I. Pro toto reasons

Broome identifies two types of normative reasons: pro toto and pro tanto. A pro toto reason for N (an agent) to F (a response) is defined by Broome as “an explanation of why you ought to F”.⁴ In this oftenquoted passage Broome claims that his definition captures the normativity of pro toto reasons:

We slide from ‘X is the reason why you ought to F’ to ‘X is the reason for you to F’, meaning exactly the same thing. The ‘reason why’ (meaning explanation) bumps into the normative ‘ought’, yielding a normative sense of ‘a reason’ that combines the meaning of both.⁵

To exemplify: say you promised to take your friend to the airport today, but now you find yourself rather reluctant to make that long drive. I tell you that you ought to take your friend to the airport nevertheless, and you ask why. “Because you promised to”, is the answer. So the reason why you ought to take your friend to the airport is your having promised to. Here we are giving an explanation of an ought. Broome claims that in the case of normative explanations like this one, the consideration following the ‘because’ plays a dual role: it is both an explainer (which by itself is non-normative) and a normative reason. We can see this by noting that ‘because you promised to’ not only explains why you ought to take your friend to the airport but is, at the same time, a reason (or even the reason) for you to take your friend to the airport. So if Broome’s suggestion is right, we can understand what a normative pro toto reason is in terms of two more basic notions -those of ought and explanation.

Let’s turn now to Dancy’s worries. The first worry is whether we can account for a normative notion such as a pro toto reason in terms of a non-normative notion such as explanation: “I do not see how he [Broome] can both say that the notion of an explanation why you ought to F is not normative, that the notion of the (pro toto) reason for you to F is normative, and that these are the same notion”.⁶

In a moment I will suggest that Broome has available a response to this worry, a response based precisely on the distinction Dancy introduces between the favoring relation and the right-making relation. Before that, however, it’s important to note that the worry is not -or at least not in the first instance- about the extensional adequacy of Broome’s definition. That is, Dancy might concede that anything that explains why you ought to F is at the same time a pro toto reason for you to F.⁷ Rather, the problem concerns whether a normative reason can be equated with (or understood in terms of) an explanatory one. Dancy⁸ claims that it can’t, because the favoring relation -which characterizes considerations that are normative reasons- is different from the right-making relation -which characterizes considerations that figure in normative explanations.⁹ Let’s illustrate this distinction with the previous example:

Favoring relation. Your having promised to take your friend to the airport favors taking your friend to the airport.

Right-making relation. Your having promised to take your friend to the airport makes it the case that taking your friend to the airport is right.¹⁰,¹¹

It’s clear from this example that, as Dancy points out, “even though one and the same feature is capable of standing on the left-hand side of both of these relations, the right-hand sides are very different”.¹² They are very different because in the favoring relation the righthand side concerns a way of acting, whereas in the right-making relation the right-hand side concerns a property of the act itself. Dancy goes on to suggest (correctly, in my view) that the favoring relation is normative and the right-making relation is metaphysical. Presumably, this means that in the former case what is at stake is the justification of the action whereas in the latter case what is at stake is the nature of that same action. Thus, Dancy’s deeper objection to Broome’s definition of a pro toto reason isn’t merely that explanation is a non-normative notion, but rather that Broome wrongly assumes that the favoring relation and the right-making relation are at bottom the same, or at least that the former can be understood in terms of the latter.

What would Broome say in response? I think he would say the following: perhaps the two relations are distinct, but we gain a deeper understanding of the favoring relation by seeing it in light of the rightmaking -or, as Broome would say, ought-making- relation. Here’s a relevant passage from Broome that suggests this line of response:

The ‘counts in favour’ formula is very commonly offered as a way of explaining what a reason is [in a footnote Broome mentions Dancy as an example]. I agree that a reason is a consideration that counts in favour of something. But my definition goes further and specifies what is the relevant sort of counting in favour.¹³

The relevant sort of counting in favor is, of course, counting in favor of some action being the action one ought to do.¹⁴ So according to Broome we gain a deeper understanding of normative reasons and of the favoring relation by noting what it is that reasons favor: reasons favor some action being the action one ought to do (or the action being the right one). In the case of pro toto reasons, a consideration favors an action because that same consideration has made it the case that one ought to perform it.

Here Dancy would object again, though. He claims that

What is favoured is acting (or at least responding) in a certain way, not the rightness of so acting, which cannot be favoured at all. (It cannot be favoured because the only things that can be favoured are responses, and the rightness of an act is not a response at all).¹⁵

So Dancy denies what Broome affirms, namely, that the counting-in-favor relation can be defined in terms of contributing to the rightness (or oughtness) of the action. It’s important to be clear that Dancy isn’t opposed to the conception of reasons as right-makers; rather, what he is objecting to is Broome’s attempt to define normative reasons (and a fortiori the favoring relation) in terms of the right-making relation, as if the favoring relation could be understood in terms of the right-making one.

As a response to Dancy, let me sketch a proposal about how the right-making relation can account for the favoring relation that is in the spirit of Broome’s account. Let’s start by asking: Why does a consideration favor an action? A plausible answer is: because the consideration makes the action the right thing to do or the action one ought to perform.¹⁶ So the two relations are related as follows. The rightmaking relation has priority: when a consideration gives the action a certain property -rightness or oughtness- it also, and because of that, favors the performance of the action.¹⁷,¹⁸

Returning to our example: why does your having promised to take your friend to the airport favor your doing so? The proposed answer is: because your having promised to do so gave the action the property of rightness or oughtness. And once your promising gave the action this property, it follows that the fact that you promised to do it favors your doing it. So even though Broome might be wrong in suggesting (if he indeed does so) that the explanatory and rightmaking relation on the one hand, and the favoring relation on the other, are the same, this doesn’t imply that these relations are so different from each other that -as Dancy thinks- it isn’t possible to understand the one in terms of the other.¹⁹ As I just suggested, this is possible: in the case of pro toto reasons, we in fact gain a deeper understanding of the favoring relation once we realize that its obtaining in a specific case is a consequence of the right-making relation being already in place. Again, the deeper understanding comes to this: a consideration favors an action because that same consideration has made the action the right thing to do (This obviously just applies to considerations that amount to pro toto reasons. I address pro tanto reasons below).

Now, what about Dancy’s point that the two relations -favoring and right-making- have different statuses, one being normative and the other metaphysical? This indeed shows that, as Dancy says, “it [is] hard to think that one of them might really be the other in disguise”.²⁰ But, as I just explained, the best way to understand the link between the two relations isn’t to claim that they are at bottom the same, but rather to note that one of them (the favoring relation) derives from the other (the rightor ought-making relation). Thus, according to this proposal, a metaphysical relation gives rise to a normative one. Is this in principle objectionable?

I don’t think so. We could say that normative reasons ‘track’ rightness or oughtness: normative reasons justify (or attempt to justify) responses, and justifying a response amounts to showing that it’s the right one or the one you ought to exhibit. So, to take the case of actions, when an action in fact has the property of rightness we can say that there is a pro toto reason for performing it; and there is this reason precisely because the action is right. Thus, the normative relation between the consideration that is the pro toto reason and the action it favors derives from the metaphysical relation between that same consideration and the property of rightness or oughtness (a relation that in turns grounds an explanation of an ought).

Let me be clear that I’m not fully endorsing this response as my own; all I have done is to suggest that the foregoing is a plausible rebuttal to Dancy’s objection that the definition of pro toto reasons offered by Broome fails because we can distinguish the favoring relation from the right-making relation.

II. Pro tanto reasons

Assume for the sake of argument that what I just sketched is a promising response (at least as pro toto reasons are concerned) to Dancy’s objection. What I want to consider now is whether this same response can make sense of the other kind of normative reasons Broome acknowledges, namely pro tanto reasons. My verdict will be negative: I will argue that if Broome’s understanding of the normativity of pro toto reasons is along the lines I suggested above, then his account can’t make sense of the fact that pro tanto reasons are supposed to be normative entities in their own right. More precisely, it can’t make sense of the fact that pro tanto reasons on the losing side of what Broome calls a ‘normative weighing explanation’ are nevertheless normative. This suggests that Broome’s account of pro toto reasons is in direct conflict with his account of pro tanto reasons.

Broome starts his discussion of pro tanto reasons in this way:

We often say there is a reason for you to F, when it is not the case that you ought to F. In these cases, the reason evidently does not explain the fact that you ought to F, since there is no such fact. It is therefore not a pro toto reason. Reasons of this sort are often called ‘pro tanto reasons’.²¹

So pro tanto reasons don’t explain oughts but, according to Broome, it doesn’t follow that they cannot be understood in terms of ought and explanation at all. They can, but the account needed here is different from the account of pro toto reasons. As we saw above, pro toto reasons are considerations that explain outright why one ought to F, whereas pro tanto reasons are considerations that play a certain role in what Broome calls ‘normative weighing explanations’. The role in question is that of a consideration adding its ‘weight’ to the weight of other considerations so that the normative balance tips one way rather than the other. The winning side is the weightier one, and one ought to do the action (or, more generally, to exhibit the response) that corresponds to it. Thus, Broome defines a pro tanto reason as “whatever plays this role in a normative weighing explanation”.²² This is Broome’s functional definition of a pro tanto reason.²³

Now let’s ask: can Broome’s functional definition capture the normativity of pro tanto reasons? Dancy claims that it can’t, because “[an] explanation is not itself normative, even if what is explained is normative” and adds that Broome “has no independent method of establishing the normativity of the explainers”.²⁴ However, as we saw in the previous section, the reply I offered on Broome’s behalf constitutes a denial of Dancy’s claim, at least as pro toto reasons are concerned: Broome can in fact establish the normativity of the explainers (that is, of the considerations that are pro toto reasons) by appealing to what grounds the relation of normative explanation, namely, the right-making relation. If a consideration explains why you ought to F, this is because the consideration makes it the case that you ought to F. And, as I have argued, Broome can say that, once a consideration gives the property of rightness or oughtness to F, that same consideration is automatically a normative reason for you to F.

But if this is Broome’s response to Dancy’s worry about the normativity of pro toto reasons (and I do think that it’s the best response he has), it ends up working against Broome himself, because it leaves him unable to explain the normativity of pro tanto reasons. Here’s why: in order to deal with Dancy’s worry about the relation of explanation being non-normative, Broome has to appeal to the right-making relation as what endows the explainers with normativity. But there is no such endowment in the case of pro tanto reasons, because a pro tanto reason doesn’t explain outright why one ought to F (otherwise, it would be a pro toto reason), and so a pro tanto reason doesn’t by itself make it the case that F is the action one ought to do. Thus, in the case of pro tanto reasons there is no ‘transmission’ of normativity from ought to explainers. So we are bound to conclude that playing the role Broome describes in a normative weighing explanation isn’t enough for establishing the normativity of pro tanto reasons, given that the right-making element (which is what endowed pro toto reasons with normativity) is absent.

Consider a possible rejoinder open to Broome. Suppose he held fast to the idea that the rightor ought-making relation is what grounds the normativity of reasons. Then he could say that the normativity of every pro tanto reason comes from the fact that, for every such reason, there is a possible normative explanation in which this reason could explain an ought. Broome could suggest that this is why pro tanto reasons are normative after all: they are capable of explaining oughts, even though they don’t do so in every occasion. In this way, the rightor ought-making relation would again enjoy pride of place in Broome’s account.

But this putative solution is problematic. For if this is how Broome is going to explain the normativity of pro tanto reasons, then he has to give up altogether the notion of a pro tanto reason -i.e., the notion of a reason that is normative even though on this occasion doesn’t explain an ought.²⁵ More precisely, he has to give up the idea, to which he is explicitly committed, that pro tanto reasons on the losing side of a normative weighing explanation are nevertheless normative.²⁶ Thus, Broome’s failure to handle at the same time pro toto and pro tanto reasons suggests that his account suffers from a critical internal tension.

One way to resolve this tension would be to just bite the bullet and accept the conclusion that considerations are normative reasons only when they actually explain an ought. In other words, Broome could get rid of the notion of pro tanto reasons and stick with pro toto ones. While this move may be revisionist, it isn’t prima facie senseless. One way to defend it is this. One could argue that considerations that on a certain occasion seem to be pro tanto reasons in favor of Fing are revealed to lack normative import or weight once one realizes that one ought not to F. That is, what initially seemed to be a pro tanto reason for Fing isn’t, on this occasion, a reason at all, given that Fing isn’t what one ought to do. On this proposal, then, real normative reasons are only those considerations that actually explain an ought, not ones that could have explained it if only others considerations weren’t in place.

For example, the fact that I’d love to eat this piece of cake seems to be a pro tanto reason to eat it. Suppose, however, that it’s actually the case that I ought not to eat it since it belongs to you. Thus, the fact that I’d love to eat it isn’t, on this occasion, a reason at all. It isn’t that the consideration is outweighed by the fact that the cake belongs to you; rather, it lacks normative weight altogether (on this occasion).

A consequence of this way of fixing the tension I identified in Broome’s account is, of course, that Broome would have to ditch the idea of normative weighing explanations of oughts. This move isn’t without precedent, however. Christine Korsgaard²⁷ advances a model of practical deliberation similar to the one I just described precisely as a way of rejecting what she calls ‘the weighing model’. In Korsgaard’s view, practical deliberation doesn’t consist in marshaling considerations for and against a certain course of action and then weighing them in order to determine what one ought to do; rather, it consists in testing one’s ‘maxim’, which incorporates the considerations that apparently favor the proposed action, against the universalizability test of the categorical imperative. Korsgaard claims that if a maxim fails this test, then every considerations that seemed to favor the proposed action “is not merely outweighed -rather, it is not a reason at all”.²⁸

So at least there is precedent if Broome wanted to argue that only considerations that actually explain an ought -that is, only pro toto reasons- are truly normative reasons and thus rejected the notion of normative weighing explanations. But I take it that he would be reluctant to follow this route. For one, he briefly considers and tentatively rejects the suggestion that moral theories that give pride of place to “rigid deontic rules” show that some deontic facts lack normative weighing explanations.²⁹ Also, and more importantly, Broome is firmly committed to the idea that what he calls ‘the central ought’ is determined by the interaction of normative requirements stemming from morality, prudence, self-interest, etc., each of which constitutes (at least sometimes) normative pro tanto reasons.³⁰ So he can’t reject the notion of a pro tanto reason altogether. For, if (apparent) pro tanto reasons weren’t normative in themselves, then how could it be that a number of them determined what one ought to do? In other words, how could an ought resulting from a weighing explanation be normative if its components -one or more pro tanto reasons- aren’t normative themselves?³¹

Therefore, Broome seems to be at the same time unable to shed the notion of a pro tanto reason and unable to give a satisfactory account of it. I don’t see how Broome could avoid this tension without adopting a very different model of practical deliberation or without embracing a primitivist view about reasons.

III. Conclusion

I will conclude by briefly asking whether Broome’s larger project in Rationality through Reasoning is challenged if we accept my arguments and thus accept that his characterization of reasons is inadequate. In the introduction to his book, Broome tells us that its overarching project is to answer ‘the motivation question’. What Broome attempts to find out is how one can motivate oneself to intend something when the starting point is a belief about what one ought to do. In a nutshell, Broome’s answer is that one can come to form the intention to F when one believes one ought to F by complying with a requirement of rationality he calls ‘Enkrasia’ which, roughly, requires one to either form the corresponding intention or abandon the cited belief. And although it often happens that one automatically complies with this requirement, one can actively come to satisfy it by reasoning one’s way to the required intention starting only from a belief about what one ought to do and a belief about the thing in question being up to one.³²

Is Broome’s answer to the motivation question compromised in some way by the arguments I have offered here? I don’t think so. Even if I’m correct that pro tanto reasons can’t be defined in terms of normative weighing explanations, that seems to have nothing to do with Enkrasia being a genuine requirement of rationality or with enkratic reasoning being genuine and correct reasoning.

One might think that my arguments are more relevant for the issue of how normative requirements coming from different normative sources determine what one ought to do on each occasion. But even here the shortcomings of Broome’s account of reasons produce limited damage. As I hinted above, in order to preserve his picture about how normative requirements determine oughts, Broome definitely can’t do without pro tanto reasons. So the right thing for him to do is, I suggest, simply to postulate (as he does) that normative sources issue normative pro tanto reasons but abandon the project of trying to define reasons in terms of weighing explanations of oughts. He can still say that reasons explain oughts, but he has to give up the ambition of capturing the normativity of pro tanto reasons in terms of explanation. The suggestion is that Broome can embrace a sort of primitivism about reasons without sacrificing the main themes of his book.

This doesn’t mean, however, that the arguments I have given here are irrelevant. For if one important attempt to avoid primitivism about reasons is shown to be unworkable, that provides indirect evidence in favor of the primitivist position.³³