection goes, the hybridist cannot supply us with a unified account of what it is to deliberate practically.

But it seems to me that there is no special problem for Source Hybridism here, either. Hybridists can begin with the least theory-laden understanding of practical deliberation that everyone ought to begin with: that practical deliberation is a mental activity through which one attempts to determine what to do by reflecting on considerations that are normatively relevant to that issue. Arpaly and Schroeder (Reference Arpaly and Schroeder2012) helpfully emphasize that that complicated mental activity will be further comprised of more discrete mental activities. For example, in deciding what to do, I might attempt to remember salient details, to predict the outcomes of potential courses of action, to enumerate what options are available to me, and so on. As such, all of these discrete activities might be thought of as parts or potential parts of practical deliberation, or practical reasoning.

That practical deliberation should already be understood as multifaceted in this way helps to show that there is nothing objectionable vis-à-vis deliberation about the fact that hybridists see normativity as grounded in multiple sources. Even if it is true that properly deliberating agents should, by hybridists' lights, at least sometimes carry out several mental activities that are different in kind when deciding what to do, this is not objectionable. Indeed, it is a familiar aspect of normal deliberation. In particular, what I think is especially familiar here is the idea that distinct kinds of mental actions might be needed in identifying one’s reasons. There is nothing particularly surprising about the fact that coming to learn which facts are normatively relevant might involve different kinds of thought processes, and so there should be nothing particularly objectionable in suggesting that agents might sometimes do best by considering both facts about value and facts about what actions would best promote one’s desires, for example, in coming to determine what reasons bear on their question of what to do.

Even the idea that hybridists will require different modes of identifying our normative reasons might be called into question, though. Or, at least, the hybridist might contend that different modes will be required less frequently than the objector is imagining. Suppose that, as some externalists think, some normatively important considerations simply strike us as so, that certain facts seem to us to count in favor or against certain actions. On this kind of view, we come to have normative knowledge through the positive epistemic status of intellectual seemings.Footnote ¹³ Now suppose that, as Scanlon famously contends, our pro-attitudes also cause certain considerations or features to strike as normatively relevant in the same way. These views are not inconsistent with each other, and, if both were true, then many externally sourced and internally sourced reasons would be identifiable as reasons in precisely the same way, because they will seem to us as being normatively relevant, and intellectual seemings of this kind carry positive epistemic status.

Perhaps, though, the worry about practical deliberation is not meant to be about identifying our differently sourced reasons, but rather in determining what to do on the basis of them, once they are identified. In this guise, the worry may just be the weighting concern with which we began, in slightly different packaging, and should be addressed in the same way, namely by admitting that there is no independent fact of the matter about which considerations will be weightiest across all situations, but that the normative significance of differently sourced reasons will be fixed holistically, as the normative significance of all reasons are.

Indeed, perhaps hybridists ought to argue that, even from the first-person point of view, considerations regarding deliberation over differently sourced reasons actually lend support to their position. First, it really does seem that we sometimes find ourselves torn between considerations grounded in our pro-attitudes and considerations grounded in an entirely different sort of thing, like objective value, for example. Part of what makes decision-making in these cases hard is that the reasons really do appear, in at least some sense, different in kind. On the other hand, though, agents are somehow capable of integrating these considerations when a choice must be made. Although it is sometimes hard to decide what to do when reasons of these kinds compete, we really do make decisions under those circumstances, and are at least often confident that we have done so in the way that is actually normatively supported by the totality of those reasons. And it is that last point that is most relevant here. It is not so important that agents in fact do adjudicate among dissimilar kinds of considerations; what is significant is that there is a wide range of cases in which there is a clearly correct first-order judgment about how to adjudicate. That I prefer for my shoes not to get muddy, for example, is normatively much less significant than the fact that I can save an innocent child by wading in to the pond. We are already justified on first-order grounds in thinking that it is possible to correctly adjudicate among seemingly dissimilar reasons; Source Hybridism adds to this a metaphysical story about the grounds of those reasons, one that can possibly explain why there seems to be a dissimilarity at all.

The foregoing comments also suggest a response to a possible modification of the first interpretation of Kolodny’s objection with which we began. Recall that on that first interpretation, what was supposed to be objectionable about hybridism is that it requires agents to identify two distinct metaphysical grounds of their reasons while deliberating. I said that this version of the objection fails because it is a mistake to think that a good theory of practical deliberation requires agents to think about the grounding relation while undertaking practical reasoning. In response to my defense, Kolodny’s objection might now be redeployed on the grounds that, although it is correct that agents are not required to consider the metaphysical grounds of their reasons, a viable theory of normative grounding must at least allow that agents could do so in a way that makes sense, and it is this constraint that Source Hybridism violates.Footnote ¹⁴

At this point, the hybridist can either reject the proposed constraint, or argue that hybridism does not violate it. The first option seems untenable to me. It really does seem right that, if I am so inclined – if I am a philosopher who works on meta-normativity, say! – it should at least be an open possibility to me that I consider not only the content of my reasons and their weight when deliberating, but also the grounds that provide their normative force. So the constraint seems like a plausible one. But I do not think that Source Hybridism violates even this. As I suggested above, I think that it is simply a familiar feature of normative conflict that we are sometimes pulled in opposing directions by reasons that are very dissimilar from one another. Rather than introducing an objectionable deliberative hurdle, Source Hybridism illuminates more clearly and explains the phenomenon of our difficulty adjudicating between our reasons in situations such as these. Again, this is not to say that the reasons are somehow irreconcilable with each other; as I said above, we do manage to see our way to a conclusion about what to do, and we should be confident that we are at least frequently getting that right. But the felt dissimilarity between choice situations of this kind, and situations in which only one kind of reason is at play, is easily accounted for by Source Hybridism: the conflicting reasons derive their normative weight from dissimilar sources.

One of the lessons that I think should be learned from this and Section 2 is that we hybridists should be careful to identify exactly what our theory of grounding commits us to. If it commits us to relatively little, as I have been suggesting here, then we may be in a good position to resist objections that charge hybridism with unsavory implications. And, again, I think it should be relatively unsurprising on the face of it that one’s preferred theory of grounding is compatible with a wide variety of theories regarding weight, practical deliberation, and so on.

4. Source Hybridism and metaphysical simplicity

The objection considered in Section 3 began with the observation that Source Hybridism is messy, insofar as it posits disparate grounds for practical normativity. Rather than attempting to show that feature of the view to be objectionable because of its relationship to practical deliberation, or some other related aspect of normativity, one might instead insist that hybridism is simply objectionably unparsimonious as such, and so ought to be rejected on purely metaphysical grounds.

The objection is perhaps best put this way: if internal and external sources could both ground the same phenomenon, there must be some feature that the sources have in common that could explain their giving rise to the very same thing. But there is no such common feature in the case of the supposed internal and external sources of normativity, so they cannot both ground that phenomenon.Footnote ¹⁵ And here, the idea seems to be that we ought not countenance disparate sources for the same phenomenon because doing so would be metaphysically unparsimonious; we would thereby postulate instantiations of the grounding-relation that are odd or superfluous.

I want to admit that there is probably something metaphysically unsatisfying about Source Hybridism. Other things being equal, it would be best to offer a completely unified account of the grounding of practical normativity. But, as I said above in the introduction – but admittedly have not argued for in this paper – other things are not equal. The alternatives to hybridism are beset by serious problems that hybridism can solve. If the problems are serious enough, then metaphysical complexity or parsimony just does not count for very much, I think. We ought to be aiming for the most metaphysically streamlined theory that can do the work that we need it to do, but increases in metaphysical complexity should always be countenanced if they are indeed necessary for that work to be done.

I realize that the foregoing line of reasoning may be unsatisfying, given that I have offered only a promissory note for arguments that are required to buttress it. So I will try now to say a bit more to make the metaphysical complexity of hybridism seem less objectionable than it might otherwise.

The objector claims that a unified phenomenon could not be grounded in disparate sources unless those sources exhibited some important commonality that explains how they could give rise to the same object, fact, state of affairs, or whatever. The hybridist could comfortably reject the objection if it could be shown that dis-unified grounding is actually more common than the objector assumes.

It seems to me that she may be able to do so if other evaluative phenomena are construed realistically. For example, if there are stance-independent truths about beauty, they might be grounded in disparate sources. The Rothko is beautiful in virtue of its stark presentation of color; the Bach cello suite is beautiful in virtue of its melodic structure. Canterbury Cathedral is beautiful in virtue of its architecture; Paradise Lost is beautiful in virtue of its linguistic imagery and meter. All of these items share the common feature of being beautiful, despite the fact that that feature is grounded in disparate sources that do not seem to share even any very broad structural similarities.

Moral features may work this way, as well, depending on the correct first-order moral theory. Unless one is a monist about bedrock moral considerations, one should think that moral features like permissibility, impermissibility, obligatoriness, supererogation, and so on can be given rise to by disparate sources. That action is morally obligatory in virtue of the overall good it produces; this other action in virtue of the just distribution it results in. If this is right, then grounding in the evaluative realm is not so infrequently dis-unified in the way that the hybridist argues that normative grounding is.Footnote ¹⁶

Perhaps the most compelling case for the hybridist to advert to in this context is that of objective list theories of well-being. Objective list theories all posit multiple intrinsic relational goods, such as pleasure, friendship, knowledge, autonomy, and so on. What the items on the list have in common is just that their presence in an individual’s life contributes to her well-being; according to these views, there need not be any further similarity in virtue of which they share this relationship to well-being (see, e.g., Scanlon Reference Scanlon2009).Footnote ¹⁷ So objective list theories of well-being recognize disparate grounds for well-being, and objective list theories are among the most popular, if not the most popular, theories of well-being.Footnote ¹⁸ At the very least this shows that a large number of philosophers are already willing to at least sometimes countenance the supposedly troubling feature of Source Hybridism under investigation here: that it posits disparate grounds for the same phenomenon.

Let me elaborate just a bit further on my appeal to objective list theories, to better clarify the similarity that I see between such theories and Source Hybridism. Recall that I began in the introduction by drawing our attention to relation R – the reasons relation, which relates facts, agents, and actions. When some fact f stands in that relation, it is a pro tanto reason for some agent to Φ. For example, that there are still workers in the building is a reason for Joel to delay the demolition. Now consider what I think is a similarly structured relation from the theory of well-being: the being extrinsically good for relation, E. We can formulate that relation in this way: E(e, A), where e is the argument place for some event, and A is the argument place for some agent. When some event e stands in that relation, it makes a pro tanto contribution to some agent’s well-being. For example, going to the baseball game is extrinsically good for Steve.

Now, suppose that we request from the objective list theorist an explanation for what metaphysically makes it the case that e is extrinsically good for A, for any instantiation of the E relation. The objective list theorist will reply that there is no univocal answer; sometimes, what makes e extrinsically good for A is some fact about e’s relationship to friendship, and sometimes to pleasure, and sometimes to knowledge, and sometimes to desire-satisfaction, or whatever. The Source Hybridist responds similarly to a request for an explanation for what metaphysically makes it the case that f is a reason for A to Φ, for any instantiation of the R relation. Sometimes it is some fact about f’s relationship to value, and sometimes it is some fact about f’s relationship to the promotion of A’s pro-attitudes.

Of course, the objective list theorist may reply to our request by saying that what makes e extrinsically good for A is just that e is appropriately related to some intrinsic good; this may make it appear that a univocal answer is available to such a theorist. But a parallel response is also available to the hybridist: what makes f a reason for A to Φ is f’s relationship to some ultimate source of normativity. But just as the hybridist denies that there is a deep unifying feature of those sources of normativity, so too does the objective list theorist deny that there is a deep unifying feature of the intrinsic goods. Since that is the putatively objectionable feature under consideration, it seems to me that Source Hybridism is at least no more objectionable than objective list theories of well-being.Footnote ¹⁹

5. Conclusion

I have argued here that three objections to Source Hybridism are not sufficiently strong to warrant rejecting the theory out of hand. The view does not have problematic implications either for normative weighting or for practical deliberation, because it is compatible with independently plausible approaches to both topics. Furthermore, even if the view is metaphysically unparsimonious to some degree relative to its competitors, that feature of the view cannot by itself warrant rejecting it. Indeed, Source Hybridism may be no more metaphysically objectionable than many other philosophical positions that enjoy wide support. Of course, as I have tried to directly acknowledge above, even if everything that I have argued for here is correct, much more work needs to be done to completely vindicate Source Hybridism.

For example, I have said nothing here about how Source Hybridism determines which reasons derive their normative force externally, and which internally.Footnote ²⁰ Strictly speaking, I think that it is appropriate for the purely meta-normative theory to remain silent on this issue, though I recognize that any interestingly developed version of the view must eventually say quite a lot about this. The way to address such an issue, though, is to engage in first-order normative theorizing, in combination with meta-normative theorizing, to determine what the particular external and internal sources of normativity in fact are; given the enormity of such a project, I will not undertake any substantive steps toward it here. Chang’s work provides us a ready model, though, of what this kind of theorizing may look like; she argues not only that Source Hybridism is true, but also that the internal source of normativity is the will. Supplying that detail on the internal side of the meta-theory enables the view to render at least some first-order judgments about what particular reasons agents have on particular occasions, and from whence their normativity is derived. To fully complete the picture, a theory of value (or of brute normativity) would need to be supplied on the external side of the theory; when in place, it would play a similar role as does Chang’s appeal to the normativity of willings.

I take no stand here on the truth, or even the plausibility, of these details of Chang’s preferred version of Source Hybridism. I offer her picture to the reader only as a lesson in what lower-order hybridist theorizing might look like, given that I have been interested thus far only in Source Hybridism as such, and have done little, if anything, to demonstrate that the theory can be filled out more substantively. As I said above, though, engaging in that project more fully is work for another time.Footnote ²¹