Abstract
In recent years, the notion of a reason has come to occupy a central place in both metaethics and normative theory more broadly. Indeed, many philosophers have come to view reasons as providing the basis of normativity itself . The common conception is that reasons are facts that count in favor of some act or attitude. More recently, philosophers have begun to appreciate a distinction between objective and subjective reasons, where (roughly) objective reasons are determined by the facts, while subjective reasons are determined by one's beliefs. My goal in this paper is to offer a plausible theory of subjective reasons. Although much attention has been focused on theories of objective reasons, very little has been offered in the literature regarding what sort of account of subjective reasons we should adopt; and what has been offered is rather perfunctory, and requires filling-out. Taking what has been said thus far as a starting point, I will consider several putative theories of subjective reasons, offering objections and amendments along the way, will settle on what I take to be a highly plausible account, and will defend that account against objections.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
The reasons I have in mind are normative or justifying reasons, as opposed to explanatory reasons (or reasons why) and reasons for which (or the agent’s reason). The term ‘motivating reasons’ is sometimes used to refer to explanatory reasons (when the relevant explanandum is an action), and sometimes to refer to reasons for which; so to be clear, the kind of reasons I am interested in here are not motivating reasons (per se), of either kind.
It might be suggested that subjective reasons do not depend solely on doxastic attitudes, but can derive from motivational attitudes like desire and certain emotions. Furthermore, one might worry that the claim that subjective reasons cannot be based on motivational attitudes begs the question against Humean theories of normative reasons, according to which a normative reason for one to ϕ requires that one have some desire that would be served by ϕ-ing. But these concerns are unwarranted. Subjective reasons have been posited precisely to account for cases like SNAKE. There is no analog of that case that would prompt us to posit subjective reasons that depend upon motivational attitudes rather than doxastic attitudes. Furthermore, Humean theories on which normative reasons require the presence of a desire are theories of objective reasons. Schroeder (2007, 2008, 2009a), for example, endorses a Humean theory of normative reasons, but believes that subjective reasons depend solely upon one’s beliefs. This is not an inconsistent position, simply because Humean theories of normative reasons are not theories of subjective reasons.
It might be claimed that there is a further distinction among subjective reasons: subjective reasons that are determined by beliefs held on good grounds, and subjective reasons that are determined by beliefs per se, whether or not they are held on good grounds. Consider, for example, a modified version of SNAKE in which I lack good grounds for believing that running away will save my life. In this case, there seems to be a sense in which I do not have a reason to run away, such that if I were to have good grounds for my belief that running away will save my life, I would have such a reason.
While I recognize the intuition at issue, I am inclined to reject this sort of bifurcated view of subjective reasons. First, accepting this distinction would, to a degree, flout ontological parsimony, and thus doing so would require rather strong evidence. It is questionable whether the intuition at issue rises to that level. Second, there is a plausible pragmatic explanation of the relevant phenomena that would preserve the simpler view of subjective reasons. Since well-grounded beliefs are, in general, likely to be true, while poorly-grounded beliefs are not, the content of well-grounded beliefs (rather than that of poorly-grounded ones) ought to be the basis of one’s attitudes and actions. Therefore, it would behoove us to instinctually restrict our focus to subjective reasons that derive from well-grounded beliefs, and treat as essentially irrelevant, and even as non-reasons, those subjective reasons that derive from poorly-grounded beliefs. If this is plausible, then it serves to explain away our intuition that a certain type of subjective reason derives only from well-grounded beliefs. In any case, I shall assume for the purposes of this paper that there is only one sort of subjective reason: the kind that derives from one’s beliefs regardless of their ground.
It should be noted that Parfit himself fails to countenance subjective reasons—he believes that what people take to be subjective reasons are only ‘apparent reasons,’ since putative subjective reasons that derive from false beliefs ‘do not have normative force, in the sense that they do not count in favour of some act.’ (2011) But Parfit provides no argument for the claim that the considerations at issue lack normative force; indeed, it is unclear whether that claim is substantively different from the claim that such considerations are not reasons. And since the way we use ‘reason’ belies the view that reasons are confined to the objective domain, we should countenance subjective reasons until we have a good reason to do otherwise. In any case, I shall assume for the purposes of this paper that subjective reasons are at least plausible enough to warrant theorizing about them.
Way suggests that it is rational to ϕ just in case one has most subjective reason to ϕ.
Specifically, Schroeder suggests that intending to ϕ entails that one has a subjective reason to take the means to ϕ-ing.
Although my discussion will be framed in terms of subjective reasons for action, I believe my account plausibly applies, at least prima facie, to subjective reasons generally, e.g., to subjective reasons for belief, desire, and emotion. If there are objections to the application of my account to any specific domain of reasons, those objections may indeed serve as sufficient reason to restrict my account, and perhaps to restrict it only to reasons for action. I would not be entirely unhappy with such a result. My claim here is that the account of subjective reasons I shall defend applies at least to reasons for action. If it also applies to subjective reasons of all kinds (as I strongly suspect it does) then that is a welcome result, but a view to which I shall remain uncommitted for the purposes of this paper.
Although Schroeder’s official statement of his account of subjective reasons in ‘Means-End Coherence, Stringency, and Subjective Reasons’ does not include a subjunctive conditional, he does subsequently discuss the account in terms of the objective reasons there would be were one’s beliefs to be true. In his ‘Having Reasons’, he invokes subjunctive conditionals exclusively. See also Way (2009) and Parfit (2011), although Parfit’s is an account of ‘apparent reasons.’
I use ‘entails’ and cognate terms to refer to strict implication, i.e., necessary material implication. That is, p entails q just in case it is impossible that p is true and q is false.
The use of entailment in an account of subjective reasons might seem too strong, since (one may claim) it is unlikely that we have beliefs whose truth would guarantee that we have an objective reason. I address this worry in Section 4.
SR-2 represents one way of interpreting the following claim made by Schroeder: ‘X has a subjective reason to do A just in case she has some beliefs which have the property, if they are true, of making it the case that X has an objective reason to do A’ (2009a, b). Schroeder never tells us what ‘making it the case’ amounts to; but a reasonable interpretation is that for beliefs to make it the case that p just is for the propositional contents of those beliefs to jointly entail that p, in which case Schroeder’s account is equivalent to SR-2. I subsequently consider whether there is a plausible account of subjective reasons according to which ‘making it the case’ denotes a metaphysically or explanatorily robust relation, such as the in virtue of relation.
This can be shown as follows. If p entails q, then for each possible world w, the p-world(s) nearest to w are all q-worlds, since all p-worlds are q-worlds. Thus, ‘p entails q’ entails ‘necessarily, if p were the case, then q would be the case.’ To see that the reverse entailment holds, assume that for each possible world w, the p-world(s) nearest to w are all q-worlds. Now assume for reductio that p does not entail q. In that case, there is a world w i in which p is true and q is false. But since each world is nearest to itself, the p-world nearest to w i (i.e., w i itself) is not a q-world, which contradicts our assumption.
I do not assume that ‘n’ must refer to a number. I allow, for example, that one can hold the proposition that p merely with a high degree of credence, in which case the corresponding proposition in one’s α is that the probability that p is high.
Indeed, I would submit that if there is any probability that running away will save your life, then there is a reason for you to run away. Of course, this may be a comparatively weak reason, depending upon the probability—if there is only a very small chance that running away will save your life, then the strength of your reason to run away will be proportionally small. This issue is discussed in more detail in the following section.
It should be noted that SR-4 is equivalent to SR-3 just in case holding p with credence n just is believing that there is n probability that p is true. Thus, my stated rationale for rejecting SR-3—that it fails to countenance subjective reasons that derive from propositions one holds but does not believe—fails as a refutation of SR-3 unless we should reject the view that holding p with credence n just is believing that there is n probability that p is true. Now, that view is (at the very least) controversial (see, e.g., Frankish 2004 and 2009 for compelling objections), and I am inclined to reject it; but I need not take a stand on that issue here. For those who are more tolerant of the view at issue, my claim can be taken as follows: accepting SR-3 requires adopting a controversial view, whereas accepting SR-4 incurs no such commitment while incurring no further cost; thus, we should prefer SR-4 to SR-3.
The relevant sort of probability is objective probability, or probability given all that is true (e.g., the 20% probability that pushing the button will result in you receiving $1000), as opposed to information-relative probability, e.g., probability given what one believes or knows.
Strictly speaking, the fact that pushing the button will result in you receiving $1000 does not entail that you have a reason to push the button. For example, if you receiving $1000 will result in the immediate destruction of the universe, and you have no desire that would be served by destroying the universe, it is quite plausible that you lack a reason to push the button. Ordinarily, however, these sorts of conditions are absent, and we can assume that they are absent in the example. Still, the right proposition to play the role of p in Reason Probability must itself include the relevant caveats, e.g., that nothing significantly harmful will result from you receiving $1000, that you want an additional $1000, etc., in order to entail that you have a reason to push the button. But we can speak loosely, and say that ‘pushing the button will result in you receiving $1000’ entails ‘you have a reason to push the button’, keeping in mind that the relevant caveats tacitly apply; and it behooves us to do so, since speaking loosely makes discussion manageably simple.
Perhaps this will strike some readers as unsatisfactory. Some might be inclined to think that there simply is no subjective reason for you to dance around like a chicken, even a reason of extremely low weight. In response, I note Schroeder’s (2007) keen observation that when we are confronted with cases of reasons of extremely low weight, we are inclined for pragmatic reasons to think that there are no reasons there at all. If this is correct, then we should be skeptical about an intuition to the effect that you lack a reason to dance around like a chicken. (These remarks also apply above, mutatis mutandis, to an intuition that there is no reason to save the drowning man who very probably is a serial killer.)
This is the relation at issue in claims like ‘q is true in virtue of the fact that p.’ Accordingly, we can think of the relevant relata as facts; e.g., the fact that p stands in the making-it-the-case relation to the fact that q. But we can also speak loosely, and talk of propositions, when they are true, making it the case. This will be convenient shorthand in what follows.
An alternative formulation would not require that all the members of L jointly make it the case that there is an objective reason for one to ϕ, but rather only that the members of some subset of L do so. That is:
SR-5.3 There is a subjective reason for one to ϕ just in case for some consistent subset L of one’s α, and some subset M of L, the members of L jointly entail that the members of M make it the case that there is an objective reason for one to ϕ.
Another formulation of this general idea would not require that the propositions that make it the case that there is an objective reason for one to ϕ are themselves in one’s α, but require only that those propositions are entailed by some consistent subset of one’s α. That is:
SR-5.4 There is a subjective reason for one to ϕ just in case for some consistent subset L of one’s α, and some set of propositions N, the members of L jointly entail that the members of N make it the case that there is an objective reason for one to ϕ.
SR-5.2, SR-5.3, and SR-5.4 are variations on a theme, in decreasing order of strictness regarding which propositions are the right ones to make it the case that there is an objective reason for one to ϕ. My remarks about SR-5.2 in what follows will apply, mutatis mutandis, to SR-5.3 and SR-5.4 as well.
For a detailed exposition of such an argument, see Schroeder (2008).
The qualifier ‘probably’ is necessary here, since it may be that one does not hold the relevant beliefs with credence 1. In order for one’s subjective reason to be the proposition that running away will save your life, rather than merely that doing so will probably save your life, one must believe with credence 1 that running away will save your life, as well as believe with credence 1 the relevant “supporting” propositions, i.e., those that entail that if running away will save your life, then there is indeed an objective reason for you to run away.
This does not rule out that there are other subjective reasons for you to push the button, e.g., that you might receive $1000 by doing so. The objection under consideration is that the proposition that you will receive $1000 upon pushing the button is not among your subjective reasons, despite the fact that SR-Complete entails that it is such a reason.
Plausibly, the weight of a subjective reason is also partially determined by the weight of the corresponding objective reason, i.e., the objective reason that is entailed by the members of the relevant consistent subset of one’s α.
For example, suppose that it is possible for you to determine that 512,927,357 is prime within three seconds, but that it is very difficult for you to do so, such that you would fail to do so most of the time. In this case, the proposition that you will receive $1000 upon pushing the button possesses a low level of accessibility, but is accessible nevertheless; and so that reason would be weightier than in PRIME. And if we suppose that it is easy for you to determine that 512,927,357 is prime within three seconds, then your reasons would be significantly weightier—it would be, e.g., like the number 3 was on the paper.
Compare Sinnott-Armstrong (1984).
It is worth nothing that, as Schroeder points out, the view that existential propositions about reasons are themselves reasons is perfectly compatible with the most important claims, and indeed the spirit, of the buck-passing account of value. In particular, the fact that value is reducible to facts about reasons (the heart of the buck-passing view) does not at all suggest, as it may appear, that propositions of the form ‘there is a reason for A to ϕ, are not themselves reasons (a view that Schroeder calls the buck passer’s negative thesis, but a view that really has nothing to do with buck-passing per se). We can pass the buck (from value to reasons) perfectly well while allowing that existential propositions about reasons are themselves reasons.
References
Cullity G, Gaut B (1997) Ethics and Practical Reason. Oxford University Press, New York
Dancy J (2004) Ethics Without Principles. Clarendon, Oxford
Dummett M (2006) Thought and Reality. Oxford University Press, New York
Frankish K (2009) Partial Belief and Flat-Out Belief. In: Huber F and Schmidt-Petri C (eds) Degrees of Belief. Springer, Synthese Library, pp 75–93
Frankish K (2004) Mind and Supermind. Cambridge University Press, Cambridge
Frege G (1918/1988) Thoughts. In: Salmon N, Soames S (eds) Propositions and Attitudes. Oxford University Press, New York, pp 33–55
Hurley S (1989) Natural Reasons. Oxford University Press, Oxford
Joyce R (2001) The Myth of Morality. Cambridge University Press, New York
Lewis D (1973) Counterfactuals. Harvard University Press, Cambridge
Parfit D (2011) On What Matters. Oxford University Press, Oxford
Raz J (1999) Explaining Normativity: On Rationality and the Justification of Reason. Ratio 12(4):354–79
Raz J (2009) Reasons: Practical and Adaptive. In: Sobel D, Wall W (eds) Reasons for Action. Cambridge University Press, Cambridge
Schroeder M (2007) Slaves of the Passions. Oxford University Press, New York
Schroeder M (2008) Having Reasons. Philos Stud 139(1):57–71
Schroeder M (2009a) Means-End Coherence, Stringency, and Subjective Reasons. Philos Stud 143(2):223–48
Schroeder M (2009b) Buck-Passers’ Negative Thesis. Philos Explor 12(3):341–7
Sinnott-Armstrong W (1984) ‘Ought’ Conversationally Implies ‘Can’. Philos Rev 93(2):249–61
Strandberg C (2008) Particularism and Supervenience. In: Shafer-Landau R (ed) Oxford Studies in Metaethics, Vol. 3. Oxford University Press, New York, pp 249–60
Way J (2009) Two Accounts of the Normativity of Rationality. J Ethics Soc Philos http://jesp.org/articles/download/TwoAccountsOfTheNormativityOfRationality.pdf
Zangwill N (2008) Moral Dependence. In: Shafer-Landau R (ed) Oxford Studies in Metaethics, vol 3. Oxford University Press, New York, pp 109–27
Acknowledgements
I would like to thank Jonathan Dancy, Mark Schroeder, David Sosa, and two anonymous referees for Ethical Theory and Moral Practice for very helpful discussion and/or comments on earlier drafts of this paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Vogelstein, E. Subjective Reasons. Ethic Theory Moral Prac 15, 239–257 (2012). https://doi.org/10.1007/s10677-011-9286-7
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10677-011-9286-7