edited by

David Bakhurst, Brad Hooker, and Margaret Olivia Little

OXFORD UNIVERSITY PRESS, 2013

---

Introduction 1 – Brad Hooker
1 Acting in the Light of a Fact 13 – John McDowell
2 Can Action Explanations Ever Be Non-Factive? 29 – Constantine Sandis
3 The Ideal of Orthonomous Action, or the How and Why of Buck-Passing 50 – Michael Smith
4 Dancy on Buck-Passing 76 – Philip Stratton-Lake
5 Are Egoism and Consequentialism Self-Refuting? 97 – Roger Crisp
6 In Defence of Non-Deontic Reasons 112 – Margaret Olivia Little
7 The Deontic Structure of Morality 137 – R. Jay Wallace
8 Morality and Principle 168 – Stephen Darwall
9 Moral Particularism: Ethical Not Metaphysical? 192 – David Bakhurst
10 A Quietist Particularism 218 – A. W. Price
11 Contours of the Practical Landscape 240 – David McNaughton and Piers Rawling
12 Why Holists Should Love Organic Unities 265 – Sean McKeever and Michael Ridge
13 Practical Reasoning and Inference 286 – John Broome
14 Why There Really Are No Irreducibly Normative Properties 310 – Bart Streumer
Afterword 337 – Jonathan Dancy

---

Introduction
BRAD HOOKER

Jonathan Dancy has published on a wide variety of topics. He wrote an extremely successful textbook on epistemology and a monograph on Berkeley. He edited an edition of Berkeley’s Principles and co-edited a Companion to Epistemology. He has written articles on topics ranging from Proust to political philosophy. His greatest contributions to philosophy, however, have focused on practical reasons.
His arguments have pushed into mainstream discussion (a) the particularist view that moral reasoning can get along well (indeed is better off ) without universal principles, (b) the holist view that an act’s having a certain property in one context might be a reason in favour of doing the act and yet in another context an act’s having that very same property might be no reason to do the act or might even be a reason against doing it, (c) the view that enabling, disabling, intensifying, and attenuating conditions of practical reasons are no part of the content of practical reasons themselves, (d) the idea that some moral reasons entice rather than insist, and (e) the view that what explains intentional action are not the agent’s beliefs or desires, but the facts that are the reasons for which the agent acts. Furthermore, his defences of non-naturalistic moral realism have done much to help rehabilitate this metaethical view. In addition, Dancy has influentially attacked the possibility of organic unities, buck-passing accounts of goodness, and the view that Derek Parfit dubbed ‘admirable immorality’.
The essays collected in this volume address all the topics of Dancy’s major contributions to contemporary theorizing about practical reasons. The collection is offered as an expression of the contributors’ affection and esteem for a man who is not only an innovative and incisive philosopher but also an inspiring and courageous person. He and his work deserve to be honoured.
John McDowell’s contribution to this volume, ‘Acting in the Light of a Fact’, considers a position about practical rationality that Dancy sets out in his book Practical Reality. Dancy contends that, when agents act in the light of facts that give them reasons, the reasons are the facts themselves rather than the beliefs that the agents have about those facts, or the fact that the agents have those beliefs. This position has provoked attack but also attracted adherents, one of whom is McDowell.
McDowell’s chapter starts by explaining what is attractive about Dancy’s position and then discusses a modification to it suggested by Jennifer Hornsby. Hornsby points to cases in which an agent’s true belief seems to provide her with reasons for action, and yet the belief is held without sufficient warrant. McDowell’s discussion considers the lessons of disjunctivism about experiential warrant for belief and mistaken conceptions of falliblism. The conclusion he arrives at is that our ideal of practical rationality cannot be as independent of theoretical rationality as Dancy’s position suggests. Rather, the ideal of practical rationality pictures the agent as acting in the light of a fact that provides a reason for action and that the agent not merely believes obtains but knows obtains.
Constantine Sandis, in his contribution ‘Can Action Explanations Ever Be Non-Factive?’, turns the spotlight to Dancy’s account of motivating reasons for action. Starting in the mid-1990s and building to Practical Reality, Dancy attacked the belief-desire view of motivating reasons on the ground that motivating explanatory reasons must at least be capable of being normative reasons, and normative reasons are not beliefs and desires, but features of the situation one is in, rather than the agent’s mental states. Dancy’s own view is ‘non-psychologism’, according to which what the agent believes about the features of the various things she might do constitute the reasons for which she acts.
A prominent objection to Dancy’s non-psychologism points to cases where the agent acted for reasons, but her beliefs were false. Jill hits Jack in the face because she believes this will make him hideous, but then his injuries turn out to make him even better looking. In this case, how can we explain Jill’s action by pointing to the facts when there was no such fact? What she believed was false, and falsehoods cannot genuinely explain things. Her belief in a falsehood might be thought to explain her action, but Dancy’s non-psychologism bars this move. So, how can Dancy’s nonpsychologism about the reasons for which the agent acts provide an explanation of the agent’s acts when the agent’s beliefs are false?
In response, Sandis argues that we should keep the non-psychologism but dispense with the idea that agents’ reasons explain action. Agential reasons are indeed considerations external to the mind, considerations which we typically take to favour actions; but agential reasons themselves do not explain actions. We explain actions instead by claims of the form: ‘His reason for joining the investment club was that it would increase his pension.’ Such a claim makes reference to agential reasons, but that is not the same as saying that the reasons are what explain the actions. The explanation, properly understood, is the whole statement that his reason for joining the investment club was that it would increase his pension. When we share the agent’s beliefs we might say ‘she did it because p’ but this is a shorthand for the true explanation, made reasonable by the fact that saying so conventionally implies that the agent (a) took it to be the case that p and (b) took this to be a consideration that counted in favour of the action. Sandis concludes by taking up Dancy’s suggestion that contrasting English with other languages illuminates this issue, and by commenting on the debate over causalism in the theory of action.
Michael Smith’s ‘The Ideal of Orthonomous Action, or the How and Why of Buck-Passing’ begins by exploring two different ways in which we might explain what makes an action an action. One way is to characterize the ideal action, even though many or all actual instances of the kind do not match the ideal. The other way is to delineate the feature that all instances of actions, whether ideal or not—that is, all approximations to the ideal—have in common.
Smith argues that ideal actions are instances of what he and Philip Pettit have elsewhere called orthonomous actions. These are actions that are motivated by an intention an agent forms as a result of exercising his capacity to recognize and respond to such reasons as there are for intending to act in one way rather than another. Moreover, Smith argues, the capacity required for orthonomous action itself decomposes into two distinct capacities. It decomposes into (a) the capacity to recognize and respond to reasons for intrinsically desiring that things be a certain way and (b) the capacity to recognize and respond to reasons for believing that a certain bodily movement available to the agent will cause things to be that way. Given that the capacity required for orthonomous action decomposes into these two distinct capacities, Smith argues that it should be no surprise that what all instances of action have in common is their being caused in the right way by a belief-desire pair, no matter what the genesis of the belief-desire pair.
Smith also argues that this conception of ideal action shows us what is wrong with Jonathan Dancy’s objections to buck-passing accounts of goodness. According to the buck-passer, something’s goodness consists in its having other properties that give agents reasons for adopting various proattitudes towards it. Dancy rejects this view, but he accepts a buck-passing view of rightness, according to which an action’s rightness consists in its having other properties that give agents reasons to do the action. But Smith argues that buck-passing views about goodness and rightness form a package deal, in that to endorse the one is to endorse the other.
In closing, Smith argues that the buck-passing view of goodness is compatible both with ‘agent-neutral’ reasons (reasons necessarily shared by everyone to favour or disfavour various ways the world might be) and with ‘agent-relative’ reasons (reasons for each person to favour or disfavour various ways the world might be that are particularly connected with that person himself). Smith contends that this makes the idea of agent-relative values much less problematic than Dancy supposes.
Philip Stratton-Lake’s ‘Dancy on Buck-Passing’ starts by explaining how Tim Scanlon’s buck-passing account differs from earlier ‘fitting-attitude’ accounts of goodness, and proceeds to refine Scanlon’s account. Scanlon distinguishes between valuing a thing, which is to take oneself to have reason to care about it, and thinking something is valuable, or good, which is to take everyone to have reason to care about it. Scanlon’s distinction looks like the distinction between agent-relative and agent-neutral reasons. Stratton-Lake accepts that both kinds of reasons exist. However, in contrast to Michael Smith, Stratton-Lake argues against the idea of agent-relative value on the ground that this idea contains a confusing ambiguity.
Then Stratton-Lake turns his attention to Dancy’s attacks on buck-passing. Here again Stratton-Lake’s chapter intersects with Smith’s. Dancy has argued that his keeping a trivial promise he made to his children about which shoelace he ties first is something he has reason to do though it is not something of any value. Smith’s view is that this act does have agent-relative value. In contrast, Stratton-Lake, having argued against agent-relative value, thinks that Dancy’s keeping his trivial promise is not good unless it has agentneutral value.
Stratton-Lake next confronts Dancy’s claim that the buck-passing view illicitly settles some basic disputes in normative ethics, in particular debates between deontologists and consequentialists. Stratton-Lake argues that buck-passing does not settle these disputes, because proponents of buckpassing are committed neither to the view that any act for which there is a reason for action is good nor to the view that maximizing is always the appropriate response to value. Stratton-Lake ends by attacking Dancy’s argument that the buck-passing account implausibly makes something’s value or disvalue contingent on the existence of beings who can take proor con-attitudes to it.
Roger Crisp’s ‘Are Egoism and Consequentialism Self-Refuting?’ addresses a problem that Dancy found in Parfit’s Reasons and Persons. That book postulates that egoism gives one the overriding aim of having one’s life go as well as possible in self-interested terms, and that impartial consequentialism gives one the overriding aim of having the history of the world go as well as possible impartially considered. But having these aims is selfdefeating: the aims are not best achieved by having them.
While not agreeing with every element of Dancy’s attack on Parfit, Crisp ends up accepting that Parfit’s formulations of egoism and impartial consequentialism are vulnerable to Dancy’s attacks. Crisp proposes that, at the most basic level, egoism and impartial consequentialism should be formulated in terms of practical reasons instead of overriding aims. Egoism holds ‘[o]ne has most reason to act in that way in which one’s life goes, for one, as well as possible’. Impartial consequentialism holds that ‘[o]ne has most reason to act in that way in which the history of the world goes as well as possible’. Crisp goes on to observe that egoism and impartial consequentialism ‘will almost always require agents to follow some strategy in living or acting other than the continual application of the theory itself ’. Egoism will prescribe, not the overriding aim that one’s life go best in self-interested terms, but whatever aims will make one’s life go best in self-interested terms. Impartial consequentialism will prescribe, not the overriding aim that the history of the world go best impartially considered, but rather whatever aims will make the history of the world go best impartially considered.
Margaret Little’s ‘In Defence of Non-Deontic Reasons’ aims to redeem the idea, defended by Dancy, that some reasons for action are merely enticing. Such reasons are not, as she puts it, ‘in the wrong-making business’: they are capable of rendering action intelligible without bringing any deontic vulnerability in their wake. As she puts the point, ‘non-deontic reasons . . . justify one action without unjustifying anything else, whether that be doing nothing, doing exactly what you had been doing before this reason happened by, or pursuing an action that is supported by a different reason. They invite, as it were, rather than tell you what to do.’
Little proceeds by first arguing against a deontically reductive view of reasons: even if all reasons impose deontic vulnerability, they cannot be reduced to such imposition, on pain of an overly thin view of agency. Reasons also carry a commendatory force—an aspect under which the will may intelligibly act even though it would not be wrong not to do so. Some reasons for emotional states (such as anger) justify an emotional state without placing one in need of justifying its absence; likewise, some reasons justify an action without placing one in need of justifying not doing the action. Little goes on to argue that the normative force of some reasons is structurally like the illocutionary force of requests, which can give their recipients a good reason to do what was requested, without going so far as to require their recipients to do what was requested. (Contrast the noninsistence in a request with the insistence in an order.)
R. Jay Wallace’s ‘The Deontic Structure of Morality’, like Little’s contribution, is interested in the distinction between ‘deontic’ reasons for action that make demands on agents and ‘aspirational’ reasons that determine what the best action in the circumstances would be without going so far as to require it (Wallace’s ‘aspirational reasons’ are what Dancy and Little call ‘enticing reasons’). One might suppose that the way to make sense of the distinction is in terms of weight or importance. But Wallace argues that this path is a dead end. Likewise, appeal to Raz’s exclusionary reasons is inadequate. The problem is that sometimes deontic reasons are outweighed or trumped by considerations they might be thought to have excluded. More generally, Wallace argues that deontic significance is potentially independent of deliberative importance.
Another way of trying to explain the difference between aspirational and deontic reasons is to point out that deontic reasons have a connection to social sanctions that aspirational reasons do not. But we cannot plausibly hold that deontic reasons are necessarily connected to extant social sanctions. In fact, deontic reasons are not always accompanied by actual threats of sanctions. Nor can we plausibly hold that aspirational reasons are never connected to extant social sanctions.
Rather, what distinguishes deontic reasons from aspirational ones is that an agent’s deontic reasons are necessarily relational in a way that aspirational ones are not. For example, if there are deontic reasons for me to do X for you, then you have a claim to my doing X for you; and because of your claim on me, I wrong you if I fail to give due regard to the deontic reasons for my doing X for you. My failure gives you a reason for resentment that those who have no claim on me do not have. ‘The basic idea is that we have a certain latitude to ignore or discount aspirational reasons, of a kind we do not have when it comes to reasons that exhibit deontic structure.’
Wallace goes on to explain the difficulties act-consequentialism has accounting for the difference between deontic and aspirational reasons, and for the ‘presumptive connection’ between wrongness and negative reactive attitudes on the part of others. Much more promising, according to Wallace, is the contractualist idea that I wrong you if my action is not justifiable to you on grounds that you could not reasonably reject.
Stephen Darwall’s ‘Morality and Principle’ begins by accepting Dancy’s holism about practical reasons and Dancy’s distinction between reasons and enablers, disablers, intensifiers, and attenuators of reasons. Darwall goes on to argue, however, that there is an attractive rationale for general moral principles that Dancy fails to consider. While Dancy does attack the view that morality is like traffic regulations in being comprised of public rules justified by their benefits, he does not consider the idea that the concepts of moral obligation and demand are conceptually tied to accountability, or the idea that ‘agents can intelligibly be held accountable only if there exist general rules and principles that are accessible to all who are morally bound as a matter of common public knowledge’.
Dancy distinguishes between the moral reasons that are strongest in a particular case and reasons that are outweighed or somehow disabled in the case. Darwall contends, however, that we must distinguish between the moral reasons that are strongest in a particular case and reasons that come from moral obligations and demands. Of course moral obligations and demands often do provide the strongest reasons in the particular situation, but in some situations the moral reasons that happen to be strongest do not come from obligations (perhaps because in the situation no obligations come into play). When reasons are supererogatory, for instance, they do not amount to an obligation and do not render the agent accountable. To hold an agent morally accountable for a wrong is to say that the agent’s action is an appropriate target of ‘reactive attitudes’, such as guilt, resentment, indignation, and moral blame. But just as in the law negative sanctions are appropriate only if there is a publicly accessible rule which agents can be required to avoid breaking, in morality people can be morally blamed only for infringing rules that they should have known they had to observe.
David Bakhurst’s ‘Moral Particularism: Ethical Not Metaphysical?’ explores the moral philosophy of David Wiggins in search of resources that might supplement Dancy’s contributions to ethics. Dancy focuses on the theory of reasons for action and takes, as his primary test cases, specific situations in which an agent or agents must respond to the reasons in play there. But an acceptable picture of a good moral agent must not only account for her responses in particular situations but also display her character and enduring commitments, which, after all, contour those responses.
Bakhurst finds in Wiggins an account of the lived ethical life, shaped by its standing commitments. At least many of these can quite naturally be expressed as principles. And some of these standing commitments will be absolute in force because to forsake them is inconsistent with anything recognizable as a moral point of view. Nevertheless, Bakhurst argues that the Dancy-supplemented-by-Wiggins position is still particularist, because, in the normal case where conflicting considerations come into play and none is absolute, uncodifable judgement is needed to resolve what to do.
A.W. Price’s ‘A Quietist Particularism’ begins by considering how Dancy places enablers, disablers, intensifiers, and attenuators of reasons outside the content of reasons. Price first points out that, if any of these sit within the very content of reasons themselves, as opposed to merely sitting within the situational context of the reasons, then ‘it would become less plausible that further information may always cancel, reverse, intensify, or attenuate a given reason’. He goes on to ask whether, even if enablers, disablers, intensifiers, and attenuators are kept out of the content of reasons, there could not be principles articulating the conditions under which these enablers, disablers, intensifiers, and attenuators obtain. Indeed, Price himself shows that morally compelling facts ‘often have a wide degree of generality’ even if they aren’t perfectly general. He also points to side-constraints, where general principles are supposed to exclude or disable any other considerations.
Price then turns to Pekka Va¨yrynen’s contention that moral reasons have a ‘normative basis’ in ‘hedged general principles’. These are principles specifying the explanation of why a fact is sometimes a moral reason and sometimes not. Though people’s grasp of such principles will often be inchoate or incomplete, Va¨yrynen suggests that the underlying reality is systematic and determinate. Price is dubious. Pointing out that ‘certain rules (and other things, such as values, and paradigms) stand firm for us’, he is sceptical that they can be given ‘any grounding of which we are more certain than we are certain of these’. When such rules or values conflict, Price suggests, there may be no correct answer about what to do. At any rate, Price uses Va¨yrynen’s arguments to question whether non-realism might fit better with particularism than Dancy’s own realism does: ‘If we really thought, like the realist, that the moral world has its contours fixed independently of our ability to map them, it would hard to be sure that Va¨yrynen’s programme might not reveal its true topography. . . . The nonrealist is immune to the fantasy of a fixed ethical world waiting to be captured by ethical systematization.’
David McNaughton and Piers Rawling’s ‘Contours of the Practical Landscape’ presents a general account of practical reasons and the relation of the normative to the non-normative. They distinguish between facts that are themselves reasons and facts that certain other facts are reasons. Facts that certain other facts are reasons are always normative: it is a normative fact that the fact that this action would relieve someone’s pain is a reason for doing it. Facts that are themselves reasons for action are sometimes normative facts. For example, the fact that an act would cause undeserved harm is a normative fact and a reason not to do that act. McNaughton and Rawling observe that some facts have invariant polarity: that an act would cause undeserved harm always counts against the act. Here we have an example of what McNaughton and Rawling are willing to call a universal reason and a principle.
They go on to systematize reasons as functions of either impersonal value, of benefits to oneself, or of benefits to those with whom one has some special connection. They hold that there are reasons to do an act if, but only if, the act is of benefit or impersonal value. But the reasons one has to benefit oneself or those with whom one has special connections can outstrip the contribution of these benefits to impersonal value. Meanwhile, impersonal value is not purely a function of benefit and harm, since, for one example, it matters whether benefits or harms are deserved. McNaughton and Rawling close by casting doubt on the existence of moral constraints.
Sean McKeever and Michael Ridge’s ‘Why Holists Should Love Organic Unities’ argues that Dancy should accept rather than reject organic unities. The doctrine of organic unities, proposed by G. E. Moore, holds that the value of a whole need not be a mathematical function (e.g. equal to the sum) of the values of all the non-overlapping parts of that whole. As Dancy explains, holism in the theory of value has the same structure as holism in the theory of reasons. Something’s having a certain property in one context might be good-making, but in another context something’s having that very same property might be not be good-making. One might have thought that Dancy would accept the doctrine of organic unities, since, like Dancy’s holism, that doctrine stresses the complexity and variability of relations between reason-constituting or good-making features and their contexts. However, Dancy argues that the doctrine of organic unities is incoherent.
Dancy’s argument is that the doctrine is incompatible with the necessary connection between values and reasons. Take a case where a necessary part of a valuable whole supposedly has no value in itself. We have no reason to preserve the part as a part, but reason to preserve the valuable whole, which depends on the necessary part. Dancy argues that, since we have reason to preserve the whole including the part, the part must have value, because of the necessary connection between reasons and values. McKeever and Ridge catalogue a number of objections to this argument. One is that Dancy’s argument threatens his own distinction between values and their enabling conditions. For Dancy does not take the fact that enabling conditions of value are worthy of protection to imply that such conditions are themselves valuable. McKeever and Ridge’s chapter concludes by offering putative examples both of organic unities and of holism about reasons.
John Broome’s ‘Practical Reasoning and Inference’ engages with Dancy’s contention that practical reasoning is not based on inference. Broome begins by laying out a theory of reasoning conceived as a mental process. Taking inference to be a relation between propositions, he explains how theoretical rationality prohibits basing belief A on beliefs B, C, D, etc., unless the content of belief A can be inferred from the contents of beliefs B, C, D, etc. Theoretical reasoning involves applying (though not necessarily explicitly knowing) an inference rule.
Broome then turns to practical reasoning, or, more specifically, reasoning to an intention. After identifying an instrumental requirement, Broome argues that practical rationality cannot be derived from theoretical rationality, and that keeping practical and theoretical reasoning separate is tricky, because normally an intention to do X is accompanied by a simultaneous belief that one will do X. Finally, he argues that practical reasoning, when conscious, is also a rule-governed operation on conscious propositional attitudes. Broome’s conclusion is that, since practical reasoning is a mental process and inference is a relation between propositions, practical reasoning is not, strictly speaking, based on inference; still, practical reasoning does involve applying rules about the permissible basing of intentions on beliefs.
Bart Streumer’s ‘Are There Really No Irreducibly Normative Properties?’ starts by explaining that an irreducibly normative property is one that is not identical with any descriptive property. He then follows Frank Jackson in supposing that normative properties must be coextensive with the (perhaps disjunctive) sets of descriptive properties, across all possible worlds, on which the normative properties supervene. Jackson holds that necessarily coextensive properties are really the same property. So he concludes, against Dancy among others, that normative properties are actually reducible to the descriptive properties on which they supervene.
The usual way of attacking Jackson’s argument is to reject its premise that necessarily coextensive predicates ascribe the same property. Streumer carefully considers prominent putative counterexamples to Jackson’s premise and argues that they are misconceived. Streumer then turns to Jussi Suikkanen’s argument against Jackson. Suikkanen points to various higherorder properties that normative properties possess but descriptive properties do not, and thus concludes that normative properties are not identical with descriptive ones. Streumer replies that the apparent difference in higher order properties can be explained in terms of modes of presentation—a normative mode and a descriptive mode—without conceding there is a difference in higher-order properties.
Then Streumer turns to David Enoch’s contentions that the belief that there are irreducibly normative properties is indispensible to our deliberation about what to do and that we think normative properties are just too different from descriptive properties to be identical to them. Streumer responds that, even if these contentions about our thinking are true, our thinking might be systematically mistaken.
Our thinking might indeed be systematically mistaken. So we should test our ideas by considering evidence, arguments, and counterarguments in rigorous, fair-minded, and imaginative ways. That is exactly what Jonathan Dancy has done over a distinguished career, with results that have greatly enriched contemporary philosophy.