UC San Diego UC San Diego Electronic Theses and Dissertations
Title Abolitionism and the Value of the Reactive Attitudes
Permalink https://escholarship.org/uc/item/78p3g0vc
Author Milam, Per-Erik
Publication Date 2014
Peer reviewed|Thesis/dissertation
eScholarship.org Powered by the California Digital Library University of California UNIVERSITY OF CALIFORNIA, SAN DIEGO
Abolitionism and the Value of the Reactive Attitudes
A dissertation submitted in partial satisfaction of the requirements for the degree of Doctor of Philosophy
in
Philosophy
by
Per-Erik Milam
Committee in charge:
Dana K. Nelkin, Chair Richard Arneson David O. Brink Nicholas Christenfeld John Martin Fischer Christine Harris
2014
The Dissertation of Per-Erik Milam is approved, and it is acceptable in quality and form for publication on microfilm and electronically:
______
______
______
______
______
______Chair
University of California, San Diego
2014
iii DEDICATION
I would not have been able to write this dissertation without the help of many friends and colleagues. All of the errors are attributable to me, but many of ideas and insights are attributable to others and to the invaluable conversations I have had with those people generous enough to read my work and talk about my ideas. I am thankful for their help and support.
iv TABLE OF CONTENTS
Signature Page……………………………………………………………………………iii
Dedication………………………………………………………………………………...iv
Table of Contents…………………………………………………………………………v
Vita………………………………………………………………………………………vii
Abstract………………………………………………………………………………….viii
Chapter 1 – Living Without the Reactive Attitudes……………………………………….1 1.1 Reactive and Objective Attitudes……………………………………………...7 1.2 An Ought-Implies-Can Problem..……………………………………………13 1.3 Abolitionism, Explanation, and the Reactive Attitudes……………………...16 1.4 Plan of the Dissertation………….…………………………………………...24
Chapter 2 – Is Abolition Impossible?……………………………………………………28 2.1 Varieties of Abolitionism…………………………………………………….29 2.2 The Impossibility Objection…………………………………………………35 2.3 Alternatives to Abolition……………………………………………………..50 2.4 A World Without Reactive Attitudes………………………………………...53
Chapter 3 – At Enmity with Unreality: Is Abolitionism Self-Defeating?……………….57 3.1 Peritropes and Peritrope Arguments…………………………………………60 3.2 Epicurus vs. Determinism……………………………………………………64 3.3 Malcolm vs. Mechanism.…………………………………………………… 69 3.4 The Peritrope Argument against Abolitionism………………………………76 3.5 Conclusion…………………………………………………………………...91
Chapter 4 – Excluding the Forces of Ruin: A Defense of Analogue Attitudes………….94 4.1 The Abolitionist Dialectic……………………………………………………96 4.2 Empirical and Normative Challenges to Abolitionism………………………98 4.3 Pereboom’s Defense of Abolitionism………………………………………104 4.4 A Broader Defense of Abolitionism………………………………………..108 4.4.1 The Empirical Problem…………………………………………...110 4.4.2 The Contingency Problem………………………………………..127 4.4.3 The Generality Problem…………………………………………..133 4.5 Conclusion………………………………………………………………….135
Chapter 5 – Between Humanity and Intelligence: Abolitionism and Personal Relationships……………………………………………………………………136 5.1 The Objective Stance……………………………………………………….138 5.2 The Personal Relationships Argument……………………………………...139 5.3 A Defense of Abolitionism…………………………………………………144
v 5.4 Value, Scope, and Incompatibility………………………………………….153 5.4.1 The Value Problem……………………………………………….154 5.4.2 The Scope Problem……………………………………………….159 5.4.3 The Incompatibility Problem……………………………………..162 5.5 Methodological Epilogue…………………………………………………...166 5.6 Conclusion………………………………………………………………….169
Chapter 6………………………………………………………………………………..171
References………………………………………………………………………………181
vi VITA
2005 Bachelor of Arts, Pomona College
2010 Master of Arts, University of California, San Diego
2014 Doctor of Philosophy, University of California, San Diego
vii ABSTRACT OF THE DISSERTATION
Abolitionism and the Value of the Reactive Attitudes
by
Per-Erik Milam
Doctor of Philosophy in Philosophy
University of California, San Diego, 2014
Professor Dana K. Nelkin, Chair
I argue for Abolitionism—the view that if no one is free or responsible, then we ought to abandon the reactive attitudes. There are two powerful reasons, one moral and one epistemological, for this conditional obligation. Just as it is wrong to punish those who are not blameworthy, so it is wrong to blame those who are not blameworthy. Moreover, if we are not responsible, then taking reactive attitudes toward one another gets the facts wrong. When we resent one another, we ascribe properties, abilities, and capacities that the other does not actually have. Against opponents of abolitionism, I argue a) that we are able to prevent reactive attitudes from forming and to eliminate them once they arise and b) that human existence would not be impoverished by their absence.
viii Chapter 1
Living Without the Reactive Attitudes
Most would agree that if no one is responsible for their actions, then many of the presuppositions upon which our criminal justice system is based must be abandoned and institutions of punishment must be reformed. Indeed, our intuitions about wrongful blaming may be strong enough for many to concede an obligation to reform even if the necessary changes are difficult and costly. This reasoning generalizes. If no one is responsible, then there are many ways in which we must act toward and conceive of one another differently, including adapting existing social practices to fit this reality. One such practice is taking reactive attitudes, by which I mean blaming and praising attitudes.
In this dissertation, I argue for a position that I call Abolitionism.1 Abolitionism is the view that if no one is responsible, then we ought to abandon the reactive attitudes. If no one ever acts responsibly, then reactive attitudes like resentment, indignation, and guilt seem inappropriate. There are two powerful reasons, one moral and one epistemological, for this conditional obligation. On the one hand, if no one is responsible, then taking reactive attitudes toward one another gets the facts wrong.
Taking a reactive attitude toward a person in response to an action that was not done freely presupposes a false psychological or metaphysical claim, namely, that the person was free, or at least that she possesses certain capacities necessary for responsible action.
1 I borrow this label from the philosophical debate about the justifications of criminal punishment (e.g., R.A. Duff 2001).
1 2
When we resent one another, we ascribe properties, abilities, and capacities that the other does not actually have. For example, I attribute to her the capacity to deliberate and choose between different actions, the capacity to understand the nature and consequences of her action; and, on the basis of those attributions, I ascribe beliefs, desires, and intentions. I may even attribute to her the ability to do otherwise—i.e., to choose between two genuinely open futures. However, if she is not free, then these attributions are misguided and the presuppositions behind them false. Taking a reactive attitude toward a person who is unfree gets the facts wrong.2 I call this the epistemological reason for abolitionism.
On the other hand, taking a reactive attitude toward a person in response to an action that was not done freely is morally objectionable. Just as it is wrong to punish those who are not blameworthy, so it is wrong to blame those who are not blameworthy.
A person who is neither free nor responsible ought not be blamed for her wrongdoing.
Therefore, if taking negative reactive attitudes toward a person involves blaming her, which it does, then one ought not take such attitudes towards persons for their unfree actions. Doing so is wrong in the same way that blaming those who are severely mentally disabled or deranged is wrong.3 I call this the moral reason for abolitionism.
2 Susan Wolf (1981) calls this undesirable state “living in discord with the facts” and contrasts it will the desirable state of “living in accordance with the facts”. I will discuss living in accordance with the facts in more detail in chapter 3. 3 Whether the moral reason for abolitionism speaks as forcefully against the positive reactive attitudes—e.g., forgiveness, gratitude, pride, etc—is a more difficult question. Praising attitudes are unfair in the sense that they praise where no praise is deserved. No one deserves praise any more than she deserves blame. However, unfair blaming involves harms that unfair praising does not. For example, being blamed may cause one to feel fear and anxiety, while being praised usually will not. Moreover, some 3
These two reasons—i.e., avoiding getting the facts wrong and avoiding wrongful blaming—support abolishing the reactive attitudes.
Abolitionism has traditionally been driven by skepticism about the possibility or actuality of free will and responsibility. Traditionally, the antecedent of the abolitionist conditional has been established by arguing that the combination of incompatibilism and determinism precludes the ability to do otherwise, which is required for free will and responsibility, and that both incompatibilism and determinism are true.4 Incompatibilism is the view that free will and responsibility are incompatible with determinism, which is itself the thesis that at any instant in time only one future is consistent with both the state of the universe at that instant and the laws of nature (Mele 2006, 3). However, theories that deny that we have the ability to otherwise are not the only threat to free will and responsibility and are not the only means of establishing the antecedent of the abolitionist conditional. (Abolitionism is notably generic in this respect.)
Other skeptics have argued that, whether or not we have the ability to do otherwise, we are neither free nor responsible because we are not the ultimate source of our actions. Galen Strawson (1994), for example, famously argues that, unless one is a causa sui, one cannot be responsible. This conclusion holds regardless of whether determinism is true and regardless of whether we have the ability to do otherwise. Derk
Pereboom (2001), who is both a free will skeptic and an abolitionist, makes yet another
philosophers have argued that the bar for praiseworthiness is not as high as the bar for blameworthiness (Nelkin 2008), so on some accounts of non-responsibility, the blaming attitudes may be inappropriate, but the praising attitudes appropriate. 4 Of course, there have been those who have denied that determinism precludes the ability to do otherwise (Moore 1912, Aune 1967, Vihvelin 2004). 4 argument against free will and responsibility, which also turns on worries about being the ultimate source of one’s actions. Pereboom argues that we are neither free nor responsible for actions done as result of direct manipulation and that causally determined actions are not relevantly different from such actions (2001, 110-17; 2008). The hard incompatibilist theories just discussed are not the most widely accepted views of free will and responsibility, but the questions they raise remain open and the threats they pose remain live.
As a matter of fact, though, the truth of abolitionism does not depend upon the success of these skeptical arguments. Abolitionism it is a conditional claim and is therefore consistent with compatibilist theories of free will and responsibility—i.e., with theories according to which free will and responsibility are compatible with determinism.
The abolitionist simply claims that if no one is responsible, then we ought to abandon the reactive attitudes, and any compatibilist view that acknowledges even the possibility that no one is free or responsible can accept this conditional. Indeed, because compatibilist requirements for freedom and responsibility are more easily met than incompatibilist conditions, compatibilists should find it even more plausible that failure to meet these conditions makes taking reactive attitudes inappropriate. The compatibilist may believe that the antecedent is significantly less plausible than the hard incompatibilist alleges, but she can accept abolitionism nonetheless.5
5 This claim does not necessarily apply to compatibilists of the sort descended from Peter Strawson (1962). Strawsonian compatibilists have a more complicated story to tell, but I will address this issue later in the chapter (sec. 3). 5
It is not my aim to defend free will skepticism or the antecedent of the abolitionist conditional and I will not do so here. What matters for my purposes is that skepticism is a serious contender in the free will debate and a plausible position in the free will landscape. Taking the possibility that we are neither free nor responsible seriously is important and one way of doing so is evaluating the truth or falsity of abolitionism.
Peter Strawson (1962) considered the possibility of abolitionism when he introduced the reactive attitudes to the free will debate and its lexicon. And, having considered abolition, he rejects it. In “Freedom and Resentment,” he gives two reasons for doing so, which, as we will see throughout this dissertation, have been the foundation for all other objections to abolitionism. He writes,
It does not seem self-contradictory to suppose that this [abolition] might happen. So I suppose we must say that it is not absolutely inconceivable that it should happen. But I am strongly inclined to think that it is, for us as we are, practically inconceivable. The human commitment to participation in ordinary inter-personal relationships is, I think, too thoroughgoing and deeply rooted for us to take seriously the thought that a general theoretical conviction might so change our world that, in it, there were no longer any such things as inter-personal relationships as we normally understand them; and being involved in inter-personal relationships as we normally understand them precisely is being exposed to the range of reactive attitudes and feelings that is in question (68).
Having stated his conviction in the practical inconceivability of abandoning the reactive attitudes, he reluctantly speculates about what we ought to do in the inconceivable scenario.
And I shall reply, second, that if we could imagine what we cannot have, viz. a choice in this matter, then we could choose rationally [between abolishing and retaining the reactive attitudes] only in light of the gains and losses to human life, its enrichment or impoverishment (70). 6
The implication, evident in the first of the two quotations, is that a life without the reactive attitudes would indeed be an impoverished life, that it would be, as Susan Wolf later described it, “a world of human isolation so cold and dreary that any but the most cynical must shudder at the idea of it” (1981, 391). I will call these two types of argument impossibility arguments and outcome arguments. These opponents argue that it is impossible to abandon the reactive attitudes and that, even if we could, on balance the losses would outweigh the gains and human existence would be impoverished.
It is to these compelling descriptions that abolitionists like Derk Pereboom (2001 and forthcoming) and Tamler Sommers (2007), driven by the moral and epistemological reasons given above, object. They reject, by turns and by degrees, different elements of the Strawsonian picture of the Reactive World. Strawson rejects abolitionism as impossible and as undesirable. Pereboom and Sommers deny both of these claims. They deny that abandoning the reactive attitudes is inconceivable and they argue, to the contrary, that we can and do abandon them, though by degrees and sometimes with great difficulty. More importantly, however, they argue that the gains and losses to human life are balanced quite differently than Strawson and his descendents have claimed.
Pereboom argues that, in a world where no one is responsible for her wrongdoing, resentment and indignation are both inappropriate and (sometimes) have devastating consequences. And Sommers rejects the implication that abandoning reactive attitudes like resentment, gratitude, and forgiveness entails the cold, uncaring world that Wolf envisions. Central to both of their claims and to the plausibility of abolitionism is the claim that the reactive attitudes do not exhaust our emotional repertoire. Divesting ourselves of reactive attitudes does not entail complete emotional repression or isolation. 7
Abandoning a responsibility-presupposing attitude like gratitude does not preclude happiness or appreciation in response to another’s kindness. It is from this claim that my own abolitionist account develops. But, in order to understand this development, I must first say more about the reactive attitudes and the objective stance with which it is contrasted.
In this introductory chapter, I lay the foundation for a defense of the heterodox view that I call abolitionism. In section 1, I introduce the reactive attitudes and the objective stance. In section 2, I address a worry about the coherence of the abolitionist claim. In particular, I consider whether the truth of the antecedent of the abolitionist conditional entails the falsity of the consequent. In section 3, I discuss the relationship between responsibility and the reactive attitudes and assess whether some accounts of this relationship are inconsistent with abolitionism. Finally, in section 4, I describe the plan of the next five chapters of the dissertation.
1.1 Reactive and Objective Attitudes
We cannot evaluate abolitionism without understanding the class of emotions that philosophers call the reactive attitudes. Given its definition, there will be as many conceptions of abolitionism as there are conceptions of the reactive attitudes.
Unfortunately, there is little consensus about what a reactive attitude is beyond agreement about a core list of examples. However, we can come to an adequate understanding of the reactive attitudes by considering three distinctions.
The first distinction is between those conceptions according to which all reactive attitudes presuppose responsibility and those according to which only some reactive 8 attitudes presuppose responsibility.6 Strawson notably includes some emotions on his list of reactive attitudes that arguably do not presuppose responsibility. For example, he includes love and hurt feelings, neither of which, on its face, must be understood as a reaction to the behavior of a responsible agent (1962, 75).7 My feelings may be hurt by the lack of affection shown me by a friend’s child or pet. Some abolitionists seem to have taken such lists at face value and argue that it is only those reactive attitudes that presuppose responsibility that must be abandoned (Pereboom 2001, 200). Any reactive attitudes that do not are irrelevant to the abolitionist. For example, Pereboom argues that neither gratitude nor forgiveness presupposes responsibility and thus need not be abolished (2001, 201-2). Other philosophers, however, have argued that Strawson’s list of reactive attitudes is too inclusive and that its heterogeneity betrays the unprincipled boundaries of his conception. R. Jay Wallace thinks it is a mistake to understand the reactive attitudes—he calls them reactive emotions—in such an “encompassing manner”
(1994, 11). Instead, he proposes a conception of the reactive attitudes that includes only those attitudes by which we attribute responsibility.8 Wallace argues that a robust and
6 It should be noted at this point that some philosophers would object to the implication that reactive attitudes can presuppose responsibility at all. Some hold that it is more accurate to say that the appropriateness of the reactive attitudes explains responsibility. I consider this view in section 3. Until then, I will assume that at least some reactive attitudes presuppose responsibility. 7 Of course, it depends on what sort of love Strawson thinks counts as a reactive attitude. Love is itself a term that can refer to any member of an extremely broad and heterogeneous class of emotions, some which may, and some which may not, presuppose that the loved one is responsible. See, for example, Kane (1996) and Pereboom (2001 and 2009) for different views on whether love requires responsibility. 8 Wallace does not define the reactive attitudes as presupposing responsibility. Nonetheless, on his view, they are, by their nature, vehicles for attributing responsibility to agents for their actions. He conceives of a reactive attitude as a response to a violated 9 principled Strawsonian account of responsibility requires that we understand the reactive attitudes in this more precise and restricted way than Strawson himself did (1994, 11).
The result is a very small set of reactive attitudes, including the blaming attitudes, but, for example, none of the praising attitudes.
I do not argue for one or the other view—i.e., that all or only some reactive attitudes presuppose responsibility. Nor do I give an account of what distinguishes attitudes that presuppose responsibility from those that do not. However, in this dissertation, I will use the term ‘reactive attitude’ to refer to only those attitudes that presuppose responsibility. By stipulation, then, abolitionism is the view that, if no one is responsible, we ought to abandon all reactive attitudes. Those who hold a more inclusive conception of the class of reactive attitudes can understand abolitionism as the view that, if no one is responsible, we ought to abandon those reactive attitudes that presuppose responsibility or those attitudes by which we praise and blame ourselves and others. We will be making the same claim. Whichever conception you prefer, the more important question remains open, namely, which attitudes in particular ought to be abandoned?
However, we can be confident that the list of proscribed attitudes will at least include at least the big three—resentment, indignation, and guilt—which, as a matter of fact, are the
demand to which the victim holds the offender, which conception entails that only negative emotions count as reactive attitudes. He says, “To hold someone to an expectation is essentially to be susceptible to a certain range of emotions in the case that the expectation is not fulfilled, or to believe that the violation of the expectation would make it appropriate for one to be subject to those emotions…Emotions that are constitutively linked to expectations, in this sense of holding someone to an expectation, are the reactive attitudes, as I will interpret them” (1994, 21). On this view, the reactive attitudes include only resentment, indignation, and guilt. 10 attitudes about which the most ink has been spilled in the brief history of the debate about abolitionism.
The second distinction is between reactive attitudes and non-reactive participant attitudes.9 Strawson describes the reactive attitudes as, “the non-detached attitudes and reactions of people directly involved in transactions with each other” (1962, 62), and the reactive stance as, “the attitude (or range of attitudes) of involvement or participation in a human relationship” (66). For Strawson, the class of reactive attitudes and the class of participant attitudes are extensionally equivalent. He implicitly denies the distinction I am drawing here. However, the class of participant attitudes is large, its boundaries are fuzzy, and its members are very different from one another. As a result, it is difficult to make precise claims about their connection to responsibility and our practice of holding responsible. This is, in part, what Wallace objects to and what motivates him to reject
Strawson’s more inclusive conception of the reactive attitudes. Wallace excludes from the class of reactive attitudes any participant attitudes that “do not seem to have any presumptive relation to expectations at all—for example, gratitude, hurt feelings, and shame” (1994, 27). This account of the distinction between reactive and non-reactive participant attitudes probably entails the smallest class of reactive attitudes. Most philosophers other than Strawson, abolitionists and anti-abolitionists alike, appear to accept this distinction, but they delineate the subclass of reactive attitudes more broadly, often including gratitude, forgiveness, pride, and some kinds of love (Wolf 1981,
Pereboom 2001, Sommers 2007, Shabo 2012a and 2012b). I follow this trend. What
9 It will become clear that each of these three distinctions bears on the other two. 11 matters for the abolitionism debate is that we a) identify many of the same emotions as reactive attitudes, b) connect those attitudes to responsibility; and c) disagree about whether they can and ought to be abandoned.
The third distinction is between the reactive attitudes (and the corresponding reactive stance) and the objective attitude (and objective stance). I will use two different terms to talk about what Strawson called objectivity of attitude. An objective attitude is an individual response to an action or event; it is analogous to an individual response of resentment. An example of an objective attitude is the attitude a psychiatrist takes in response to the pathological actions of a patient, insofar as she refrains from taking his attitudes and actions personally. As with the reactive attitudes, there is no agreed upon definition, but objective attitudes can usefully be understood either by means of such examples or by means of generalizations from those examples. Some hold that the characteristic feature of an objective attitude is its detachment (Strawson 1962, Wolf
1981). Whereas reactive attitudes are participant attitudes taken by one person toward another (or toward herself), objective attitudes are attitudes of detachment or non- participation. A commonly cited example of the detached nature of an objective attitude is our emotional detachment from the inappropriate behavior of a cognitively disabled person. Others hold that the characteristic feature of such attitudes is that an objective attitude is the kind of attitude we take toward “natural objects” (Downie 1966, Nagel
1979, Sommers 2007). On this view, taking the objective attitude differs from merely abandoning the reactive attitudes. It requires denying that anyone is responsible in addition to refraining from taking attitudes, like resentment and guilt, that presuppose responsibility. And it requires viewing oneself and others as nodes in the causal web and 12 not as somehow separate from it. Proponents of this view claim that this is how we understand non-responsible individuals. In this dissertation, I accept the “natural objects” conception of the objective attitude, though I consider other conceptions when these alternatives bear on the plausibility of an argument.
More important than the objective attitude, for the abolitionist, is the objective stance. The objective stance is a general adoption of an objectivity of attitude. For example, while I may take an objective attitude toward my demented grandmother in response to a particular action, I may also take the objective stance toward her generally, for all of her actions or for all actions that I think result from the cognitive incapacities that are due to dementia. My stance is objective, though I may take reactive attitudes from time to time. A stance, in this context, is a disposition to react in a particular way, but which allows for exceptions. As a matter of fact, the reasons for taking an objective attitude typically will apply generally and not merely to a particular action. This is because the reasons for taking an objective attitude are usually extended in time or scope of influence. For example, dementia rarely improves and tends to affect a range of cognitive functioning. Some formulations of the abolitionist position hold that the moral and epistemological reasons for abolitionism are also reasons to adopt a wholly objective stance. I discuss these “strong abolitionist” positions in chapters 3 and 5.
Because I am considering the case of universal non-responsibility, I will typically use ‘objective stance’ to mean the universal adoption of the objective stance toward all actions of all agents. Those who argue that the truth of determinism requires adopting the objective attitude mean that it requires the adopting the objective stance in just this sense.
Susan Wolf, for example, asks us to, “Imagine for a moment what a world would be like 13 in which we all regarded each other solely with the objective attitude” (1981, 390; my emphasis). Seth Shabo also considers the implications of achieving “exclusive objectivity of attitude” (2012a, 133). As I said, adopting the objective stance will be the topic of debate in chapters 3 and 5.
Both of the above ways of characterizing the objective stance—as one of detachment or as a perspective on natural objects—might lead one to believe that it is simply the kind of attitude that we take in response to all non-responsible actions.
However, this is not the standard view. Strawson notes that we reserve the objective stance for those who are not responsible for their actions in virtue of some incapacity.
Thus, we take the objective stance toward a schizophrenic, but not toward a person who acts wrongly under duress. Neither is responsible for her action, but only one is proper target of an objective attitude or stance. With regard to cases of accidental or coerced, negligent (and presumably also reckless) wrongdoing Strawson writes, “None of them invites us to suspend towards the agent, either at the time of his action or in general, our ordinary reactive attitudes” (1962, 64). Whether or not we find the distinction between situational inability (e.g., duress) and personal incapacity (e.g., mental disability) ultimately compelling as an aid for identifying reasons for taking the objective stance, we can take Strawson’s view as at least a default assumption.
1.2 An Ought-Implies-Can Problem
In the last section, I provided rough sketches of what we will mean, going forward, by reactive attitude and objective stance. The purpose of these conceptions is to establish the meaning of the abolitionist conditional, so that we can evaluate it and its 14 implications. However, in this chapter, I address a worry about abolitionism that arises regardless of how we understand the reactive attitudes. This worry can be expressed as follows: abolitionism holds that if no one is responsible, then we ought to abandon the reactive attitudes. But if no one is responsible, then no one can do otherwise. And, if ought implies can, then the claim that one ought to abandon the reactive attitudes is illicit if no one can do otherwise than take them. In short, with the regard to the abolitionist conditional—if no one is responsible, then we ought to abandon the reactive attitudes— the consequent is false whenever the antecedent is true.
There are two responses to this worry. First, abolitionism is not the claim that, if we are unable to do otherwise (than we will do), then we ought to abandon the reactive attitudes. As I said above, inability to do otherwise is only one of many states of affairs that (purportedly) undermines freedom and responsibility. For example, one might hold a view like that of Galen Strawson (1994), according to which responsibility is impossible regardless of whether determinism is true and regardless of whether one can do otherwise. It is therefore a mistake to claim that abolitionism is incoherent because determinism precludes the ability to do otherwise, and the latter fact, together with the
Ought-Implies-Can Principle, undermines any ought claims. So long as it is not an inability to do otherwise that undermines responsibility, the abolitionist can coherently say that one ought to abandon the reactive attitudes.
Second, inability to do otherwise precludes some but not all ought claims. Many would agree that, for example, “If S ought not to have performed an action a, then S could have refrained from performing a” (Nelkin 2011, 100). However, even if we accept this version of the Ought-Implies-Can Principle, there are other ways of 15 interpreting the consequent of the abolitionist claim. For example, Pereboom
(forthcoming), following Gilbert Harman, distinguishes an ‘ought to do’ claim, like
Nelkin’s, from an ‘ought to be’ claim. A claim of the latter type—e.g., the global incidence of malarial infection ought to be lower—“evaluates a state of affairs and does not by itself imply that any particular agent has a reason to contribute to bringing about that state of affairs” (Harman 1977, 87; quoted in Pereboom forthcoming ms, 16). If we understand abolitionism in this way, we get the claim: if no one is responsible, then it ought to be the case that no one takes reactive attitudes. This is a weaker claim, but still interesting and controversial. And the same anti-abolitionist worries remain to be rebutted.
Pereboom also suggests that ‘ought to be’ statements sometimes imply reasons for an agent to act, but that the existence of these reasons still does not imply a way for the agent to “realize what ought to be” (Pereboom forthcoming ms 16). That is, these reasons for an agent to act do not imply that the agent can actually bring about what ought to be the case. For example, if the fact that the global incidence of malarial infection ought to be lower is of particular relevance to me but not to you, then I may have a reason realize this state of affairs, while you do not. And it may be neither inappropriate nor unfair to recommend that I do something to lower the global incidence of infection. Likewise, it may be neither inappropriate nor unfair to claim that we all ought to abandon the reactive attitudes, so long as we understand that this ‘ought’ derives from an ‘ought to be’ claim and not an ‘ought to do’ claim. This ‘ought’ is an ‘ought’ of what Pereboom calls ‘axiological evaluation’ and not an ‘agent specific demand’
(Pereboom forthcoming ms 18). 16
1.3 Abolitionism, Explanation, and the Reactive Attitudes
In section 1, I described what the reactive attitudes are. In this section, I consider different accounts of the relationship between the reactive attitudes and responsibility. In particular, I assess whether some accounts of the relationship between the reactive attitudes and responsibility are inconsistent with abolitionism.10
In the mid-twentieth century, a number of philosophers defended a revisionist view of responsibility that Richard Arneson (2003) has called responsibility as
“influenceability”.11 This account of responsibility has two main elements. First, it aims to provide a justification for holding responsible rather than identifying the conditions
(properties, abilities, or capacities) that make a person responsible for a particular action.
That is, it is concerned primarily with a normative rather than a psychological or metaphysical question. Second, its justification is deflationary and often consequentialist in flavor. On this view, the primary justification for holding a person responsible is the positive influence that doing so will have both on those held responsible and on others.
For example, blaming a person for her wrongdoing may prevent her from acting wrongly again and may deter others from following her bad example. This forward-looking justification is deflationary insofar as it is incompatible with conceptions of praise and blame that appeal to what the agent deserves in response to her action.12 The revisionist
10 Many accounts of responsibility and the reactive attitudes will be inconsistent with the antecedent of the abolitionist conditional. However, as I said above, my aim here is to discover whether any accounts are inconsistent with the conditional itself being true. 11 For example, J.J.C. Smart (1961). 12 However, it is not incompatible with a forward-looking justification of the desert-based practices. People may be most easily influenced, and influenced in the right direction, by practices that appropriate the vocabulary of desert (Smart 1961, 68-9). 17 view refers only to whether the wrongdoer (or others) can be influenced to avoid future wrongdoing. J.J.C. Smart, for example, recognizes this feature of his view and proposes replacing the vocabulary of “praise” and “blame” with the less metaphysically and normatively loaded notions of positive and negative grading (1961, 69-71). However, this is not to say that it does not matter at all for the influenceability view that the person actually committed the action in question. For it may be that certain means of influencing a person via moral blame are effective only if the person blamed is the same person who did the action. For example, it may be that blaming attitudes only make greedy people less disposed to act on their greed when those attitudes are responses to actual greed-based wrongdoing. If so, then influenceability views are also backward- looking in a sense.
Peter Strawson was responding, in part, to this kind of view when he introduced the notion of reactive attitudes. He suggested, among other things, that the deflationary conception of responsibility as influenceability fails to capture the particular backward- looking nature of our responsibility judgments. When we hold a person responsible, we judge not only that she can and ought to be influenced by our blaming practices, but also that she deserves blame for her action. The reactive attitudes—e.g., resentment, indignation, and guilt—can serve both influencing and blaming functions. For example, when Sally resents Jesse for standing her up at dinner, her resentment is not only directed at Jesse, it is also (hopefully) recognized by Jesse the next time they meet or speak to one another as being about the fact that she stood Sally up. On the influenceability view, the nature and content of the wrongdoing is irrelevant to the judgment that Jesse ought to be held responsible, except insofar as we must determine whether we can disincentivize her 18 future wrongdoings and how best to do so. A reactive attitude looks backward to the offense rather than forward to its own consequences for its warrant.
Many contemporary philosophers, both compatibilist and incompatibilist, recognize a relationship between taking the reactive attitudes and whether a person is responsible. However, for those who conclude that we are neither free nor responsible, the reactive attitudes, like other institutions of blaming such as prisons, become targets for abolition. Thus, abolitionists like Derk Pereboom and Tamler Sommers argue that we ought to abandon the reactive attitudes to whatever extent is possible.
Strawson’s account of the reactive attitudes has been incredibly influential and has been adopted in various ways and to varying degrees by many philosophers working on questions of free will and responsibility. A common account of the connection between responsibility and the reactive attitudes is expressed by the Strawsonian biconditional, which states that the reactive attitudes are appropriate if and only if the targets of these attitudes are responsible.13 Amongst those who hold this view, two groups can be distinguished by their diverging explanatory claims. Both groups accept the above biconditional, but they diverge with respect to which side has explanatory priority. Following Brink and Nelkin (2013), I will call the two groups ‘realist
Strawsonians’ and ‘response-dependent Strawsonians’ or simply ‘realists’ and ‘RDSs’.
Realists hold that whether one is responsible explains the appropriateness of taking
13 This statement is adapted from what Brink and Nelkin call ‘Strawson’s thesis’. However, like others, they hold that only some reactive attitudes presuppose responsibility. As a result, they state Strawson’s thesis differently. “Reactive attitudes involving blame and praise are appropriate just in case the targets of these attitudes are responsible” (2013, 287). Fischer and Ravizza affirm essentially the same biconditional (1998, 7-8). 19 reactive attitudes toward her. The RDS holds that the appropriateness of taking reactive attitudes toward a person explains that fact that she is responsible.14
For the realist, the issue of whether we ought to abandon the reactive attitudes arises as soon as we have reason to doubt that we are free or responsible. Given the
Strawsonian biconditional, if a person is not responsible, then it is not appropriate to take reactive attitudes toward her. This is true whether the realist is an incompatibilist or a compatibilist. For the RDS, however, the abolitionist worry does not arise so easily. If the fact that one is morally responsible is explained by the fact that one is an apt target of the reactive attitudes, then the claim that we should abandon the reactive attitudes because we are not responsible appears confused.
In order to see exactly how the RDS thesis conflicts with abolitionism, we must look beyond the biconditional itself. The biconditional states that the reactive attitudes are appropriate if and only if the targets of these attitudes are responsible. The abolitionist claim is that if no one is responsible, then we ought to abandon the reactive attitudes. This claim is very similar to a conditional that is directly entailed by the biconditional. If it is not the case that the targets are responsible, then it is not the case that the reactive attitudes are appropriate (and we ought to abandon them). Thus, it looks as if anyone, including the RDS, who accepts the biconditional should accept abolitionism.
At first glance, then, we might conclude that the disagreement between the realist and the RDS is orthogonal to the abolitionist claim. If the RDS agrees with the
14 Brink and Nelkin describe these two types of Strawsonians (2013, 285-288). 20 biconditional, then she can agree that if no one is responsible, it is inappropriate to take reactive attitudes toward anyone. As an analogy, consider the claim that X is a biological parent of Y if and only if X has directly passed genes to Y. This is a true biconditional, but explanation only goes in one direction. It is not true that being a biological parent explains the direct gene transmission. It is only true that directly passing genes explains being a biological parent. Nonetheless, anyone who has reason to deny that X is the parent of Y also has reason to deny that X has directly passed genes to Y.15 Similarly, anyone who has reason to deny that we are responsible also has reason to deny that it is appropriate to hold people responsible by taking reactive attitudes toward them.
However, the problem created by the competing explanatory interpretations is deeper than this. The RDS defends the claim that the fact that the reactive attitudes are appropriate explains the fact that their targets are responsible. The reason that the target of a reactive attitude is responsible is because the reactive attitude is appropriate. The problem arises because the abolitionist conditional is also an explanatory claim. The abolitionist claims that reactive attitudes are inappropriate because their targets are never responsible. In short, we have two mutually exclusive explanatory claims. The RDS claims that whether reactive attitudes are appropriate explains whether their target is responsible (Strawson 1962, Wallace 1994). The realist claims that whether their target is responsible explains whether reactive attitudes are appropriate (Fischer and Ravizza
15 Thanks to Dana Nelkin for suggesting this example. Or consider an analogous case. Suppose the sun is 30° above the horizon. The height of a flagpole is X if and only if the length of its shadow is √3X. In this scenario, anyone who has reason to doubt that the length of the shadow is √3X also has reason to doubt that the height of the pole is X. And this is true even though it is the height of the pole that explains the length of the shadow and not the other way around. 21
1998, Pereboom 2001, Brink and Nelkin 2013). If abolitionism is an explanatory claim, then it is incompatible with the RDS explanatory claim. If abolitionism is not an explanatory claim, then it is compatible. Thus, if we want abolitionism to be an explanatory claim, then we must accept realism. While all other accounts of responsibility and the reactive attitudes can, in principle, accept abolitionism, the RDS must deny it.16
Abolitionism is an explanatory claim. It is the view that we ought to abandon the reactive attitudes because we are neither free nor responsible. The abolitionist is therefore committed to realism and to the denial of the RDS thesis. While it appeared that abolitionism, being a conditional, was consistent with any account of freedom and responsibility, in fact it is inconsistent with a Response-Dependent Strawsonian compatibilism. As a result, the arguments of this dissertation do not apply to the RDS and I will be assuming a realist view of the reactive attitudes from here forward.
Luckily, there are good reasons to reject the RDS thesis. I cannot adequately refute the view here, but let me briefly offer some reasons for rejecting RDS in favor of a realist view of responsibility. Brink and Nelkin note that, “Because the realist believes
16 There is actually a third option besides a) accepting RDS and denying abolitionism or b) accepting realism and accepting that abolitionism may be true. Michael McKenna defends a view according to which the two explanatory claims are not mutually exclusive; both are true, but neither has explanatory priority. He writes, “Perhaps it is mistaken to think that on a Strawsonian approach, in facing up to our Euthyphro-type question, we can defend an exclusive one-way-only direction of explanatory priority. Assuming one does not wish simply to deny the truth of the biconditional, there is a third option, which is to deny that there is an exclusive one-way-only order of explanatory or metaphysical priority. Neither is more basic” (2012, chapter 2). If neither explanatory claim has priority, then abolitionism is compatible with the RDS explanatory claim. As a result, the abolitionist can claim that her view neither presupposes nor excludes a particular account of responsibility and the reactive attitudes. 22 that the reactive attitudes presuppose responsibility, she can appeal to our practices of exemption and excuse to help understand the conditions under which we are responsible”
(2013, 288). The RDS cannot do this. According to the RDS, a non-responsibility claim—e.g., S is not responsible for A—must be grounded in a claim that a reactive attitude is inappropriate. And the inappropriateness claim must be grounded in a claim about the failure of the reactive attitude to be consistent with the general pattern of responses to others—i.e., consistent with the general pattern of reactive attitude taking.
As Brink and Nelkin put it, “Particular expressions of a reactive attitude might be corrigible as inconsistent with a pattern of response, but the patterns of response are not themselves corrigible in light of any other standard” (287). For the RDS, explanations of inappropriateness bottom out in the failure of an attitude to fit the pattern of our attitude taking. Meanwhile, for the realist, explanations of inappropriateness can appeal to incapacities or situational factors that diminish control. Noting this difference is less an argument than an explication of the RDS commitment, but it is (to me at least) a much more intuitive sense of what it means for one fact to explain another.
Angela Smith (2007) provides another reason for rejecting the RDS explanatory claim.17 Recall that, according to this view, the fact a person is responsible is explained by the fact that she is an appropriate target of a reactive attitude—i.e. that it is appropriate to hold her responsible by, say, resenting her. Smith objects to this because, as she says,
Our intuitions about whether and when it would be fair to react negatively to another are sensitive to a host of considerations that appear to have little or nothing to do with an agent’s responsibility or culpability for her
17 Michael McKenna (2012), in turn, considers this objection to the RDS view and offers his own defense, against Smith, of the RDS position. 23
attitudes or behavior. If this is correct, then theories which make attributions of responsibility dependent upon the appropriateness of our reactions as moral judges will turn out to be fundamentally misguided (2007, 466-7).
She goes on to say that, “Blaming is a way of responding to faults in ourselves and others, and can be unfair or inappropriate for any number of reasons” (472). On Smith’s view, whether it is appropriate to take a blaming attitude toward a person depends upon, for example, one’s standing to blame, the significance of the offense, and the how the offender herself has responded to her offense (478-483). Intuitively, none of these factors bears upon whether the offender is responsible, but the RDS view cannot capture this intuition because whether a person is responsible is entirely explained by the appropriateness of taking reactive attitudes toward her. This counterintuitive implication of the RDS view provides another reason to reject it in favor of realism.
I do not take these considerations to be decisive against the RDS view. However, by noting them, I hope to show that we have good reasons independently of abolitionism to prefer realist interpretations of the Strawsonian biconditional over response-dependent interpretations. Abolitionism is not compatible with every view about free will and responsibility, but it is compatible with most. And, as such, we have good prima facie reasons, both moral and epistemological, to investigate whether it is true. In this dissertation, I will henceforth assume a realist interpretation of the Strawsonian thesis— i.e., of the relation between reactive attitudes and responsibility.
24
1.4 Plan of the Dissertation
In the following chapters, I argue for abolitionism: if no one is responsible, then we ought to abandon the reactive attitudes. Abolitionism begins on its back foot, so making the case for this heterodox view requires first addressing a number of arguments against it. Each of the following four chapters will address an important attack on abolitionism. However, while these chapters dispel the force of the main anti-abolitionist argument, they also contribute to the positive case for abolitionism. The present chapter has articulated the moral and epistemological reasons driving abolitionism.
Understanding the unsuccessful objections to abolitionism will not only emphasize the force of these driving reasons, it will also illustrate what it would be like abolish the reactive attitudes. Each chapter thus contributes to a more accurate picture of life without the reactive attitudes.
Four main arguments are marshaled against abolitionism, each with roots in
Strawson’s seminal account. I call these broadly Strawsonian objections the
Impossibility Argument, the Peritrope Argument, the Social Harmony Argument, and the
Personal Relationships Argument. The first and second are impossibility arguments.
They claim that it is impossible to abandon the reactive attitudes. The third and fourth are outcome arguments. They claim that on balance the losses of abandoning the reactive attitudes would outweigh the gains and human existence would be impoverished. I offer novel responses to each of the four and argue that none of them is sound. Having dispelled the anti-abolitionist worry, I conclude by summarizing the benefits retained and gained by abandoning the reactive attitudes. 25
In chapter 2, I consider and reject the Impossibility Argument. Proponents of this objection to abolitionism claim that we cannot abolish the reactive attitudes (Strawson
1962). Therefore, if ought implies can, we cannot be required to abolish them.18 But human beings can abandon the reactive attitudes. Not only do attitudes like resentment and guilt often fade away, we also actively eliminate such attitudes when we, for example, excuse actions we once believed to be blameworthy. Moreover, empirical studies of purposeful emotion regulation suggest that we can eliminate not only occurent reactive attitudes, but also our dispositions to such attitudes (Ekman et al. 2005).
In chapter 3, I address what I call the Peritrope Argument. Proponents of this argument claim that acting on one of the two driving reasons for abolitionism, namely, getting the facts right, is self-defeating. Susan Wolf (1981) argues that we are not required to abandon the reactive attitudes because doing so does not actually allow us to get the facts right. But neither does continuing to take the reactive attitudes. If no one is responsible, then we must live in discord with the facts whatever we do. This surprising and counterintuitive conclusion rests on a subtle and ingenious argument about the commitments entailed by our reasoned actions, including the reasoned abandonment of the reactive attitudes. However, while the Peritrope Argument does undermine some formulations of the abolitionist position and thereby demonstrates the limits of the view, it nonetheless remains inadequate as a refutation of abolitionism.
18 This objection is different from the one considered earlier in this chapter, which also appealed to the Ought-Implies-Can Principle. The Impossibility Argument denies the consequent of the abolitionist conditional—i.e. it claims that we are not obligated to abandon the reactive attitudes because we cannot do so. The incoherence argument considered above claimed that whenever the antecedent of the abolitionist conditional is true, the consequent is false—i.e., it claimed that the conditional itself is always false. 26
The Social Harmony Argument and the Personal Relationships Argument attack the moral reason for abolitionism. In chapter 4, I address the Social Harmony Argument proposed and defended by Shaun Nichols (2007) in response to the abolitionist proposals of hard incompatibilists like Derk Pereboom (2001 and 2009) and Tamler Sommers
(2007). Nichols argues, on empirical grounds, that reactive attitudes perform important psychological and social functions and that they do so more effectively than any non- reactive attitudes that an abolitionist might suggest to replace them. I dispute the empirical evidence offered in support of this claim. Moreover, even if reactive attitudes are more effective, the relative effectiveness of different attitudes at performing a given function will inevitably be due to contingent biological, cultural, and historical facts. I argue that our emotional repertoire could have been different but equally functional.
Given that moral and social progress often begins by exposing a system ill-suited to supporting the necessary reforms, I argue that we cannot assess the gains and losses to human life of living without the reactive attitudes solely by reference to their usefulness within our present system of practices.
In chapter 5, I consider and reject the Personal Relationships Argument.
Proponents of this argument claim that abandoning the reactive attitudes, in particular by adopting the objective stance, precludes the formation and maintenance of valuable personal relationships.19 They claim that the value of such relationships—in particular reciprocal romances and mature friendships—is sufficient reason not to abandon the
19 This is a particularly interesting argument because it is one that divides the abolitionist camp. Pereboom agrees with Strawson and the anti-abolitionists that adopting the objective stance precludes personal relationships (2001, 199-200), while Sommers does not. 27 reactive attitudes, even if there are good reasons to do so. I contend that this is simply not true and that taking the objective stance is compatible with the deployment of a rich and flexible emotional repertoire that can provide the foundation for these valuable relationships. And I argue that this anti-abolitionist argument exaggerates the importance of reactive attitudes for relationships and ignores or minimizes other factors, independent of the reactive attitudes, which play an important role in human social life.
In chapter 6, having defended abolitionism against these four arguments, I summarize the upshots for abolitionism and paint the picture of a world without reactive attitudes that has emerged from each of the preceding chapters.
Chapter 2
Is Abolition Impossible?
It is the act of an ill-instructed man to blame others for his own bad condition; it is the act of one who has begun to be instructed, to lay the blame on himself; and of one whose
instruction is completed, neither to blame another, nor himself.”
- Epictetus, Enchiridion
When I introduced the abolitionist position at the beginning of the previous chapter, I drew an analogy between abolishing the reactive attitudes and abolishing state institutions for punishment. One might reasonably believe that state punishment of criminals is just only if the victim is guilty of responsible wrongdoing. But if criminals are never responsible for their wrongdoing, then perhaps we ought to abolish prisons, or at least eliminate the retributive element of criminal punishment. And we might make similar claims about institutional practices of reward. If no one is responsible, then perhaps we ought not, for example, award prizes for literary or artistic accomplishments.
Radical institutional changes of this sort may seem inconceivable. However, the magnitude of institutional reform necessary to bring our practices in line with the above suppositions pales in comparison to the scope of the implications of those suppositions for our interpersonal practices. For one might just as reasonably believe that responding with resentment or indignation to a wrongdoer is just or fair only if that person acted
28 29 responsibly and that if no one is responsible, we ought to abolish the reactive attitudes. I have called those who take this view abolitionists and their view abolitionism.
In this chapter, I lay the foundation for the abolitionist project by arguing that abolitionism is possible. In section 1, I identify different formulations of the abolitionist view and describe the dimensions along which they differ from one another. In section 2,
I argue that it is possible to abolish the reactive attitudes. In particular, I argue that it is not only physically (or nomologically) possible, but that we are able—we have the ability—actually to abandon these attitudes. In section 3, I consider alternative solutions to the abolitionist worries raised by the facts (if they are facts) that we are neither free nor responsible. In particular, I consider the case for partial abolition. Finally, in section 4, I briefly speculate about what it would be like to live without the reactive attitudes. I suggest that the bleakest imagined scenarios—those that depict humans divested of reactive attitudes as robots or emotional incompetents—are exaggerated, but also that we must not ignore the reasons for the pessimists’ disquiet.
2.1 Varieties of Abolitionism
Abolitionism is a conditional claim. If no one is responsible, then we ought to abandon the reactive attitudes. In this chapter, I defend abolitionism against what I call the Impossibility Argument. Proponents of this argument hold that abandoning the reactive attitudes is impossible and that, because ought implies can, it cannot be the case that we ought to abandon them. I contend, to the contrary, that abolition is possible.
However, abolitionism is less a single thesis than a class of views. The different members of this class differ with regard to what abandoning the reactive attitudes is 30 thought to entail. It is important to understand these different formulations of abolitionism because the Impossibility Argument makes a stronger or weaker claim depending on what abolition entails.
Let me introduce and explain the different ways in which we might understand abolitionism. Five distinctions will help to illuminate the different dimensions of the abolitionist position. First, abolition can be understood either as the inhibition or abandonment of incipient or occurent attitudes or as the prevention or elimination of the disposition to experience reactive attitudes. Second, it can be achieved either by preventing the development of dispositions to reactive attitudes or by eliminating existing dispositions. Third, it can be understood as promoting either the abolition or merely the disavowal of the reactive attitudes. Fourth, it can be understood as promoting either full or partial abolition of the reactive attitudes. And, finally, it can be understood as promoting either the abolition of the reactive attitudes alone or, in addition, the adoption of the objective stance.
The attitude/disposition distinction. There is some ambiguity about what it means to abolish a reactive attitude. One interpretation holds that abolition simply requires constant vigilance against incipient and occurent attitudes. On this view, a person abolishes the reactive attitudes just in case, when she spontaneously experiences an instance of a reactive attitude (or recognizes its onset), she eliminates that attitude (or prevents it from arising). For example, Sally abolishes resentment if, when she resents
Raphael for treating her poorly, she somehow exorcizes and ceases to feel that resentment. 31
Abandoning an attitude can itself be understood in two ways, which can be made explicit by reference to the distinction between occurent and standing beliefs—i.e., between a presently conscious belief and a belief that one holds continuously but which is not always conscious. One can abandon an occurent attitude—e.g., a present experience of resentment—or one can abandon a standing attitude—e.g., a persistent but not always presently conscious resentment about the same offense. One can abandon occurent resentment simply by distracting oneself from the offense that caused it. For example, if my friend stands me up at the movie theater, I can cease to feel resentment simply by going to the movie and devoting my attention solely to that. However, if the resentment is a standing attitude, it will return when I am reminded of her offense. I can abandon the standing attitude by, for example, finding out why she stood me up. If I learn that she was very ill, then I will excuse the fact that she did not show up and will have resolved the conflict in such a way that, if I think about it again later, my resentment will not return.20
A second interpretation of what it means to abolish a reactive attitude holds that abolition requires that a person not be disposed to spontaneously experience reactive attitudes in the first place—i.e., that one lack what I will henceforth call a reactive
20 Alternatively, I may eliminate the standing attitude by forgiving the person. For example, if my friend apologizes for standing me up and takes responsibility, I may forgive her and cease to have either occurent or standing resentment. There is even empirical evidence which suggests that a victim who works through the conflict, whether doing so results in an excuse or in forgiveness, is less likely to have her resentment return than one who merely distracts herself (Larsen et al. 2012). 32 disposition.21 On this view, a person abolishes a reactive attitude just in case she (or someone else) controls her cultural, developmental, or social circumstances in such a way that she is not disposed to it. For example, Jesse has abolished a disposition to resentment, or it has been eliminated in her, if she no longer is prone to experience this attitude toward seemingly appropriate targets of it. Recognizing this distinction, and the relative difficulty of abolishing dispositions to, as opposed to instances of, the reactive attitudes, is crucial to assessing the possibility of abolitionism.
The prevention/elimination distinction. Implicit in my explanation of the previous distinction was the distinction between preventing a reactive attitude or disposition and eliminating a reactive attitude or disposition. We prevent reactive attitudes by avoiding people or circumstances likely to cause them or by developing the ability both to recognize their mental precursors and to regulate our emotional states before they actually arise. We eliminate reactive attitudes by, for example, reflecting on whether they are justified by the circumstances of the action (and agent) to which they are responses. For example, I eliminate resentment by recognizing that the person who spilled juice on my laptop was a small child. Similarly, we prevent reactive dispositions by inhibiting their development. This can be done in various ways, including by manipulating the environment within which children develop their emotional repertoires.
And, finally, we can eliminate existing reactive dispositions. This is perhaps the most difficult of the four tasks identified by the previous two distinctions—preventing and
21 I take this to be what Strawson means when he speaks of a “proneness” to the reactive attitudes (1962, 74) and what Wallace refers to as our “susceptibility” to this type of emotion (1994, 21). 33 eliminating attitudes and dispositions—and one that I will discuss in more detail in the next section.
The abandon/disavow distinction. It is also important to distinguish between abandoning an attitude or disposition and disavowing it. To abandon an attitude is to prevent or eliminate an instance of, or disposition to, that attitude. To disavow an attitude is to remove the implicit endorsement that comes with experiencing it. For example, the master guru who trains herself never to experience resentment has abandoned the disposition to this attitude. The novice who has not yet eliminated the disposition will sometimes experience incipient or occurent resentment. However, the novice can act on her commitment to abolitionism not only by continuing her efforts to eliminate her disposition to resentment, but also by disavowing instances of resentment when they arise. For example, when she recognizes the heat of resentment at those who take advantage of her charity, she can step back and withdraw her endorsement from this attitude, as we do when we get angry at our computers or the weather. This distinction is important because it may be that complete elimination of all dispositions to all reactive attitudes is impossible. If this is the case, then while complete abandonment of the reactive attitudes cannot be required—assuming that some version of the Ought-Implies-
Can Principle is true—we can nonetheless be required to disavow those intransigent instances of resentment, guilt, and pride.
The full/partial distinction. It is not obvious whether the reasons for abandoning the reactive attitudes support full or partial abolition—i.e., abolition of all or of only some reactive attitudes, respectively. At first glance, and given my stipulation that all reactive attitudes presuppose responsibility, it would appear that taking any reactive 34 attitude requires both that we get the facts wrong and that we unfairly blame or praise people. Thus, with regard to the epistemological reason for abolition, if we are required to abandon the reactive attitudes, we are required to abandon all of them. However, two considerations support a partial requirement. First, with regard to the moral reason, unwarranted praise, via attitudes like gratitude or pride, is not morally objectionable in the same way that unwarranted resentment or indignation are. The latter are both unfair and harmful. The former are perhaps unfair, but not harmful. Or, more precisely, unfair praising is only harmful to those who are unfairly excluded from praise. Another consideration in favor of partial abolition is the possibility that, while partial abolition is possible, full abolition is not. It may be that some reactive attitudes simply cannot be removed from our natural emotional repertoire. Or, it may be that abolishing some attitudes precludes the possibility of abolishing others. I will discuss the possibility of such “interaction effects” between attitudes in the next section.
The strong/weak distinction. Some abolitionists argue that we should abandon the reactive attitudes, but that we should not adopt the objective stance. Rather, we should replace the reactive attitudes, where necessary, with non-reactive analogues that do not presuppose moral responsibility. For example, Derk Pereboom proposes that we abandon guilt for a close cousin, which resembles regret, that does not presuppose responsibility and that we abandon resentment altogether and resort to moral sadness or concern (2001,
200). I call this view weak abolitionism or just abolitionism. Tamler Sommers, on the other hand, argues that we should both abandon the reactive attitudes and adopt the objective stance toward other agents and their actions (2007). I call this view strong abolitionism because it has an additional requirement, namely, adopting the objective 35 stance. Both versions have critics and in subsequent chapters, I will defend both abolitionism and strong abolitionism. Strong abolitionism is stronger or weaker depending on what it means to take the objective stance. I discussed two different views of the objective stance in chapter 1, section 1. One view holds that the objective stance is a detached stance; the other holds that it is a stance whereby one views other beings, including fellow humans, as “natural objects”.
Summary. In this chapter and the three to follow, I argue that, if we are neither free nor responsible, we can and should abolish the reactive attitudes to the greatest degree possible. In the next section, I consider the degree to which we can abolish reactive attitude and the extent to which we must settle for partial abolition and/or disavowal of reactive attitudes and dispositions.
2.2 The Impossibility Objection
Abolitionism about the reactive attitudes is a position that is firmly contained within the contemporary free will debate. However, the claim that we ought to regulate our emotional lives in light of, and even by means of, theoretical discoveries has a long history. Both the Epicureans and the Stoics advocated emotional restraint for various therapeutic reasons. Epicurus saw his metaphysical views as tools for eliminating fear of death, chance, and the gods.22 Similarly, many Stoics argued that emotions like anger, resentment, disappointment, and grief arise as a result of our disordered opinions and can be eliminated with significant benefits. For example, in the Enchiridion, a Stoic
22 See his Letter to Menoeceus in Brad Inwood’s The Epicurus Reader (1994, 28-31). 36 handbook of Epictetus’ ethical teachings compiled by Arrian, Epictetus says that, “Men are disturbed not by things which happen, but by [their] opinions about the things” (§5).
If you can control these opinions, including desires and aversions, which are your own, rather than trying to control the actions of others, which are not, “no man will ever compel you, no man will hinder you, you will never blame any man, you will accuse no man, you will do nothing involuntarily (against your will), no man will harm you, you will have no enemy, for you will not suffer any harm” (§1). Seneca defends similar claims in his De Ira. The Stoics, Epicureans, and others who counseled emotional regulation and abolition of various forms and to various degrees all assumed that that these manipulations were possible, at least for some people. I turn now to an assessment of this crucial assumption.
In the previous section, I described different formulations of the abolitionist position. Philosophers have raised a number of objections to abolitionism. The most important of these are the subject of the following three chapters. However, before addressing these concerns, one fundamental objection requires immediate attention. This objection, which I have called the Impossibility Argument holds that abolishing the reactive attitudes is impossible and that, because ought implies can, we cannot be required to abolish them.
In this section, I consider whether it is possible to abandon the reactive attitudes.
In order to answer this question we must be clear about the scope of the abolition suggested and the nature of the modal claim. Regarding the scope of abolition, our question is most naturally understood as a question about whether we can abandon all the reactive attitudes, that is, whether a person can live without ever (or ever again) 37 experiencing a reactive attitude. We want to know whether we can eliminate all reactive dispositions and prevent humans, by whatever means, from developing them in the first place.
Regarding the nature of the modal claim, lacking or eliminating the reactive attitudes is certainly not logically or conceptually impossible.23 There is nothing contradictory or incoherent in imagining their absence from our emotional repertoires.
Nor is it metaphysically or physically impossible. We can conceive of possible worlds in which they are absent and it is consistent with the physical laws of the universe, as we know them, that a human being never experience a reactive attitude. (Indeed, depending upon how significantly the reactive attitudes and dispositions depend on cultural influences, early anatomically modern humans may have lacked such emotions in the actual world.) Reactive attitudes are a class of emotions, which makes them empirical phenomena. Their existence in our emotional repertoire is contingent on evolutionary, cultural, and developmental forces. With regard to the modal claim, we are interested in what human beings can be like psychologically and how they can change themselves, both in the present and in the future.
One way of putting the modal question, then, is to say that we are interested in psychological possibility. This expresses our interest in what our psychology can be like, how much it can differ from the current norm and still be recognizably human.24
23 Strawson himself allows that, “It does not seem self-contradictory to suppose that this [the decay or repudiation of participant reactive attitudes] might happen. So, I suppose we must say that it is not absolutely inconceivable that it should happen” (1962, 68). 24 More revisionary abolitionists might not require so much. They might be interested solely in physical possibility. They might not care whether human beings without 38
However, we cannot speak of psychological possibility in the same way we would speak of, say, physical possibility because few recognized psychological regularities plausibly warrant the label of ‘law’. We identify a state of affairs as physically possible when it is conceivable and consistent with the physical laws of the universe. For example, it is physically possible for me to run a marathon in under two hours, in part, because it would violate no laws of nature. Unfortunately, this model cannot be applied to psychological possibility. We can argue that some imagined scenario is conceivable and consistent with psychological facts, but this cannot be our standard for psychological possibility because what we want to show is exactly that something inconsistent with the current facts could be the case. We are interested in consistency with a subset of the psychological facts.
We want to know whether we can be psychologically different than we are but still recognizably human.
Another way of putting the modal question, then, is to say that we are interested in ability, specifically our ability to prevent and ability to eliminate reactive attitudes and dispositions. Do we have the ability to prevent ourselves from experiencing reactive attitudes and from developing reactive dispositions? And do we also have the ability to eliminate the occurent and standing attitudes and existing dispositions to them? The difference between conceiving of the question as one of psychological possibility versus
reactive attitudes can live lives that are at all similar to those we live now. Rather, they might only be interested in what allows human beings to live the best lives possible and, if that means living lives that are not recognizably human, then so be it. To value ‘recognizably human lives’ over the human lives we might have if we purged reactive attitudes entirely from our species’ emotional repertoire is like valuing native ecosystems solely because they are native rather than because of the services native ecosystems provide. It gives unwarranted preference to the status quo. 39 one of ability is merely the difference between asking, “What can we be like?” and
“What can we do?” Conceiving of the question in terms of ability is most natural when asking whether we can change how we are now. Conceiving of the question in terms of psychological possibility is most natural when asking whether we can be a particular way in the future or whether we could have been different than we are.
In this section, then, my task is to answer two questions. First, are we able to eliminate all reactive attitudes or dispositions thereto? Second, are we able to do so while living recognizably human lives? If the answer to both questions is yes, then, the consequent of the abolitionist conditional is possible. If the answer to either question is no, then there is some degree to which we cannot be required to abolish the reactive attitudes.
Strawson famously remarks on the possibility of abandoning the reactive attitudes, “I am strongly inclined to think that it is, for us as we are, practically inconceivable” (1962, 68). And Thomas Nagel agrees that, “We are unable to view ourselves simply as portions of the world…We cannot simply take an external evaluative view of ourselves” (1979, 185). Strawson goes on to say that, “This commitment [to ordinary interpersonal attitudes] is part of a general framework of human life, not something that can come up for review as particular cases can come up for review within this general framework” (1962, 70). It may be possible to inhibit particular instances of a reactive attitude, like resentment. It may even be possible to weaken one’s disposition to resentment. However, it is not possible to abandon the entire framework.
Tamler Sommers is more optimistic about the possibility of abolition, though he acknowledges that progress will take time. 40
[I]f I am right, there exists a continuum of progress one can make in the rejection of moral responsibility and the full-time adoption of the objective attitude. Perhaps there are some who have achieved total success. Perhaps not. But there is no denying that our natural commitments, convictions, and intuitions on this issue can soften to a large degree, sometimes without our even noticing it (2007, 340).
His optimism does not lead him to conclude that full abolition is possible, but he acknowledges that the basis for such a conclusion will be the degree to which our emotional repertoire is contingent and alterable. In what follows, I try to establish just how malleable we are in this respect.
My aim in this section is to argue that there is good reason to think that we can abandon the reactive attitudes. In order to do so, I describe scenarios that, if they are conceivable and consistent with the subset of psychological facts that we think are required for a recognizably human life, provide such justification. These scenarios do not purport to show that the process of actually abandoning the reactive attitudes, either individually or as a society, will be quick or easy.
Attitudes. I begin by suggesting that a person is able to abandon a particular instance of a reactive attitude. This can mean either eliminating an occurent or a standing reactive attitude. We know that one can eliminate an occurent attitude because it happens. People often cease to feel resentment because they intentionally distract themselves and cease to think about whatever offense prompted it. Moreover, the process of abandoning resentment in this way is largely under our control, at least to the degree that we are capable of such self-manipulation. For example, Jesse may cease to feel resentment toward Sally for cheating on him because he has recreationally medicated himself and dulled his emotional responses. (Some may be reticent to use active verbs 41 like “abandon” or “eliminate” to describe what Jesse does. I am not so reticent, but I do recognize that we sometimes do not play an active or conscious role in the elimination of an attitude.)
A similar case can be made for abandoning standing attitudes. While an instance of occurent resentment might be eliminated, perhaps by distracting oneself, the same standing resentment can and often does return when we turn our attention back to the offense. If Jesse distracts himself, he may temporarily cease to feel resentment, but the same resentment returns when he is no longer distracted (Larsen et al. 2012). However, even standing attitudes can be and are in fact abandoned. This is the case when, for example, Jesse forgives Sally’s wrongdoing, perhaps because he learns that she is remorseful and is no longer willing to disregard his interests. Not only has he ceased to resent Sally, he has also resolved the conflict that gave rise to the attitude in the first place. This does not ensure that his resentment has vanished forever—for he might later decide that what he thought was a sincere apology actually was not—but his standing resentment has been abandoned. We can see this more clearly by analogy with a standing belief. Suppose I cease to believe that charities working in the developing world are ineffective. The fact that I lack this belief does not mean that I will never reassess the evidence and change my mind, but it still seems reasonable to say that I have abandoned my standing belief about the effectiveness of such charities.
It appears, then, that we can and do abandon both occurent and standing instances of reactive attitudes. However, it may be the case that one instance of a reactive attitude is always replaced by another. If so, then it would still be true that we can never truly abolish reactive attitudes. This would be the case, for example, if resentment were 42 always replaced by forgiveness. But reactive attitudes are not always replaced with others. While we often replace resentment with forgiveness, we also sometimes excuse an offense and, in doing so, we do not necessarily replace resentment with another reactive attitude. I conclude, then, that we can eliminate particular reactive attitudes. I contend that we can also prevent these attitudes. However, the primary mode of doing so is by manipulating our reactive dispositions.
Dispositions. Now let us ask whether it is psychologically possible, or whether we are able, to prevent or eliminate a disposition to experience a particular reactive attitude. Sommers argues for the possibility of divesting from one’s reactive dispositions by drawing an analogy to leaving behind one’s religious beliefs. He imagines a theist who, in light of compelling arguments against the existence of God, abandons both the belief in God and, over time, the habits and behaviors of a believer. And he suggests that, “It is certainly possible that someone could come around to live the denial of RMR
[robust, desert-entailing moral responsibility] in the same way” (2007, 339). I want to move a little bit more slowly from such cases to a possibility claim. Living the denial of a personal God is a life-changing transition for a theist, but it is still importantly different from living the denial of responsibility. Nonetheless, I agree with Sommers that it certainly seems possible to abandon our reactive dispositions. Consider the following imagined scenario.
Ungrateful. Commodus has grown up as the only son of a powerful Roman ruler. For as long as he has had desires, every one of them has been provided for. Nor has he been led to believe that what he receives is anything other than what is owed him by basic right. Perhaps he cannot even conceive of having a desire that is unsatisfied for longer that a short while. As a result of such pampering, Commodus is ungrateful. He is ungrateful not in the sense that he is displeased with his retainers for their 43
efforts, but in the sense that he never feels gratitude toward them. He is not at all disposed to gratitude, nor has he ever felt the emotion, disconnected as he is from the belief that the objects of his desire might be withheld.
I suggest that the story of Commodus is conceivable, that it is consistent with the subset of psychological facts that describe the (admittedly fuzzy) class of recognizably human lives, and that it provides good reason to believe that a person can lack a disposition to a common reactive attitude like gratitude. (Again, it is a separate question whether such a scenario is desirable. I will address questions of this sort in chapters 4 and 5.)
Our next step, then, is to ask whether a similar story can be told for each and every reactive disposition. Unfortunately, many of the reactive attitudes are importantly different from one another, so we cannot infer from the possibility of abandoning any particular reactive disposition that we can abandon all of them. Moreover, some attitudes are likely to be more intransigent than others, so it is worth considering the more difficult cases. However, if we could show that even the most intransigent reactive attitude can be abandoned, we would be one step closer to justifying the belief that all attitudes can be.
If we can abandon the most intransigent disposition to a reactive attitude, then we can abandon any such disposition. Presumably, the most difficult reactive attitudes to abolish are those with the greatest evolutionary and cultural importance. Some have suggested that the anger-based blaming attitudes, which are purportedly necessary for social cohesion, best fit this description (Nichols 2007), so let us turn to one such attitude.
Consider another imaginary scenario, this time concerning the disposition to resentment.
Misfortune. Thomas has lived a poor life. An orphan, he has never known a family or had any regular source of love or comfort. Living on 44
the streets since he was child, he has more often than not been the target of contempt, derision, and ill treatment. Nor has he been given any reason to think that he can expect or that he deserves any better treatment than he has received. As a result, he does not respond to the slings and arrows of outrageous fortune with resentment. His expectations of others are so low that they are always met. He is not surprised when the grocer sells him bad vegetables or the police beat him up. He is unmoved by the reality that life is brutish, people are nasty, and opportunities are few and poor. However, it is not the case that Thomas cannot feel anger. Rather, he has not experienced decent treatment or good fortune of the sort that might provide evidence that his misfortune is deserving of frustration and his ill treatment deserving of resentment.
Like Ungrateful, this scenario is both conceivable and within the range of recognizably human lives, albeit at the far end of the likelihood and desirability scales.25 We can imagine states of affairs that have existed throughout history that, when combined, might have produced people like Thomas. There have been times when life was held cheaply and ties of affection were the exception rather than the rule, even between kith and kin.
Some may object that these imagined scenarios tell us very little. This objection deserves both a concession and a defense. I concede that such scenarios at most give us reason to believe that human beings can be made to lack dispositions to certain attitudes, including intuitively intransigent ones. Because they both depend on a person never experiencing the material conditions necessary to form expectations of the relevant sort, they only support the claim that we (or circumstances) can prevent people from developing reactive dispositions. They do not give us reason to believe that an individual can eliminate an existing disposition.
25 Jean Hampton describes cases of severe self-abnegation amongst women raised in traditional gender roles. Such self-abnegation, which, again, is extreme in the cases she describes, seems to result in just the sort of distorted repertoire of reactive dispositions that might preclude, to a significant degree, resentment or indignation. See, in particular, Hampton’s discussion of Terry (1993, 135-6) and the poem, “Severed Heart” (160-1). 45
However, in defense of our power to eliminate reactive dispositions, I would respond that these imagined scenarios should be taken to complement more concrete evidence of our ability to actually abolish a disposition to a reactive attitude. Consider practitioners of Buddhist and other sorts of meditation. One of the four noble truths of
Buddhism is the claim that suffering can be avoided by abandoning one’s desires and thereby avoiding the frustrations that arise when they are not satisfied. Figures like the
Buddha who purportedly achieve this state or something like it are plausibly described as lacking dispositions to attitudes like grief and resentment. Moreover, psychologists who study Buddhist meditative practices agree that practitioners are attempting and accomplishing a sort of emotional regulation.
The initial challenge of Buddhist meditative practice is not merely to suppress, let alone repress, destructive mental states, but instead to identify how they arise, how they are experienced, and how they influence oneself and others over the long run. In addition, one learns to transform and finally free oneself from all afflictive states. This requires cultivating and refining one’s ability to introspectively monitor one’s own mental activities, enabling one to distinguish disruptive from nondisruptive thoughts and emotions (Ekman et al. 2005, 60).
This description not only identifies a group of people who practice the elimination of various “afflictive” emotions, it also describes how the practitioners do so, namely, by careful introspection and monitoring. Moreover, unlike the imagined scenarios above,
Ekman et al. describe a practice of abolishing preexisting dispositions to these emotions as opposed to preempting the development of such dispositions. While the class of
“afflictive emotions” is not the same as the class of reactive attitudes, it seems likely that 46 at least some of the latter are members of the former.26
I contend that such cases, both real and imagined, justify the belief that we can abolish dispositions to particular attitudes. Indeed, if some of the afflictive emotions from which Buddhists succeed in freeing themselves by meditation are reactive attitudes, then we have a concrete demonstration of the actuality of such abolition. Finally, if anger-based attitudes like resentment are the hardest to abolish, and those toward which we are most prone, and it is possible for a person to prevent or eliminate all instances of and dispositions to these attitudes, can we infer that it is also possible that a person could prevent or eliminate all reactive dispositions? Unfortunately, no.
Here we face a different obstacle. For there could be interaction effects between the absence of one attitude and another, whereby the absence of one makes the abolition of another harder than it would otherwise be. Ungrateful and Misfortune provide an example of one possible interaction effect. If these scenarios are taken as guides for how we might abandon gratitude and resentment, respectively, then it would appear that the material conditions that make possible the absence of gratitude are incompatible with the material conditions that make possible the absence of resentment. These cases demonstrate is that reactive attitudes require expectations of their targets.27 Gratitude
26 Thus, for example, some Buddhists appear to believe that, “When the mind is obsessed with resentment, it is trapped in the deluded impression that the source of its dissatisfaction belongs entirely to the external object (just as, in the case of craving, the mind locates the source of satisfaction in desirable objects). But even though the trigger of one’s resentment may be the external object, the actual source of this and all other kinds of mental distress is in the mind alone” (Ekman et al., 2005, 61). Notice also the similarity to Epictetus’ view quoted in the epigraph to this chapter. 27 An expectation is not simply a prediction. To say that Sally expects Jesse to care for her in times of need is not merely to say that she predicts that he would do so. Her 47 toward a person requires that one’s expectations be exceeded in some sense. Resentment requires that one’s expectations be unmet in a way that one believes constitutes wrongdoing. The situation that purportedly precludes gratitude is one in which expectations can never (or rarely) be exceeded because they are so high. The situation that purportedly precludes resentment is one in which expectations of the sort to which we typically hold others can never (or rarely) be formed in the first place because others’ behavior never supports their formation. It seems impossible to experience both
Commodus’s life and Thomas’s life and so no one person can lack the disposition to both attitudes.
Similar interaction effects may exist between other attitudes. More generally, it may be that the reason we have such a complex and subtle repertoire of reactive attitudes is that we live in a world of moderate hardship and moderate good fortune, a world where things and people could be both better and worse and where we require attitudes with which to manipulate ourselves and others in order to make our lives as good as possible.
There are two responses available to the worry about interaction effects. First, one might just accept the core of the objection and acknowledge that the circumstances that make possible abolition of one type of attitude preclude abolition of another. (Of course, we should also acknowledge that while some interaction effects may preclude abolishing both attitude X and attitude Y, other effects may guarantee the abolition of
expectation has a normative dimension; she also believes that he owes her something. R. Jay Wallace’s account of holding to expectations could explain why reactive attitudes require expectations (see chapter 1, footnote 8). On his view, it is impossible to hold to expectations and lack a reactive disposition. However, this is only one way of explaining the connection between reactive attitudes and expectations. 48 both. For example, a common view of forgiveness is that it is, by its nature, the foreswearing of resentment (Butler 1900, Murphy 1988). On such a view, abolishing resentment entails abolishing forgiveness.) Or, one might propose some form of partial abolitionism. One plausible formulation, given intuitive support by the above discussion, might require only abolition of the negative reactive attitudes. I discuss this alternative form of abolitionism in the next section. Second, one might accept the claim that scenarios like Ungrateful and Misfortune are incompatible with one another, but deny that these are the only scenarios that might allow for the abolition of, for example, gratitude and resentment.
I find this second response plausible. For, while interaction effects of the sort described above are possible, it is also possible that replacing the abolished reactive attitudes prevents these effects from obtaining. For example, the scenario we must imagine is not one in which we lack gratitude and anything like gratitude and also lack resentment and anything like resentment. Rather, according to those abolitionists who propose replacing reactive attitudes with analogues, what we should imagine is a scenario in which gratitude persists in essentially the same form and function minus an assumption that its target was responsible for her good deed. Meanwhile resentment is replaced by the quite dissimilar but (purportedly) equally effective analogue of moral sadness. This would not require a scenario like Ungrateful and thus would not open the abolitionist’s possibility argument to the worry about interaction effects. This defense requires that we articulate and defend the account of non-reactive analogues of the reactive attitudes.
However, most parties to the debate accept the existence of such analogues and only 49 disagree about their relative effectiveness, so I will delay discussion of them and the case for their effectiveness until chapter 4.
I have discussed a type of interaction effect between attitudes that precludes or makes more difficult the abolition of both. However, we should also acknowledge that in some cases, the opposite seems more likely and that, if anything, it would be easier to abandon some attitudes together or to abandon one once another had been abandoned.
For example, if resentment and indignation are the same but for their target—i.e., if they are both constituted in part by appraisals of wrongdoing and feelings of anger but are experienced on behalf of different victims (oneself or another)—then it would seem that abolishing resentment would quite naturally coincide with the abolition of indignation.
Even guilt, which differs in important ways from the other two blaming attitudes, would seem less stable or entrenched once resentment had been abandoned. For guilt, too, involves an appraisal of wrongdoing and negative affect felt on behalf of a victim that may be either oneself or another.
Thus far, I have argued for the possibility of abolition by pointing to imagined and actual cases in which we have prevented or eliminated reactive attitudes and dispositions. I want now to make a brief speculative point about the inconceivability of abolition. I suspect that one reason why abandoning the reactive attitudes appears inconceivable to some is that we fail to distinguish between abandoning reactive attitudes and abandoning our basic feelings. For, while it may be impossible to abolish the feeling of or the disposition to anger, it is much easier—which is not to say easy—to abolish moral anger about responsible wrongdoing, which is a more complex emotion and one that requires that many more conditions be met. The reactive attitudes are, ultimately, a 50 rather idiosyncratic subset of our rich emotional repertoire and abolition does not entail the bleak featureless emotional landscape that its opponents sometimes envisage.
In this section, I have argued that there is good reason to believe that we can abandon the reactive attitudes. There is good evidence that we can prevent and eliminate individual instances of reactive attitudes as well as reactive dispositions. Moreover, there is good reason to believe that, even if a person cannot eliminate all of her existing reactive dispositions, we can nonetheless prevent the development of such dispositions by intervening in their development, as in our imaginary cases. Finally, we should remember that, even if reactive dispositions cannot be either eliminated or prevented, we can nonetheless disavow those attitudes that arise as a result of undesirable dispositions.
Thus, while more may need to be said in defense of the possibility of full abolitionism, none of the objections considered thus far give us reason to conclude that it is impossible and there is good, albeit not decisive, reason to believe that it is possible.
2.3 Alternatives to Abolition
In the previous section, I argued that abandoning the reactive attitudes is possible.
If this is right, and if no one is either responsible, then the moral and epistemological worries raised by the abolitionist are good reasons to pursue abolition. However, just how pressing abolition is depends on whether there are alternatives that also respond to these concerns. Abolitionism is driven by the worry that, if we are neither free nor responsible, then we both get the facts wrong and wrongfully blame and praise individuals by taking reactive attitudes toward them. However, abolitionism may not be the only way to avoid such problems. Consider the following two alternatives. 51
First, one might argue that we ought to abolish only the negative reactive attitudes. The positive attitudes are on the same shaky epistemic ground as negative attitudes, but the moral case for abandoning them is much less compelling because wrongfully harming a person (by blaming her) is more objectionable than wrongfully benefitting her (by praising her). Thus, the upshots of partial abolition are that it allows us to maintain beneficial praising attitudes and to avoid the difficult task of abolishing the positive reactive attitudes. Moreover, a partial abolitionist position of this sort avoids the worry about interaction effects discussed earlier, namely, that the conditions necessary for preventing the development of (some) positive reactive dispositions preclude the conditions necessary for preventing the development of (some) negative reactive dispositions. Second, one could argue that the worries that motivate the abolitionist require only that we disavow, and not that we abandon, whatever reactive attitudes we experience and whatever reactive dispositions we possess. On this view, we are not required to eliminate or prevent ourselves from experiencing resentment or guilt so long as we disavow such emotions when they arise.
In response, I would first note that the reason for pursuing these two partial abolitionist alternatives depends greatly on whether we conclude that humans can achieve full abolition. If full abolition is impossible or prohibitively difficult, then both alternatives are appealing. If it is not, then there seems to be no reason to settle for partial abolition.
With regard to the first alternative, it is the positive attitudes that have, thus far, received the least defense from anti-abolitionists. Far more ink has been spilled in defense of resentment and indignation than forgiveness and gratitude. I suspect that the 52 reason for this is that the abolitionist case for replacing gratitude and forgiveness with non-reactive analogues is stronger, at least on its face, than the case for replacing resentment and indignation. However, the plausibility of these analogues provides a response to the worry that abolishing both positive and negative attitudes and dispositions is impossible. For the hypothesized interaction effect between, for example, Ungrateful and Misfortune would not exist if gratitude could be replaced without anything like the scenario in Ungrateful obtaining. In short, the challenge raised to full abolitionism by the possibility of interaction effects of abolishing positive and negative reactive attitudes is undermined by, and in direct proportion to, the existence of an effective, non-reactive replacement.
Finally, both alternatives ignore the strength of the epistemological reason for abolitionism, namely, that by taking reactive attitudes, whether positive or negative, we get the facts wrong. Unless we are actually unable to abandon both positive and negative attitudes, then both alternatives are unsatisfactory in this regard. It is not obvious how highly we do (or ought to) value getting things right, but some at least view it as a non- trivial good. For example, Robert Nozick claims that one lesson we should take away from his famous Experience Machine thought experiment is that we not only want to experience events, we “want them to be so” (1989, 106). We will have more to say about this in the next chapter, but for now it suffices to say that neither of the alternatives to abolition presently on offer considers the value of getting things right.
I conclude that, while there is support for both alternatives to full abolitionism, the reasons in favor of these alternatives rely for their force on whether full abolitionism 53 is possible in the first place. We ought, then, abandon the reactive attitudes to the greatest degree possible.
2.4 A World Without Reactive Attitudes
Suppose that we are able to abandon the reactive attitudes entirely and that there are neither reasonable alternatives that avoid the problems of attributing moral responsibility nor alternatives that are preferable to full abolitionism. How should we feel about our (purported) requirement to abandon these attitudes? Should we approach a life without reactive attitudes with trepidation, with optimism, or with indifference? In this final brief section, I consider what such a life would be like. In doing so, I comment on the scenarios imagined by various pessimists and optimists. I consider whether pessimists have exaggerated the implications of abolishing the reactive attitudes and whether optimist advocates for analogues have papered over the pessimists’ problems.
Many opponents of abolitionism imagine a world without the reactive attitudes as a grim place. Susan Wolf writes
We can see that the abandonment of all the reactive attitudes would make a very great difference indeed. To replace our reactive attitudes with the objective attitude completely is to change drastically—or, as most would say, reduce—the quality of our involvement or participation in all our human relationships (1981, 390).
It is such a world, so much bleaker and more barren than our present world, to which the pessimist [anti-abolitionist] fears the truth of determinism [given incompatibilism] would rationally force us (1981, 392). 54
I suspect that such a bleak outlook is unwarranted. But neither are we justified in ignoring, as some abolitionists seem to, the intuitions behind such foreboding fantasies.
In response to what he sees as misplaced pessimism, Sommers writes
When you take the objective attitude toward other human beings, you do nothing more than see them as natural things. But a human being is still a human being—the most exciting, infuriating, unpredictable, lovable, loathsome natural thing in the world. So, when we adopt the objective attitude…we choose friends as we choose human friends—that is all (2007, 326).
While I sympathize with Sommers’ resistance to such pessimism, I nonetheless think that he moves to quickly. I do not think that taking the objective stance precludes close friendships, but friendship may be different from the objective stance and it behooves us not to consider what those differences might be. For every conceivable world in which humans have transformed themselves into heartless robots by abandoning gratitude, forgiveness, and the resentment that forgiveness requires, there is another conceivable world in which war is abolished and mutual respect between all is the norm. Neither seems especially likely. The first scenario smacks of the same failure of imagination that suggests that committed consequentialists cannot form or maintain mature friendships.28
However, the second scenario also lacks imagination. For example, on many views, the mutual respect we all have (or ought to have) for one another is based on presumptive
28 Alastair Norcross (2012) argues, to the contrary, that, while there may be a tension between the various character traits of the committed consequentialist friend, such a friend may reasonably (and even rightly) judge that no alteration of these traits or commitments would be optimal. David Brooks recently made a similar argument, though it is not explicitly about consequentialists, in a New York Times op-ed called “The Way to Produce a Person”. There he claims that people who enter financially rewarding careers in order to give generously to effective charities are less likely to “cultivate a deep soul” than those who pursue careers and charity through more commonplace means and motives. 55 attributions of autonomy and/or rationality of a sort potentially precluded by the objective stance. Abolitionists often acknowledge that full abolition cannot be guaranteed, but they do not often attempt to imagine the complex emotional interrelationships upon which our various interpersonal relationships and engagements supervene.
Which end of the spectrum is more likely? One way is to identify, as best we can, the role that reactive attitudes play in maintaining social harmony and to assess both the specific benefits provided and the need for a reactive attitude, as opposed to some other emotion, to provide them. This is the task of chapter 4. Another way is to consider the role of the reactive attitudes in our most important personal relationships—i.e., our friendships and romantic relationships—and to assess whether such relationships are possible without them. This is the task of chapter 5. In these two chapters, I will argue that, while reactive attitudes perform an important function in mediating many of our social interactions, their benefits are not unique and they can be replaced. In particular, I argue that we can form and maintain important personal relationships without taking, or being disposed to take, reactive attitudes toward one another. On the basis of these arguments, I conclude that abolitionism is true and that, if we are neither free nor responsible, then we ought to abandon the reactive attitudes.
However, before addressing the purported outcomes of abandoning the reactive attitudes, I turn, in the next chapter, to a different kind of impossibility argument. In this chapter, I argued that abolishing the reactive attitudes is possible in sense that we are able to do it. In the next chapter, by contrast, I defend abolitionism against the claim that abandoning the reactive attitudes and taking the objective stance is self-defeating. This 56 objection states that taking the objective stance, while not strictly impossible, involves a performative contradiction.
Chapter 3
At Enmity with Unreality: Is Abolitionism Self-Defeating?
Those are the enviable people who live at enmity with unreality; and those are the
pitiable who are knocked on the head by the thing done without knowing or caring. So that when I ask you to earn money and have a room of your own, I am asking you to live in the presence of reality, an invigorating life, it would appear, whether one can impart it
or not.
- Virginia Woolf, A Room of One’s Own
Strong abolitionism, as I described it in the last chapter, is the view that if no one is responsible, we ought to abandon the reactive attitudes and adopt the objective stance toward other agents and their actions. Its opponents, then, hold that even if no one is responsible, we are still not required to abandon the reactive attitudes and adopt the objective stance. They deny that we must abandon these attitudes and argue that a purported lack of free will need not undermine our practice of using them to hold one another responsible. They argue that it would still be rational (and permissible) to take reactive attitudes toward one another, even if it is shown that determinism (or some other thesis) undermines free will. In this chapter, I focus on what I call the Peritrope
Argument against abandoning the reactive attitudes and adopting the objective stance. (A peritrope argument is an argument that attacks a claim as self-refuting.) Against the abolitionist, the proponent of the Peritrope Argument claims that even if human beings
57 58 are never free or responsible, this fact neither requires nor supports abandoning the reactive attitudes and taking the objective stance. The argument turns on the claim that the strong abolitionist’s reasons for doing so are self-refuting, which claim in turn relies on an account of the relation between taking the objective stance and acting for reasons.
Proponents of the Peritrope Argument make two main claims that I deny. First, the most plausible formulation of the argument relies upon what I claim is a false dichotomy, namely, that a person must take either reactive attitudes or objective attitudes—either a reactive or an objective stance—toward another’s attitude and actions.29 Whether we accept or deny this dichotomy obviously has implications for the plausibility of strong abolitionism. However, non-reactive attitudes are the topic of the next chapter, so I will not discuss them here. Second, the argument asserts that taking the objective stance is incompatible with living in accordance with the facts—i.e. it is incompatible with believing and acting in a way that is consistent with the facts about how the world is. In order to understand the purported incompatibility we must understand what the objective stance is and what it means to live in accordance with the facts. I will explain further what it means to live in accordance with the facts and how we ought to value this state of affairs when we come to the argument in section 4. We have already discussed, in chapter 1 section 1, what taking the objective stance might be thought to entail and the plausibility of the Peritrope Argument depends on whether we conceive of the objective stance as a “detached” stance or as the stance we take toward
29 Peter Strawson (1982) and Susan Wolf (1981) both seem to accept this dichotomy. R. Jay Wallace (1994), Derk Pereboom (2001), Shaun Nichols (2007), and others deny it and accept that our emotional repertoire contains non-reactive participant attitudes. 59
“natural objects” or in some other way. As we shall see in section 4, there is a conception of the objective stance, descended from Strawson, according to which the Peritrope
Argument succeeds, but others according to which it does not. For the time being, we need only recall that, according to Strawson and according to most accounts, the objective stance is a fitting response to cases in which non-responsibility is due to something like incapacity and not to situational factors, like acting by accident or under duress.
The Peritrope Argument is an impossibility argument and, as such, is very different from other anti-abolitionist arguments. Outcome arguments, as we shall see in chapters 4 and 5, dispute where and to what degree the “gains and losses to human life” accrue, whether to those who take reactive attitudes or those who abandon them.30 By contrast, the Peritrope Argument denies that we can live in accordance with the facts and
“get things right” whether we take the objective stance or not. Outcome arguments accept the epistemological and moral reasons cited by abolitionists as reasons but deny that they outweigh reasons against abolition, much less that they are decisive in favor of abolition. The Peritrope Argument denies that abandoning the reactive attitudes in order to live in accordance with the facts has any weight at all. In this way, the argument is a fitting subject for this third chapter. Having established, in chapter 2, that abolition is possible, this chapter considers an argument that purports to show that abolitionism is nonetheless self-defeating.
30 Talk of the “gains and losses to human life” comes from Strawson (1982, 70). 60
The present chapter is structured as follows. In section 1, I introduce the notion of a peritrope argument. I describe different forms that peritrope arguments can take and give examples. In section 2, I present and evaluate an Epicurean formulation of a peritrope argument against determinism and explain how this argument bears on the abolitionist debate. In section 3, I do the same for Norman Malcolm’s (1968) peritrope argument against mechanism. I suggest that both Epicurus and Malcolm’s arguments are predecessors to Wolf’s claim that taking the objective stance is self-defeating. In section
4, I introduce and consider two arguments that together make the case against strong abolitionism—Susan Wolf’s (1981) Dilemma Argument and Peritrope Argument—and note the similarities between them and the previous examples. Finally, I evaluate the
Peritrope Argument and assess the strong abolitionist position in light of the challenge it poses.31
3.1 Peritropes and Peritrope Arguments
A peritrope is a self-refutation. A peritrope argument is an argument that adduces as the grounds for its conclusion a self-refuting claim made by its opponent. A peritrope argument notes that an opposing claim or argument can be turned against itself.
(The noun peritrope means reversal; and the verb peritrepein means to turn around or turn over.) More specifically, a peritrope is a refutation in which “the thesis to be refuted
31 I am concerned in this chapter only with strong abolitionism, as it was defined in chapter 2. Unless explicitly noted, all references of abolitionism refer to strong abolitionism. 61 serves as a premise for its own refutation” (Burnyeat 1976, 48).32 I use the Greek term for the anti-abolitionist argument, rather than simply calling it a self-refutation argument, because the self-refuting character of the claim that we can take the objective stance resembles a purported peritropic thesis identified by Epicurus in Democritus’ view of free will and determinism. I will discuss Epicurus’ claim and its status as a predecessor to the anti-abolitionist argument in the next section.
Here are three types of peritrope and corresponding peritrope arguments that illustrate both the forms they can take and the contexts in which they tend to arise. All three are discussed by Burnyeat (1976) and found in the work of Sextus Empiricus, specifically his Outlines of Pyrrhonism and Against the Mathematicians. The first type of peritrope involves the claim that all appearances are true to reality, or accurately reflect the ways things really are in the world.
One cannot say that every appearance is true, because of its self-refutation [peritrope], as Democritus and Plato urged against Protagoras; for if every appearance is true, it will be true also, being in accordance with an appearance, that not every appearance is true, and thus it will become a falsehood that every appearance is true [Sextus Empiricus, M 7.389-90] (Burnyeat 1976, 47).
Protagoras claims that every appearance is true. This claim seems unobjectionable for some cases. For example, a rose appears red to Sally. If every appearance is true, then the rose is truly red. Fine. But other cases are problematic. It also appears to Sally that not every appearance is true, perhaps because she perceives a stick as bent when in water but perceives the same stick as straight when removed from the water. If every
32 For more discussion of the word and its philosophical origins and uses see Burnyeat 1976, especially pp. 47-48. 62 appearance is true, then not every appearance is true. It is clear that Protagoras’ claim is self-refuting. In this case, the peritrope argument—i.e., the argument attributing self- refutation to Protagoras—involves multiple premises.
P1. Every appearance is true.
P2. It appears that not every appearance is true.
C. Not every appearance is true.
The second type of peritrope is a claim that refutes itself without the addition of any other premises. For example, the claim that “there are no truths” can be understood as a peritrope because a contradiction follows directly from the claim (Sextus Empiricus, M
7.399 cited in Burnyeat 1976, 50).
P1. There are no truths.
C. It is true that there are no truths.
The third type of peritrope also comes from Sextus Empiricus.
If the Skeptic cites a reason why there is no such thing as a reason…he refutes himself/is reversed, and in the act of saying there is no such thing as a reason he lays it down that there is (PH 3.19; M 9.204) (Burnyeat 1976, 51).
This type of peritrope differs from the other two, as Burnyeat notes, in that the view is not “directly falsified by its own content,” but rather “it is the Skeptic’s undertaking to establish his thesis by reason that falsifies it, for his thesis is that there is no such enterprise to undertake”—i.e., no such enterprise as establishing a thesis by reason
(Burnyeat 1976, 51). The previous two types of peritrope made self-refuting statements.
This third type involves a performative contradiction or self-refutation. In order to 63 support the claim, its proponent must do something that conflicts with—e.g., that presupposes the contrary of—that very claim.
This third type of argument can be presented formally.
1. If the skeptic attempts to convince her opponent of a position, then the skeptic
assumes that she is offering reasons for her opponent to believe her.
2. If there is no such thing as a reason, then there are no reasons to offer for any
position.
3. If a person believes that there is no such thing as a reason, then she believes that
there are no reasons to offer for any position.
4. If a person attempts in good faith to convince another of a position, then she
believes that that position is correct (that its claims are true).
5. If the skeptic attempts in good faith to convince her opponent that there is no such
thing as a reason, then the skeptic both assumes that she is offering reasons to her
opponent and believes that there are no reasons to offer.
This reconstruction illustrates the form of the peritrope argument and the way in which its proponent establishes the performative contradiction. We will see this form again in peritrope arguments made against those who argue for determinism and mechanism.
The three examples above all concern truth and knowledge, but peritropes and peritrope arguments are not limited to epistemology. Peritrope arguments also have a long history in the free will debate. Let me now introduce some of these arguments.
64
3.2 Epicurus vs. Determinism
In this section and the next, I consider two peritrope arguments that are directly related to the Peritrope Argument against abolitionism, which I will consider in section 4.
The two arguments below, one made by Epicurus in his (mostly lost) work On Nature and the other by Norman Malcolm (1968), both suggest that determinism is a self- refuting thesis. In support of this claim each presents a peritrope argument, which concludes that the act of asserting or arguing for determinism is a performative contradiction. These related arguments provide insight into the anti-abolitionist argument presented in section 4 below and, in particular, to the nature of the commitment that I suggest is at the foundation of all three—i.e., Epicurus’ argument against determinism,
Malcolm’s argument against mechanism, and Wolf’s argument against abolitionism.
Epicurus makes a peritrope argument in the extant fragments of book 25 of his magnum opus On Nature.33
For if one were to attribute to admonishing and being admonished the mechanical necessity of what always on any occasion [happens to] affect oneself, one would never in this way come to an understanding [lacuna of a few words] by blaming or praising…But if one were to do this, one would be leaving the very action which, being in our power, creates the basic grasp of responsibility, and thereby in some respect having changed his doctrine [long lacuna, of 45 or 50 words] of such error. For this type of argument is upside-down and can prove that all things are like what are called ‘necessitated events’. But he quarrels about this very topic on the assumption that his opponent is responsible for being foolish. And if he [goes on] indefinitely saying again [and again], always on the basis of arguments, that he does this by necessity, he is not reasoning it out [properly] as long as he attributes to himself responsibility for reasoning well and to his opponent responsibility for reasoning badly. But if he were
33 The word Epicurus uses to describe the determinist argument is perichato, which David Sedley (1983) translates as ‘self-refuting’ and Brad Inwood (1994) translates as ‘upside-down’. 65
not to stop [attributing responsibility] for what he does to himself and [rather] to assign it to necessity, he would not…[lacuna of about 30 words].34
I reconstruct the argument contained in the above fragments as follows:
1. If one person attempts to convince another of her position, then she is committed
to the claim (that it is true) that he is capable of being convinced by an argument
and thus that he is responsible.
2. If determinism is true, then no one is responsible.
3. If a person is committed to determinism and determinism implies that no one is
responsible, then she is committed to the claim that no one is responsible.
4. If a person attempts in good faith to convince another of her position, then she is
committed to the claim that that position is correct (that its claims are true).
5. If a person in good faith attempts to convince another of determinism, then she
both is committed to the claim that he is responsible (because capable of being
convinced by an argument) and is committed to the claim that he is not
responsible (because determined).
By attempting to convince her opponent, the determinist commits herself to the denial of her own conclusion.
The form of the peritrope here is similar to the form in the third example described in the previous section. Just as a skeptic who gives a reason for thinking there are no reasons undermines his thesis, so does a determinist who makes an argument to
34 This translation is by Brad Inwood from the text prepared by David Sedley. It can be found in The Epicurus Reader translated and edited by Inwood (1994, 76). 66 her (purportedly non-responsible) opponent. Both the skeptic and the determinist have made a performative contradiction.
The connection between making an argument for determinism and (thereby) implicitly denying determinism is not as tight as the performative contradiction involved in giving a reason for believing that there are not reasons. The former requires further premises, including the premise that a person who is determined is not responsible, and the premise that making an argument to a person assumes that he is responsible.
Nonetheless, with the requisite additional premises, we have a clear performative contradiction and the peritrope argument is valid. However, I will argue that it is not sound.
Consider Epicurus’ argument. An opponent can avoid the accusation of performative contradiction in three ways. I call these three arguments the Compatibility
Objection, the Closure Objection, and the Interpretation Objection, respectively. First,
Epicurus’ opponent can dispute premise 2 and argue that determinism is compatible with responsibility. Second, his opponent can dispute the closure-type principle stated in premise 3. That is, the determinist can argue that a commitment to one claim C does not entail commitment to another claim C*, even if C entails C*. Third, his opponent can deny the connection, implicit in premise 1, between assuming that a person is capable of being convinced by an argument and assuming that she is responsible.
Let us start with the Compatibility Objection. Epicurus claims that determinism—in his words, “being necessitated”—is incompatible with responsibility.
The obvious response is simply to object to Epicurus’ argument that determinism and responsibility are compatible and propose one of the many contemporary compatibilist 67 views. The proposed view need not even hold that we are free and responsible, only that they are compatible. I think this is a reasonable objection to Epicurus’ peritrope argument. However, the debate about abolitionism concerns a conditional the antecedent of which is the denial of free will and responsibility, so Epicurus’ arguments may still e instructive even if we accept compatibilism. If determinism and responsibility are indeed incompatible, then an interesting obstacle faces those who advocate determinism.
Epicurus’ argument is also instructive insofar as it allows us to see a benefit of Wolf’s argument, which, as we shall see, does not necessarily presuppose incompatibilism.
Next consider what I have called the Closure Objection. Premise 3 of Epicurus’ argument states that, if a person is committed to determinism and determinism implies that no one is responsible, then she is committed to the claim that no one is responsible.
The plausibility of this premise turns on how we understand the status of “being committed to”. If being committed to C (where C entails C*) implies that one cannot or does not believe ~C*, then the premise is clearly false. For we can and often do believe things that are inconsistent with the entailments of our other beliefs. For example, it may be that if the Principle of Sufficient Reason is true, then a self-existent being must exist.
Nonetheless, the fact that I believe the Principle of Sufficient Reason to be true does not imply that I cannot or do not deny existence of a self-existent being. However, if being committed to C simply implies that, on the basis of one’s commitments, one ought to believe C*, then the premise seems plausible. If a person ought to believe in determinism, then, if determinism precludes responsibility, she ought also to believe that no one is responsible. Understood in the former, stronger sense, the premise is false and the peritrope argument is unsound. Understood in the latter, weaker sense, the premise is 68 true, and the determinist is faced with two mutually exclusive obligations—to believe that her interlocutor is responsible and not to believe that he is responsible. This is indeed a performative contradiction. However, it is weaker than it initially seemed and is not a decisive objection to the determinist. After all, we accept theories that yield exclusive obligations. Any moral theory that allows for genuine moral dilemmas is plausibly such a theory. And so is a scientific theory that, because incomplete, provides equally good reason to hold two mutually exclusive beliefs.
Finally, consider what I earlier called the Interpretation Objection. One might object to Epicurus’ first premise. It may be true that a person who attempts to convince another of some view is committed to the view that her interlocutor is capable of being convinced. However, it is not clear that believing that a person is capable of being convinced by an argument requires also believing that she is responsible. On most conceptions of responsibility, to judge someone to be a responsible agent, or responsible for a particular action, involves commitments beyond simply the ability to be convinced by an argument. For example, responsibility is often taken to require that the agent have some manner of control over her actions or beliefs, and perhaps that she be able to do or believe otherwise than she does.
One could respond that premise 1 is true on a thin conception of responsibility, whereby to be responsible is simply to be rational or responsive to reasons in this basic sense. On this view, a person is responsible so long as he can comprehend the content of his interlocutor’s claims and perceive that she is reasoning from one claim to the next.
However, if we apply this conception of responsibility in premise 1, we must also apply it to the remaining premises. But premise 2 is much less plausible on this thin 69 interpretation of responsibility. For it is not implausible that there are people who are determined, but whom we reasonably believe possess this sort of basic rationality. If so, then it seems that whatever conception of responsibility renders premise 1 most plausible, renders premise 2 much less plausible, and vice versa. However, this is not the last word on the matter. We will consider this particular issue in more detail when we consider
Wolf’s peritrope argument. It may be that Wolf finds a way around this objection both for herself and for Epicurus. Nonetheless, I defer this discussion until section 4.
3.3 Malcolm vs. Mechanism
Norman Malcolm (1968) makes a similar argument to that of Epicurus. The target of his argument is mechanism—which he takes to be a “special application of physical determinism” (1968, 127)—and mechanical explanations of human action.35 It should be noted that, while it is true that some mechanist views are also determinist, mechanism does not imply determinism, nor does determinism imply mechanism.
Mechanism does not imply determinism because it is possible for a mechanistic process to be indeterministic. Consider a toy example adapted from Daniel Dennett. A computer is connected to a Geiger counter that sends a signal to the computer only if a radioactive atom decays. The computer is programmed to initiate some process (say printing a test page) every time it receives a signal from the Geiger counter and to refrain from initiating the process if it does not. The computer is a strictly programmed mechanism, but its
35 The thesis that Malcolm calls mechanism should not be confused with the prominent but unrelated view in the philosophy of biology also called mechanism. See, for example, Machamer, Darden, and Craver (2000). 70 behavior, printing or not, is undetermined. Determinism does not imply mechanism simply because mechanism (and not determinism) is committed to the truth of determinism, but also to the truth of physicalism—where, for our purposes, physicalism is the thesis that “If something x, has some mental property M (or is in mental state M) at time t, then x is a material thing and x has M at t in virtue of the fact that x has at t some physical property P that realizes M in x at t” (Kim 1998, 74).
Malcolm summarizes his conclusion that mechanism is inconceivable in the following way.
We have uncovered two respects in which mechanism is not a conceivable doctrine. The first is that the occurrence of an act of asserting mechanism is inconsistent with mechanism’s being true. The second is that the asserting of mechanism implies that the one who makes the assertion cannot be making it on rational grounds (1968, 149).
I reconstruct his argument for this conclusion as follows:
1. Mechanical and purposive explanations are each complete and independent.
2. If a purposive and a mechanical explanation of an action A are each complete and
independent of one another, then each is a sufficient explanation of A.
3. If the mechanical explanation of A is sufficient, then the purposive explanation of
A is not necessary.36
4. If the purposive explanation of A is not required in order to explain A, then the
purposes averted to in the explanation (i.e., the agent’s intentions, etc.) are not
causes of A.37
36 Malcolm argues that if a comprehensive neurophysiological theory of human action were true, then “[desires and intentions] would not ever be sufficient causal conditions nor would they ever be necessary conditions” (1968, 136). 71
5. An intentional action is an action that is caused (at least in part) by an intention.
6. If a mechanical explanation of A is sufficient, then A is not an intentional action.
7. If mechanism is true, then every event, including human actions, can be given a
sufficient mechanical explanation.
8. If mechanism is true, then intentional action is impossible.
9. Arguing and asserting are always intentional actions.
10. If one argues for or asserts anything, then intentional action is possible.
11. If one argues for or asserts anything, then mechanism is false.
12. If one argues for or asserts the thesis of mechanism, then mechanism is false.
By arguing or asserting that mechanism is true, one commits oneself to the claim that mechanism is false.
Here too the form of the peritrope is similar to the performative contradiction, identified by Sextus Empiricus, made by the skeptic who gives reasons for believing that there are no reasons. Malcolm’s argument—in particular, the argument contained in premises 8-12—is also very similar to Epicurus’. While Epicurus claimed that determinism implies that no one is responsible, Malcolm holds that mechanism (a species of determinism) implies that no one acts intentionally or purposively (a requirement for responsibility). Malcolm avoids the objections I raised to Epicurus’ argument. His argument does not require closure like Epicurus’ third premise. Moreover, it does not
37 Malcolm summarizes the argument thus far, “Yet a mechanist explanation of behaviour rules out any explanation of it in terms of the agent’s intentions. If a comprehensive neurophysiological [i.e., mechanist] theory is true, then people’s intentions never are causal factors in behavior” (1968, 142). 72 presuppose incompatibilism because it is, in part, attempting to argue for incompatibilism.
However, Malcolm’s peritrope argument is also ultimately unsound. Malcolm sees the performative contradiction arising from conceptual claims about explanation as well as from the physicalist commitments of mechanism, which I will elaborate in a moment.38 Malcolm’s argument relies on two claims. He claims that mechanism implies that there are two complete and independent explanations of an action, one purposive and the other mechanical. And he claims that these explanations are incompatible or exclude one another. A purposive explanatory framework explains actions in terms of beliefs, desires, and intentions or purposes. A mechanical explanatory framework explains actions in terms of physical states and processes. However, Malcolm’s opponent can dispute this combination of views in four ways.39 First, his opponent can deny that the two explanations are independent and argue instead that the purposive is dependent on the mechanical (or vice versa).40 Second, she can deny that one or the other of the two explanations is complete. That is, she can argue that one explanation is incomplete and,
38 This may be true for Epicurus as well. David Sedley (1983) suggests that Epicurus is making an anti-mechanist argument of the same sort as Malcolm. A closer reading of Epicurus would consider this possibility in more detail. However, my interpretation takes Epicurus’ comments about necessity to be of primary importance in the passage quoted above. Moreover, there is nothing in the extant fragments to suggest anything like the explanatory exclusion worries that drive Malcolm’s argument. I do not deny that Epicurus could have been an anti-mechanist, but I will treat him as an anti-determinist. 39 This summary of the logical space available to Malcolm’s opponent can be found in Kim (1989). 40 It is true that the abolitionist can undermine Malcolm’s argument by claiming that mechanical explanations are dependent on purposive explanations. However, I will ignore this alternative because Malcolm’s mechanist opponent presumably views the purposive as dependent on the mechanical. 73 therefore, does not exclude the other, complementary explanation. Third, she can argue that, while purposive and mechanical explanations appear to have the same explananda
(e.g., an action A) and thus appear to exclude one another, in fact they always have different explananda. That is, what appears to be action A described in purposive terms and action A described in mechanical terms are actually two different events, A and A*.
Fourth, she can simply deny that two complete and independent explanations exclude one another.
I will take these in reverse order. Let me begin with a brief note about the fourth objection, the denial of explanatory exclusion. Jaegwon Kim (1989 and 1998) argues in favor of explanatory exclusion of the sort Malcolm requires and for the sake of brevity I will take this view for granted. Nonetheless, it is worth noting that instances of overdetermination provide putative counterexamples to, or at least difficult cases for, the explanatory exclusion thesis. For example, suppose two assassins simultaneously hit their target with a fatal gunshot, perhaps one in the heart and one in the brain. Each of the causal sequences they initiated has equal claim to be called a complete and independent explanation of the victim’s death.41
Let us turn, then, to the other three objections. It may be that a mechanical and a purposive explanation do not exclude one another. There are three ways in which these two types of explanation could be compatible. First, contrary to appearances, the two types of explanation could have different explananda. A purposive explanation might
41 See Kim (1989, 91-2) for his discussion of the relevance of overdetermination cases for the mechanism debate. See also Alvin Goldman (1969) and Daniel Dennett (1973) for objections to Malcolm’s initial argument and particular to the explanatory exclusion thesis. 74 explain phenomenon B (Sally buying a sandwich), while a mechanical explanation explains phenomenon B* (a particular set of bodily movements). One might argue that B is not the same phenomenon as B* and that the two explanations therefore explain different things. Whether or not such a story is plausible, however, it seems clear that neither the mechanist nor the purposivist would characterize their aims in this way. Both parties take themselves to be explaining the same events, not to be explaining ontologically separate realms of phenomena. In this case, both are offering explanations of Sally’s purchase of a sandwich. Taking the two explanatory frameworks seriously requires foregoing this objection to Malcolm. So this objection is inadequate
Second, the purposive explanation could be incomplete.42 For example, it might only be possible to give partial explanations of human actions (like B) in terms of agent’s beliefs, desires, and intentions. Explaining Sally’s purchase of a sandwich in terms of her beliefs, desires, and intentions might inevitably leave part of the phenomenon unexplained. Actions are, after all, quite complex. If the question is, “Why did Sally buy a sandwich?” a suitable answer is, “Because she was hungry.” However, a suitable purposive explanation may not be available for a more specific question. “Why did Sally buy that very sandwich rather than another of the seemingly identical sandwiches available?” It may be that at a certain level of specificity, purposive explanations of a given action are unavailable. For example, the answer to the above question might be something like, “Sally bought that sandwich because it was in the upper right of her
42 Malcolm’s peritrope argument would be undermined if either of the explanations is incomplete. I discuss the possibility of an incomplete purposive explanation simply because this is part of the mechanist thesis that Malcolm is challenging. Malcolm does the same thing in his paper. 75 visual field and human visual processing contains a bias toward items in the upper right of the visual field.” If this is the claim, and if all actions can be conceived in such specific terms, then it seems quite plausible that purposive explanations of an action are inevitably incomplete. If this is true, then there is room for mechanical explanations to do unique explanatory work and Malcolm’s worry about explanatory exclusion is misguided and his peritrope argument unsound.
Third, the purposive explanation could be dependent upon the mechanical explanation. The most obvious case of dependence would be the case of mind-body identity. If every belief, desire, and intention is identical to some neuronal state (or other microphysical state), then every explanation of a human action in these purposive terms is dependent upon the corresponding explanation in neurophysiological terms. If such dependence were the case, then the purposive explanation, far from excluding the mechanical, would require it. So, is the purposive explanatory framework dependent on the mechanical? This seems likely. If any reductive physicalist view of the mind and body is true, then purposive explanations are dependent on mechanical explanations of some sort (whether neurophysiological or otherwise). Purposive explanations are also dependent upon mechanical explanations on some non-reductionist views, including epiphenomenalism, though not on others.43
If any of the last three objections is correct, then Malcolm’s argument is unsound.
In particular, premise 1, and therefore premise 8—that if mechanism is true, intentional
43 Epiphenomenalism holds that physical events cause all mental events, but that mental events cannot cause physical events (Kim 1996, 129). On this view, any explanation given in terms of mental events (e.g., purposive explanations) will be dependent upon an explanation in terms of physical events (e.g., neurophysiological explanations). 76 action is impossible—is false. The third objection seems especially plausible.
(Moreover, as I mentioned, we might deny the thesis of explanatory exclusion itself.)
Thus, while his argument may be valid, it does not demonstrate that mechanism is self- refuting. While nothing I have said above constitutes a decisive refutation of Malcolm’s peritrope argument, a very plausible case has been made against the claim that mechanical and purposive explanations are complete and independent of one another, and, therefore, that they are incompatible with one another. Those who defend the purposive explanatory framework as both complete and independent appear to be committed to at least one implausible view about the relation of mental to physical states.
3.4 The Peritrope Argument against Abolitionism
In this section, I present a contemporary formulation of a peritrope argument offered by Susan Wolf (1981) against the abolitionist view that if we are not free, then we ought to abandon the reactive attitudes and adopt the objective stance.44 Wolf’s main argument is not itself a peritrope argument, but depends upon one that is similar in form to the two we have seen.
However, before considering the merits and deficiencies of this particular peritrope argument, it is important to clarify as much as possible what is at issue. Wolf argues that abandoning the reactive attitudes and adopting the objective stance in order to live in accordance with the facts is a self-defeating enterprise. Living in accordance with
44 What follows is a reconstruction of Wolf’s argument. I think it is charitable and consistent with her apparent intentions in the article. However, even if it does diverge from her intentions, it remains an important argument and is worth considering as an arrow in the anti-abolitionist quiver. 77 the facts means “getting things right” about the world. Wolf herself has little to say about what it means to live in accordance with the facts. She says in a footnote that living in discord with the facts does not entail logical inconsistency. So, one might have two logically consistent beliefs, but nonetheless be living in discord with the facts if one or both of these beliefs cannot be justified according one’s own accepted standards of justification (1981, 393n1). For example, I might be living in discord with the facts if I believed that communicative extraterrestrial life exists despite having nothing but the uncertain probability given by the Drake Equation.45
Wolf claims that taking reactive attitudes toward one another despite believing that no one is free or responsible is an instance of living in discord with the facts. She says, “If, despite the fact that this is our status, we choose to retain our reactive attitudes, we choose to live as if we were a kind of being that we know we are not. In doing this, we choose something akin to self-deception” (1981, 393). Wolf typically refers to the choice to, for example, take or not take reactive attitudes. Presumably, however, one can live in discord with the facts without explicitly choosing to do so. I may live in discord with the fact (if it is one) that there is no communicative extraterrestrial life even if I am unaware either that there is no such life or that my beliefs do not meet my own standards of justification.
The relevance of the state of affairs that Wolf dubs “living in accordance with the facts” is that she thinks it is quite valuable. That is, she takes the abolitionist’s
45 The Drake Equation calculates the probability of the existence of a communicative extraterrestrial civilization given a number of other probabilities. However, many of the component probabilities cannot be known with any confidence at all—e.g., the fraction of civilizations that actually produce detectable interstellar transmissions. 78 epistemological reason to be a strong reason. Again, however, she does not say explicitly why she thinks this. While acknowledging that the value of so living can be overridden, she states that doing so, “is to choose to leave a value unrealized, a value, moreover, that is of considerable depth and importance. To choose to act against, or contradict, a value as deep as this one, is inevitably to suffer a significant loss” (1981, 393). I think that the best way to understand this value is by reference to Robert Nozick’s Experience Machine case. The person in the Experience Machine is living in discord with the facts, particularly the fact that she is hooked up to a machine that generates all of her experiences. Almost none of her experiences are veridical. She is not writing a book, eating a sandwich, playing tennis, or wandering through the Musée D’Orsay. The person not hooked up to the machine is, at least to some degree, living in accordance with such facts. When she believes that she ate a sandwich, she is almost always correct. She is, more than her wired counterpart, “getting things right” about the world. Nozick describes the value we find in getting things right as follows:
Few upon hearing this description [of thorough and systematic deception] would exclaim, “What a wonderful life! It feels so happy and pleasurable from the inside.” That person is living in a dream world, taking pleasure in things aren’t so. What he wants, though, is not merely to take pleasure in them; he wants them to be so. He values their being that way, and he takes pleasure in them because he thinks they are that way. He doesn’t take pleasure merely in thinking they are (1989, 106).
Presumably something like this is what Wolf has in mind when she says that, “we place a primitive and unanalysable value on ‘getting things right’” and that, “From this value, the value of living in accordance with the facts would follow as a direct corollary” (1981,
401). Most important for my purposes is that the state of living in accordance with the 79 facts is what the abolitionist aims to secure by taking abandoning the reactive attitudes and taking the objective stance.46
Abolitionists claim that living in accordance with the fact that no one is free or responsible is good, that it can be achieved, and that one can benefit from it by abandoning the reactive attitudes. Proponents of the Peritrope Argument may (or may not) agree that living in accordance with the facts is good, but disagree (in either case) that this state of affairs can be achieved by abandoning the reactive attitudes. That is, they deny that abandoning the reactive attitudes and taking the objective stance is a means to the end of living in accordance with the facts. They do not see the objective stance as a means to achieve the agreed upon end. (It is thus consistent for the proponent of the Peritrope Argument to accept abolitionism for a different reason. For example, they may agree that avoiding wrongful blaming via reactive attitudes is sufficient to establish the abolitionist position. However, this side of the abolitionist debate is not my concern here.)
46 Interestingly, Nozick also has a view about the value of emotions that could be taken to bear on the value of taking or not taking reactive attitudes in world where no one is free or responsible. He writes, “Emotions provide a kind of picture of value, I think. Something’s being valuable involves its having a certain mode of structural organization to a certain degree—for example, a degree of organic unity; the responding emotion would be a psychophysical entity with a similar or parallel mode of organization…An intense emotion that is fitting [i.e. justified] is a close response to particular value, and is valuable in itself…[The] combination of emotion in relation to value gives us a further integrated structure, added to the integrated structure of the value itself. If such additional integrated structures count as valuable—as I think they do—that gives us a second value. So it is a valuable thing that there exist fitting [i.e. justified] positive emotions” (1989, 93-4). In short, the value of emotions depends, in part, on whether they are justified and not only on their usefulness or effectiveness in their natural cognitive role. 80
The Peritrope Argument against abolitionism is only concerned with the above described value of living in accordance with the facts. Whether a human being can be made better off in hedonic terms (or in some other way) by living in accordance with the facts is not at issue. Thus, for example, it is not a valid objection, and the defender of the abolitionist argument need not deny, that a person who forswears guilt will benefit from feeling fewer negative emotions. What is at issue, at least for the present formulation of the argument, is whether living in accordance with the facts secures this unique value for a person who takes the objective stance.
Let me turn now to the argument itself. Wolf appears to conceive of her anti- abolitionist argument as a dilemma. She argues that we value believing and acting in accordance with the facts about the way the world is (1981, 393). If we are not free and responsible, then if we continue to take reactive attitudes, we forego the benefit of living in accordance with the facts because we act in discord with the fact that we are neither free nor responsible. Taking a reactive attitude presupposes that its target is responsible, which is by hypothesis false. However, if we abandon the reactive attitudes and take the objective stance, we still forego the benefit of living in accordance with the facts because we still act in discord with the fact that we are neither free nor responsible. Taking the objective stance would appear to avoid an illicit presupposition of responsibility, but
Wolf claims that this appearance is deceiving (1981, 398). She argues that taking the objective stance in order to benefit from living in accordance with the facts is self- defeating.
I understand Wolf’s main argument as follows. I will call it the Dilemma
Argument. 81
1. We must either take the objective stance or not.
2. If we do not take the objective stance, then we do not live in accordance with the
facts.
3. If we do take the objective stance, then we do not live in accordance with the
facts.
4. Therefore, we cannot live in accordance with the facts.
As I mentioned above, this is not itself a peritrope argument. Rather, in the Dilemma
Argument, the peritropic element is what grounds the counterintuitive premise 3. As I will show, this premise follows from a peritrope argument which concludes that the abolitionist position is self-refuting in the performative sense, though not in the sense that its content implies a contradiction. The rationale for taking the objective stance is to live in accordance with the facts (and benefit from doing so), but, according to Wolf, taking the objective stance purportedly does not allow one to do so and indeed may make doing so impossible.
I note but do not comment here on the fact that premise 2 is true only if we assume that the objective stance is the only stance one can take that does not presuppose responsibility. If there is a third type of attitude, besides reactive and objective, that does not presuppose that its target is responsible (or make any other false claims), then not taking the objective stance does not entail taking a reactive attitude and does not entail living in discord with the facts. It is clear that Wolf, following P.F. Strawson, assumes the reactive/objective dichotomy, though she does not explicitly claim that not taking the objective stance in response to one’s own and others’ actions implies taking reactive attitudes towards them. The important premise for my present purposes is the third. 82
Consider premise 3. Wolf suggests that taking the objective stance in order to live in accordance with the facts involves a performative contradiction (1981, 399).
Taking the objective stance for this reason commits one to a claim that is inconsistent with the reason one gives for taking it. A person who takes the objective stance in order to live in accordance with the facts thereby presupposes that she is responsible for doing so and thereby fails to live in accordance with the facts.
We can see that the Peritrope Argument behind premise 3 is parallel to the arguments made by Epicurus and Malcolm.
i. If a person takes the objective stance, then that person is committed to the claim
that no one is responsible.
ii. If a person acts for reasons, then that person is committed to the claim that
someone (at least herself) is responsible.47
iii. Taking the objective stance is an action that is done for reasons.
iv. Therefore, if a person takes the objective stance, then that person is committed
both to the claim that no one is responsible and to the claim that someone (at least
herself) is responsible.
It seems, then, that taking the objective stance involves a performative contradiction. The goal of taking the objective stance is to live in accordance with the facts, but this goal
47 Wolf makes this point through the example of a completely programmed robot, “For the robot, in taking an objective view of himself, necessarily leaves a part of himself out of this view—specifically, he leaves out the part of himself which is taking the objective attitude.” She makes a similar point about a determined (and unfree) human agent, “In taking any attitude toward ourselves, including the attitude that we are not free or responsible beings, we would be asserting ourselves as free and responsible beings” (Wolf 1981, 398-399). 83 cannot be achieved if any attempt to do so requires contradicting oneself. It is the above
Peritrope Argument and this fact that together purport to justify the otherwise counterintuitive premise 3—that if one takes the objective stance, then one does not live in accordance with the facts.
It is worth noting that if either of Wolf’s conclusions is true—either the conclusion of the Dilemma Argument or the conclusion of the Peritrope Argument— abolition is challenged. The conclusion of the Dilemma Argument undermines the epistemological reason for abolitionism, i.e., that we can avoid getting things wrong by abandoning the reactive attitudes and taking the objective stance. And the conclusion of the Peritrope Argument undermines the possibility that one could consistently adopt a wholly objective stance—though one could inconsistently do so.
The anti-abolitionist, via the Peritrope Argument, holds that we need not instantiate the impoverished world that, on her view, would exist if we all abandoned the reactive attitudes because taking the objective stance for this reason commits one to a claim that is inconsistent with the reason one gives for taking it. The cost of not adopting the objective stance in a world where we are neither free nor responsible is that we must abandon any hope of “getting things right”—of living in accordance with the facts— which is itself a significant sacrifice (Wolf 1981, 401-402). However, the cost of adopting the objective stance is the same. Living in accordance with the facts is impossible for rational beings like us, if we are neither free nor responsible. Together these arguments strike at one of the foundations of the abolitionist position. By itself, it does not offer reasons that compete with reasons for taking the objective stance or dispute the weight of such reasons. Rather, it denies that a purported reason for taking the 84 objective stance—i.e., the epistemological reason we have to avoid commitment to false claims—is a reason that one can consistently act on.48 The Peritrope Argument is a purely negative argument. It does not offer any reason for retaining the reactive attitudes.49 The argument denies that we have any robust reason, grounded in the value of living in accordance with the facts, for taking the objective stance because the value of living in accordance with the facts eludes us whether or not we take the objective stance.
If we take reactive attitudes, then we straightforwardly fail to live in accordance with the facts and thereby fail to gain any value from doing so. If we take the objective stance, then we also fail to live in accordance with the facts. This threatens the abolitionist position not by supporting the anti-abolitionist, but by undermining one of the two reasons cited by the abolitionist in favor of abandoning the reactive attitudes and thereby shifting the burden of proof (at least in part) back to the abolitionist.
But is the Peritrope Argument sound? I think it is not.
Epicurus and Malcolm offer peritrope arguments against determinism and mechanism, respectively, which fail for different reasons. However, both are relevant to
Wolf’s Peritrope Argument insofar as each provides a model for understanding the purportedly self-refuting nature of commitments to both non-responsibility and various forms of intentional action, particularly arguing and asserting. Wolf’s argument develops, whether purposefully or not, a weaker thesis than Epicurus and takes a different
48 Wolf overstates this point in her discussion, claiming that none of the purported reasons for taking the objective stance is a reason at all (1981, 393-4). 49 Wolf does make other positive arguments. For example, she follows Strawson in suggesting that a world in which no one ever took reactive attitudes would be bleak and impoverished (1981, 390-1). 85 tack than Malcolm. Her claim is about non-responsibility rather than determinism and about attitudes rather than purposive mental states generally.
Regarding the Dilemma Argument, premise 3 is quite counterintuitive. However, if it is false, it will be because Wolf’s Peritrope Argument is unsound. This argument avoids the weaknesses from which Epicurus and Malcolm’s peritrope arguments suffered. It does not require incompatibilism, or any claims about the nature of mental events, or explanatory exclusion. However, on its face, it appears to rely on ambiguity about what is meant by responsibility.
Consider Wolf’s dilemma. This argument is part of a larger argument against abandoning the reactive attitudes. More specifically, it is part of a response to the following question: If we are not responsible agents, are we required to abandon our reactive attitudes? The dilemma supports the claim that we are not so required. It is most natural to understand this question as about moral responsibility in particular. Moral responsibility is typically what is taken to be at issue when discussing the reactive attitudes. And indeed Wolf seems to have moral responsibility in mind throughout her discussion. For example, she emphasizes consequences of abandoning the reactive attitudes for moral virtues like courage and charity; she speculates on how we would react to injustice and cruelty having abandoned the blaming attitudes; and the targets of the attitudes she describes are thieves and murderers (Wolf 1981, 390-391).
However, looking at the Peritrope Argument, we can interpret the term responsibility in one of three ways. First, because we think that the dilemma is about moral responsibility, we might think that the Peritrope Argument must also be interpreted as about moral responsibility. Premise (i) of the Peritrope Argument is plausible on this 86 interpretation because it is part of what it means to take the objective stance that one does not hold others morally responsible as praiseworthy or blameworthy. But premise (ii) is not plausible on this interpretation. Suppose that premise (ii) was about moral responsibility and call this revised premise (ii*).
ii*. If a person acts for reasons, then that person is committed to the claim that
someone (at least herself) is morally responsible.
This premise is subject to numerous counterexamples. Normal adults routinely act solely on nonmoral reasons. For example, Raphael buys a pizza (action) because he is hungry
(nonmoral reason). On even the weakest conception of “commitment”, it seems clearly false that this action commits Raphael to the claim that he (or anyone else) is morally responsible. Moreover, there may be those who never act on moral reasons, because they lack the capacity to either recognize or react to them, but who nonetheless act for reasons.
Psychopaths are sometimes described in this way (Talbert 2012).50 While such people are not responsive to moral reasons, and are thus not held morally responsible, they are acknowledged to act for nonmoral reasons. Any of these cases, normal or abnormal, if accurately described, is a counterexample to premise (ii*). I conclude, then, that Wolf’s combination of Dilemma and Peritrope Argument fails when the responsibility at issue is
50 However, David Brink has argued that, whatever deficits psychopaths have, “these deficits do not prevent the formation of moral concepts and the ability to recognize moral and criminal norms” (2013, 33). In addition, he argues, the “common and significant failings of psychopaths to conform [their behavior to reasons they recognize] do not automatically show that they lack volitional competence” (34). If Brink is correct in this assessment, then psychopaths are not a useful example, here. However, the incorrect description of psychopaths would still provide a useful description of a type of deficit relevant to the current discussion. 87 understood only as moral responsibility. Premise (ii*) is false and this formulation of her
Peritrope Argument is unsound.
It is equally unsound if we interpret the argument as being solely about epistemic responsibility—i.e. responsibility for believing and reasoning justifiably or unjustifiably.
For if we interpret premise (ii) as referring to epistemic responsibility, similar counterexamples apply.
ii**. If a person acts for reasons, then that person is committed to the claim that
someone (at least herself) is epistemically responsible.
The fact that Raphael bought a pizza because he was hungry no more implies that he holds someone epistemically responsible, than that he holds someone morally responsible.
Finally, consider a third interpretation of responsibility in premise (ii), according to which to say that a person is responsible is to say that she is rational in the basic sense of being able to recognize and act for reasons.51 On this view, it is true that by buying a pizza because he is hungry, Raphael is committed to the claim that someone (at least himself) is responsible. By acting on his reason for buying a pizza, Raphael plausibly affirms his own basic rationality. Indeed, on this interpretation of responsibility, premise
(ii) is almost tautological. If a person acts for reasons, then that person is committed to the claim that someone (at least herself) is reasons-responsive, i.e., able to recognize and act for reasons.
51 Thanks to David Brink for suggesting this alternative interpretation. 88
The question, then, is whether the rest of the premises in Wolf’s Peritrope
Argument remain true on this interpretation. In particular, we might ask whether the relevant interpretation of premise (i), call it (i^), is true.
i^. If a person takes the objective stance, then that person is committed to the claim
that no one is reasons-responsive.
At first glance this looks false. For we might think that, while we take the objective stance in response to non-responsible actors, nonetheless not all non-responsible actors are non-reasons-responsive. After all, many conceptions of moral responsibility require that other conditions be met besides reasons-responsiveness (e.g., Kane 1996, Pereboom
2001). However, a closer look reveals a more serious objection and a possible reason to believe that premise (i^) remains true on this interpretation of responsibility.
Suppose that we take the objective stance because we believe that both incompatibilism and determinism are true and that, therefore, no one is free or responsible. On some views the combination of incompatibilism and determinism undermines responsibility because it undermines basic reasons-responsiveness. It precludes the ability to reason in any but the way one does reason, which may not be sufficiently responsive for any form of responsibility. If lack of reasons-responsiveness is the reason for taking the objective stance, then we might, merely by taking this stance, commit ourselves to the view that no one is reasons-responsive.
Now, it certainly is true that whenever we take the objective stance toward a person, we do so because we believe that she is not responsible. However, is it also true that whenever we take the objective stance toward a person, we do so because we believe that she is not reasons-responsive? Strawson sometimes seems to think so. He states, “If 89 your attitude toward someone is wholly objective, then though you may fight him, you cannot quarrel with him, and though you may talk to him, even negotiate with him, you cannot reason with him” (1982, 66). And it is true that sometimes when we take the objective stance toward a person, we do so because we believe that he is not reasons- responsive. For example, Strawson rightly notes that we take the objective attitude toward those whose behavior is purely compulsive. Such behavior is almost by definition not reasons-responsive. Moreover, it is not sufficient for taking the objective stance that we judge the person not to be responsible. Not all non-responsible beings are proper targets of the objective stance. We do not believe that those who act under duress are responsible, but we do believe them to be reasons-responsive. Indeed, it is only useful to hold someone under duress if you believe the person to be reasons-responsive. And this is precisely why Strawson says of duress and other similar situations that, “None of them invites us to suspend toward the agent, either at the time of his action or in general, our ordinary reactive attitudes” (1982, 64). We might argue that some cases of non- responsibility do not warrant the objective stance precisely because the agent is still reasons-responsive.
I agree that if we accept Strawson’s conception of the objective stance and his account of when and why we take it, and if we accept a view according to which determinism and incompatibilism preclude reasons-responsiveness, then we must also accept that taking the objective stance commits one to denying reasons-responsiveness. I agree, therefore, that, on this view of the objective stance, Wolf’s Peritrope Argument is sound and premise 3 of her Dilemma Argument is supported. It thus appears that strong abolitionism may indeed be self-defeating and that the Peritrope Argument provides 90 some reason for taking seriously the bleak conception of a world without reactive attitudes that we considered in the previous chapter. And, at the same time, it provides some reason to think that Sommers’ claim that such pessimism is unjustified was itself a hasty assessment.
However, three points remain relevant for the abolitionist. First, the abolitionist may deny that determinism and incompatibilism together preclude responsibility by precluding reasons-responsiveness. For example, this is not held to be an implication of these two theses by either Galen Strawson (1994) or Derk Pereboom (2001). And Saul
Smilansky’s illusionist position—that no one is free or responsible, but that we ought to live under the illusion that we are (2000, 6)—seems to presuppose that we are reasons- responsive. Thus, at most, this objection identifies one form of skepticism about freedom and responsibility according to which strong abolitionism is impossible in the sense of entailing a performative contradiction.
Second, as I mentioned above in my brief discussion of the Dilemma Argument, premise 2 is arguably false. (Recall, premise 2 says: If we do not take the objective stance, then we do not live in accordance with the facts.) While it is not immediately counterintuitive in the way that premise 3 is, premise 2 does implausibly assume that the objective stance is the only alternative to taking reactive attitudes. One either takes the reactive stance or the objective stance. However, many abolitionists think that there is a middle ground, that we ought to abandon reactive attitudes, but replace them with non- reactive participant attitudes (e.g., Pereboom 2001). Even non-abolitionists accept that such a middle class of attitudes exists (e.g., Wallace 1994). Denying premise 2 allows the abolitionist to retain the force of the epistemological reason for abolitionism. That is, 91 we can still argue that we ought to abandon the reactive attitudes in order to live in accordance with the facts. However, this concession potentially forces the strong abolitionist to retreat to the less radical abolitionist position that does not require adopting the objective stance.
Third, the strong abolitionist can argue that Strawson’s conception of the objective stance, and of when and why we take it (and ought to take it), is mistaken, perhaps because it is infected by the same false dichotomy between reactive and objective stances that undermines premise 2 of the Dilemma Argument. If a good argument can be made for abandoning Strawson’s conception of the objective stance, then we would have to reconsider Wolf’s Peritrope Argument and, thus, the fate of strong abolitionism. The plausibility of Strawson’s conception is brought into question by the arguments of the following two chapters.
3.5 Conclusion
The strong abolitionist argues that, if we are neither free nor responsible, then we ought to abandon the reactive attitudes and adopt the objective attitude. In this chapter, I have considered an argument against strong abolitionism, which charges that it is self- defeating. I have concluded that there is a form of skepticism about responsibility and a conception of the objective stance which, when taken together, support Susan Wolf’s
Peritrope Argument. Her argument succeeds (under these assumptions) because, unlike other peritrope arguments before it, it establishes a genuine performative contradiction.
In this chapter, I considered three peritrope arguments about free will and responsibility—from Epicurus, Norman Malcolm, and Susan Wolf. Only one of these— 92
Wolf’s—directly challenges the abolitionist position, but the other two provide insight into the form of argument and the purported source of the tension that underwrites all three challenges. All three arguments attempt to establish a performative contradiction.
The source of the tension in each argument is something like the objective stance.
Despite the anachronism of this label in the case of Epicurus, I believe that it accurately describes how these three philosophers understand their respective opponents. The determinist, mechanist, and abolitionist each views the world from a perspective (the objective stance) which appears to be self-refuting. And Epicurus, Malcolm, and Wolf each view the objective stance as entailing a different set of commitments for their respective opponents. Epicurus takes acceptance of determinism to imply an objectivity of attitude on the part of the determinist toward his interlocutor. For Epicurus and Wolf, taking the objective stance seems to have implied nothing more than the denial that anyone is responsible, though they may have conceived of responsibility differently. For
Malcolm, the objective stance is forced upon a mechanist by a conjunction of theses including explanatory exclusion, and taking it seems to have implied a view of the world as entirely non-purposive. Thus, these three arguments, while not all concerned with abolitionism, do all bear on one source of anti-abolitionist concern, namely, the objective stance.
Having evaluated the challenge to the abolitionist position posed by the Peritrope
Argument, let me briefly mention two problems with this argument that are independent of my preceding objections to it. First, the argument does not address or affect the moral reason for abolition, namely, that we ought not blame the blameless. Whether or not taking the objective stance allows us to live in accordance with the facts, taking the 93 objective stance does allow us to refrain from wrongfully blaming those who are not blameworthy. This alone might be sufficient reason to abandon the reactive attitudes.
Second, even if the Peritrope Argument is sound, taking the objective stance is not self-refuting if one does not take it toward oneself. One can take the objective stance toward everyone’s actions and attitudes except one’s own and thereby live almost entirely in accordance with the facts and also avoid wrongly blaming others.
Chapter 4
Excluding the Forces of Ruin: A Defense of Analogue Attitudes
“It is easier to exclude the forces of ruin than to govern them, to deny them admission
than to moderate them afterwards.”
- Seneca, De Ira
We typically assume that others are free and responsible agents. This assumption grounds many of our interpersonal and social practices. For example, our system of criminal punishment assumes not only that it is useful to punish those who break the law but also that lawbreakers in some sense deserve punishment for freely choosing to do so.
Arguably, at least on retributivist justifications of punishment, this judgment of deservingness, and the practice of criminal punishment that acts upon it, presupposes moral responsibility. If no one is responsible, then this assumption is mistaken and social practices like criminal punishment that are based upon it are inappropriate.
Suppose that no one is free or responsible. If we wish to live in accordance with this fact, then it appears that we must act toward and conceive of one another differently, including adapting certain existing social practices to fit this reality. I contend that the fact (if it is one) that no one is either free or responsible requires that we revise our judgments and adapt our practices in such a way that they are consistent with this reality.
In this chapter and the next, I attempt to paint a more detailed picture of a world without the reactive attitudes and I defend abolitionism against objections concerning the
94 95 possibility and value of such a world. The present chapter focuses on the plausibility of replacing the reactive attitudes and the value of their proposed replacements. The next chapter, by contrast, defends abolitionism against the claim that valuable personal relationships are impossible if we abandon the reactive attitudes.
In this chapter, I defend abolitionism against an objection which states that reactive attitudes perform their psychological and social function more effectively than any mental states that might replace them. Shaun Nichols (2007) argues for unnoticed benefits of the reactive attitudes and especially for benefits that accrue independently of the fairness or unfairness of taking reactive attitudes toward non-responsible agents.52 He argues that the abolitionist proposal that we replace unjustified and unjust reactive attitudes with appropriate analogues is unworkable. In doing so, Nichols responds to
Derk Pereboom (2001) and Tamler Sommers (2007), who suggest that analogue attitudes53—emotions that perform similar functions to reactive attitudes, but which do not presuppose that their targets are responsible—can adequately perform the function of their reactive counterparts.54 For example, an emotion similar to the forgiving attitude
52 He does not ignore fairness entirely and comments briefly on the possibility that the unfairness of taking blaming attitudes toward non-responsible agents trumps any other benefits we might identify (2007, 422-424). 53 Strawson himself uses this term to refer to “vicarious analogues” of the reactive attitudes—i.e., those attitudes felt on behalf of a victim who is not oneself (1962, 70-1). 54 It is important to recall the important distinction between two types of abolitionism identified in chapter 2. Pereboom’s abolitionism holds that we ought to abandon the reactive attitudes and, where possible, replace them with analogue attitudes that do not presuppose moral responsibility. Sommers proposes a stronger form of abolitionism, however, which holds that we ought to abandon the reactive attitudes, replace them with analogues where possible, and also adopt the objective stance toward one another. In this chapter, I defend Pereboom’s proposal against Nichols’ largely empirically driven objections. In the next chapter, however, I defend the strong abolitionist proposal. 96 can promote reconciliation between victim and offender and can help provide therapeutic closure for the victim about the offense, while also not holding the offender responsible as genuine forgiveness does (2001, 201).
In section 1, I briefly introduce the debate between the abolitionist and the anti- abolitionist, including the abolitionists’ reasons for thinking we can live without the reactive attitudes. In section 2, I explain Nichols’ objections to the abolitionist proposal.
I will call this objection Social Harmony Argument. In section 3, I consider Pereboom’s responses to these objections and address further objections to Pereboom’s position.
Finally, in section 4, I provide an empirical and philosophical defense of abolitionism against Nichols’ argument. This defense itself has three parts. First, I argue that many of the analogue attitudes proposed by Pereboom and Sommers (e.g., regret) have stronger empirical support than either Nichols or Pereboom allow. Second, I argue that, even if
Nichols is right that abandoning the reactive attitudes would have negative effects (in the short term), this fact alone does not undermine the abolitionists’ claim that they are morally objectionable and ought to be abandoned. Finally, I argue that, if it succeeds,
Nichols’ argument proves more than he intends and that, taken together, its implications are implausible.
4.1 The Abolitionist Dialectic
P.F. Strawson (1962) describes reactive attitudes as those attitudes we take when we engage personally with other individuals. They are attitudes of engagement or participation. Strawson identifies resentment, indignation, guilt, forgiveness, gratitude and others. Derk Pereboom’s (2001) main contribution to the discussion of the reactive 97 attitudes is to suggest that non-reactive attitudes may play a similar role in our engagement with others—i.e., that reactive attitudes can, at least in some cases, be replaced by non-reactive analogues.55
Pereboom’s proposal is meant to dispel the concern, raised by Strawson and maintained by others, that abandoning the reactive attitudes would result in a world without genuine human society, a world characterized by an essential detachment between individuals. On the abolitionist view defended by Pereboom and Sommers, a reactive attitude may serve a particular purpose, but this purpose may not always require that particular reactive attitude. Our emotional repertoire is rich and varied enough to provide replacements that do not presuppose moral responsibility. Emotions are like faces. I may effectively, if petulantly, communicate my disdain for your soapbox derby racer by sticking out my tongue at you. However, I may communicate my disdain just as effectively, and just as petulantly, by raising my eyebrows and pouting my lips.
Similarly, the function of some reactive attitudes can be performed either by different reactive attitudes or by non-reactive attitudes. Moreover, while some functions may require reactive attitudes, some of those functions may not be necessary for a flourishing inter- and intrapersonal existence.
Pereboom and Sommers suggest a number of analogues to the reactive attitudes.
Negative analogues like moral sadness and regret are meant to replace responsibility- presupposing attitudes like resentment and indignation, on the one hand, and guilt, on the
55 As I have noted, Pereboom was not the first to suggest that there are non-reactive participant attitudes. R. Jay Wallace (1994) also argues that the reactive emotions are a subset of the class of participant emotions. However, he does not consider the possibility of replacing reactive attitudes with participant attitudes. 98 other. Similarly, forgiveness and gratitude, divested of their presupposition of responsibility, are meant to replace their reactive counterparts. Shaun Nichols (2007), however, contends that the imagined benefits of non-reactive replacements for the reactive attitudes are just that, imagined. These benefits are not supported by empirical evidence.
While the dialectic between the abolitionist and her opponent is relatively straightforward, the details of the debate are not. The rules of the game for evaluating an analogue and comparing its efficacy to that of a reactive attitude are not explicit, nor is it even clear whether there are implicit standards of evaluation beyond the utility of the emotion broadly conceived. I want to distinguish between two types of criticism one can make of analogues and two related types of defense one can give of them. One can criticize either how well the analogue performs its required function or its inability to provides various downstream benefits. For example, one can criticize moral sadness either for its inability to communicate protest or for failing to galvanize group resistance to oppression. Likewise, one can defend analogues along these same dimensions, either by showing that they do perform their identified function or by showing that the same downstream benefits do accrue.
4.2 Empirical and Normative Challenges to Abolitionism
Nichols (2007) does not deny that analogues could replace reactive attitudes for some purposes. Rather, he denies that these replacements would be as effective as their reactive counterparts. Nichols makes both an empirical and a normative claim about the relative effectiveness of Pereboom’s analogues. The empirical claim is that the analogue 99 attitudes do not in fact perform all the functions of their corresponding reactive attitudes.
The normative claim is that the failure of the analogues to perform these functions constitutes a significant loss to human life and that, for this reason, we ought not abandon or replace the reactive attitudes.
Evaluating each of these claims requires both a technical vocabulary and a clear understanding of the success (and failure) conditions for each argument, so let me address these questions first. Establishing the empirical claim requires demonstrating that there are reactive attitudes whose function cannot be performed by an analogue. Refuting it requires showing that they are replaceable. Establishing the normative claim requires demonstrating that reactive attitudes have benefits that their proposed analogues lack and that these benefits outweigh their moral costs. Refuting this claim requires showing that all analogues are as beneficial as their reactive counterparts, whether because they yield all the same benefits, because they yield a different but comparable set of benefits, or because the particular reactive attitude that an analogue replaces is itself not beneficial.
Understanding the two claims in this way, we can see that the easier the empirical claim is to refute, the harder the normative claim is refute. If many emotions can qualify as a genuine analogue—i.e., an emotion that performs the same function as some reactive attitude—then many of those emotions will not be effective analogues—i.e., they will not be as beneficial as their reactive counterparts. And vice versa. The stricter the conditions are for counting as a genuine analogue the more likely it is that a genuine analogue will be as beneficial as its reactive counterpart.
Let us turn briefly to Nichols’ technical vocabulary. Nichols treats the reactive attitudes as a special type of emotion. Understood as emotions, reactive attitudes have a 100 cognitive element (appraisal), an affective element (feeling)56, and an associated behavior or action (Kalat and Shiota 2007, 14-19). For example, the basic psychological profile of fear involves an appraisal (that is a dangerous spider), an affective response (the feeling of fear), and a resulting behavior (escape the spider). On Nichols’ view, reactive attitudes and their analogues have the same basic structure. Thus, for example, the resentment I feel at being robbed may involve an appraisal of the thief’s action as wrong, a feeling of anger toward her, and a tendency to retaliate against her (Nichols 2007, 413-
414).
We can understand the benefits of an attitude in terms of the behaviors it prompts.
Experiencing an emotion inclines one to act in a particular way. For example, when I am afraid of a spider, I am inclined to run away. Emotions may have many behaviors associated with them. For example, when I am angry with someone for lying to me, I may be inclined either to confront that person about her wrongdoing or to ignore her.
That B is the behavior associated with emotion E simply means that, for subject S, the presence of E significantly increases the likelihood that S will do B. Experiencing an emotion has all sorts of consequences, only some of which will be identifiable.
We can understand the function of an attitude as that set of benefits that it is particularly suited to provide. For example, we can say that the function of moral anger is to prevent, deter, and protest perceived wrongdoing by others. This is fine, so long as we understand the limits of the locution. We must keep in mind that our emotional
56 The affective component of an emotion should not be confused with the emotion itself, though the same words may, and sometimes must, be used to describe them. For example, the feeling of fear is different from emotion fear of which it is one component. 101 capacities have evolved like many other human abilities and that they provide not only the benefits that drive their evolutionary selectiveness but also other benefits (and costs) that are accidental and due, for example, to cultural or historical context. Experiencing an emotion can have beneficial consequences beyond those associated with what we would call its (evolutionary) function. And it is also true that emotions can have harmful side effects, as when moral anger facilitates vicious cycles of excessive retribution—as was the case in feuds like those between the Hatfields and the McCoys.
Let me turn now to Nichols’ empirical claim that the analogue attitudes do not perform the functions of their corresponding reactive attitudes. The relative effectiveness of reactive attitudes and their analogues is judged by reference to the behaviors that a given attitude purportedly produces. For it is one’s behavior—the motivation and the action, not the appraisal and the feeling—that contributes to a rich social existence. For example, if an appraisal of a dangerous snake as beautiful and a concurrent feeling of joy could produce the required motivation to flee, then the bizarre emotion constituted by these elements would have the same relative effectiveness as fear. Nichols’ criticism of
Pereboom’s proposed analogues is that they do not produce the useful behaviors identified as the function of the reactive attitudes.
Nichols focuses on the class of reactive attitudes of moral anger, which he takes to include resentment and indignation. He argues that moral anger is useful both for oneself and for others. Taking an attitude of moral anger toward an offender purportedly benefits the victim by signaling intolerance for wrongdoing and discouraging further mistreatment (Nichols 2007, 417). When a friend breaks her promise to me, expressing my anger communicates to her that she has done wrong, that I am upset with her, and that 102
I will be angry with her in the future if she breaks further promises. All of these messages are important for building and maintaining relationships with others who, realistically, will wrong us from time to time.
Moral anger purportedly benefits others (non-victims) in a similar way. Nichols draws on research in experimental economics that punishment, understood as an expression of moral anger, has a significant effect on the behavior of participants in public goods games—e.g., games in which participants can choose to invest in and jointly profit from a common fund or selfishly free ride on the investments of others. Various studies have shown that humans are willing to punish those who refuse to abide by the norms of such cooperative games, that the reason for their decision to punish is anger at a perceived injustice, and that the result of such punishment is greater compliance with the norms by free riders and would be free riders (Fehr and Gächter 2002). Moreover, participants will punish non-compliance even when it is certain that they will not benefit personally from future compliance by free riders (Fehr and Fischbacher 2004).
Against these benefits Nichols describes the inadequacies of Pereboom’s analogues, particularly moral sadness and regret. The purported problem with moral sadness is that the behavior it prompts is withdrawal (Nichols 2007, 420). Those who respond with sadness tend to give up rather than protest in the face of obstacles. The problem with regret, Pereboom’s analogue for guilt, is different. Nichols’ primary criticism of this emotion is that it is poorly defined and poorly understood. We simply do not know whether regret motivates people to act in the same useful ways that guilt does.
Nichols also criticizes Pereboom’s suggestion that analogue attitudes might be supplemented by moral resolve. Pereboom suggests moral resolve as a sort of intensifier 103 of existing moral commitments, an independent motivational force beyond the motivation internal to the moral judgment (2001, 205-6). For example, regret about having lied to a friend might motivate one not to lie in the future, but that motivation may be supplemented, and made stronger, by an independent resolution not to lie. The internal disincentive to lying provided by the emotion of regret is supplemented by an artificially generated motivation to stick by the moral resolutions one endorses. Nichols’ objection to this suggestion is primarily anecdotal. He is suspicious of the claim that mere resolution is an effective motivator and he cites as support for this suspicion teenagers’ frequent failure to adhere to their commitments to abstinence and unproductive communist labor forces, labor forces presumably motivated by moral resolve and egalitarian commitments rather than wage market incentives (2007, 421).
What is the significance of these findings and suspicions for Nichols’ normative claim that we ought not abandon the reactive attitudes? Nichols argument is that, if our emotional framework is not flexible enough to permit effectively replacing the reactive attitudes, then we must assess the losses of abandoning them as strongly outweighing the gains. For example, if we cannot incentivize good behavior in others or promote the interpersonal behaviors necessary for valuable relationships without moral anger, then we must consider the universal abandonment of this attitude as a significant and unbalanced loss. This is not to imply that Nichols believes that our typical anger-based responses are optimal deployments of this emotion. He may think that the level of anger we feel and express is excessive. Nichols is arguing that if analogue attitudes are ineffective, or significantly less effective, then we ought not abandon the reactive attitudes. And he concludes that, while there are reasons for abandoning the reactive attitudes, abolitionism 104 is not the optimal solution to the problem raised by the fact (if it is one) that no one is free or responsible.
4.3 Pereboom’s Defense of Abolitionism
Pereboom (2009 and forthcoming) has responded to Nichols’ worries about abolitionism as well as other worries about the adequacy of the analogue attitudes. In response to Nichols’ argument in favor of retaining moral anger, Pereboom suggests three reasons for thinking that moral anger, while effective, is not the optimal tool for its job (2009, 176). First, while self-interested anger is useful in motivating cooperation, we would prefer that cooperation have more virtuous motives. For example, we would prefer that parties cooperate in a joint venture because doing so is fair rather than because they fear reprisal.
Second, anger is a blunt instrument. If the benefit of anger is that it motivates wrongdoers to behave appropriately, then it is optimal only if the amount of anger expressed is the minimum amount needed to motivate the wrongdoer.57 Given that moral anger, as an aggressive and blaming emotion, is harmful, being optimally angry—getting the necessary beneficial motivation with a minimum of anger—is morally important. But we typically do not have this sort of fine-grained control over our anger. This is especially bad if, as seems to be the case, excessive or unrestrained anger can itself breed
57 This intuition is the basis for punishment policies that attempt to set punishments at the minimum level (e.g. minimum number of years in prison) necessary to deter potential wrongdoers. For example, the appropriate punishment for armed robbery is held to be X years in prison, such that if the punishment was less than X years, more (or substantially more) people would commit armed robbery than would if the punishment was X years. For a more detailed discussion see Yaffe (2010, 293-6). 105 resentment and begin a cycle of anger. Pereboom suggests that this is often what happens both interpersonally and in many political and ethnic conflicts between groups.
Two points concerning this particular response to Nichols deserve some attention before we move on. As it happens, Nichols does not say anything to indicate that he would disagree with Pereboom’s claim that anger is a blunt instrument. He may agree but insist that anger is nonetheless a necessary part of our emotional repertoire and of our set of communicative tools. He may even condemn our tendency to respond with disproportionate anger to those who harm us. Nichols might also point out that the proposed analogues to anger are unlikely to be any more finely calibrated than the reactive attitudes they replace.58 While this is probably true, it does not entirely undermine the force of Pereboom’s criticism. For he might respond, rightly I think, that the deployment of a blunt instrument like anger is more harmful, and perhaps significantly more harmful, than the deployment of an equally blunt sadness or disappointment.
Third, and related to Pereboom’s second point, anger does more than motivate self- and other-respecting behavior. It also produces fear, which can then have its own effects on how the subject and the recipient of that anger relate to one another and to others. These negative side effects of moral anger can directly undermine its benefits.
Pereboom also defends his analogue attitudes. Against Nichols’ charge that moral sadness promotes inaction or withdrawal, Pereboom notes that Nichols’ main empirical evidence for this claim concerns infants rather than adults (Pereboom 2009, 178; Nichols
58 Thanks to David Brink for drawing my attention to both of these important points. 106
2007, 420). Such evidence is inadequate given the significant differences in rational capacities between infants and adults. Because emotions require various cognitive and physical capacities, like the ability to make a fist, run away, or recognize others’ thoughts as different from one’s own (Kalat and Shiota 2007, 82-3), mature emotions of sadness are substantially different than infant sadness.59 In response to Nichols’ claim that regret is a poorly understood emotion and thus a suspect candidate for replacing guilt,
Pereboom suggests Hilary Bok’s (1998) conception of non-desert-based guilt. Bok views guilt, and Pereboom regret, as an attitude that hurts not because we are blaming or punishing ourselves, but because we recognize that we have “slighted what we take to be of value, disregarded principles we sincerely think we should live by, and failed to be the sorts of people we think we should be” (Bok 1998, 168-9; quoted in Pereboom 2009,
182-3). On this view, which is just as comprehensible as desert-based conception of guilt, the motivational force of guilt, which is shared by regret, is not the pain of self- directed moral anger, but rather the pain of disappointment. Finally, in response to
Nichols’ suspicion about the motivational power of moral resolve, Pereboom cites his own anecdotal evidence that, at least in certain contexts, a commitment to just and fair treatment combined with moral resolve to do the right thing even when it is difficult can replace moral anger as the foundational emotional element and produced the same or greater benefits. As an example, he cites widespread replacement of punitive (anger- based) institutional punishments with rehabilitative systems and the benefits that these
59 I agree with Pereboom that the empirical evidence about infants does not support the general claim that the expression of moral sadness by adults cannot coincide with self- respecting behavior or motivate others to cease or refrain from inappropriate behavior. 107 systemic changes have had in countries like Finland and the Netherlands (Pereboom
2009, 179).60
Before moving on from Pereboom’s defense of abolitionism and the analogue attitudes, it is worth identifying in his defense what we might can an accommodationist strand of abolitionism. Accommodationist abolitionists are those who aim to demonstrate that life without the reactive attitudes is really not that different from life with them. This strand of abolitionism can be contrasted with what we might call a revisionary strand.
Revisionary abolitionists are those who accept the anti-abolitionist claim that life without the reactive attitudes will be quite different—e.g., characterized by different social practices and relationships. Nonetheless, they hold that we have strong moral and epistemological reasons for changing our world in this way. Of course, accommodationist and revisionary proposals should be understood as scalar rather than binary. Pereboom is closer to the accommodationist end than, for example, Sommers insofar as Pereboom rejects the need to take the objective stance. Likewise, he is accommodationist in the sense that his goal in response to Nichols is primarily to demonstrate that the analogue attitudes can support a similar set of practices to those already in place.
In the next section, I offer a deeper defense of abolitionism in which I pursue both accommodationist and revisionary strategies. I provide a novel contribution to
Pereboom’s case for the effectiveness of non-reactive analogues, but I also argue that we
60 I call this anecdotal evidence because it appeals to the phenomenon of prison reform as an instance of the efficacy of moral resolve, but it does not offer an interpretation of the phenomenon (or the data about it) that gives us reason to think that it clearly applies to the debate about moral resolve. 108 cannot properly assess the effectiveness of these analogues without considering the significant revisions to our social practices that they imply.
4.4 A Broader Defense of Abolitionism
Pereboom disputes each of Nichols’ empirical claims about the analogue attitudes as well as the normative claims that purportedly follow from them. In this section, I offer three additional, and more general, responses to the claim that analogue attitudes are inadequate replacements for the reactive attitudes and I argue that the latter ought not be abandoned.
First, I argue that Nichols is wrong about the relative effectiveness of reactive vs. analogue attitudes. He has not assessed the variety of analogues there are and his assessments of their relative effectiveness are incorrect. Second, I argue that even if
Nichols is right that reactive attitudes typically are more effective in their attitudinal function than their non-reactive counterparts, this fact is contingent upon historical, cultural, and developmental circumstances. If things had been different or become different in the future, the effectiveness of these attitudes may have been different or will become diminished. Third, I argue that Nichols’ argument applies more broadly than he suggests. For example, it may be that a parallel argument would justify taking what we would intuitively consider to be clearly inappropriate attitudes. If his account does generalize in this way, then Nichols’ initial conclusion is less compelling.
A disclaimer is in order before engaging with Nichols’ empirical arguments against abolitionism. One might rightly worry that an empirical defense of the effectiveness of the analogue attitudes will be prone to all the same inadequacies that it 109 identifies in its opponent. For example, it may be the case that the relevant evidence from social psychology establishes neither Nichols’ attack on moral sadness nor
Pereboom’s defense of regret nor, for that matter, my own empirical defense of regret below. In light of this worry one might be tempted to abandon the entire project of empirically supported arguments about our emotions, at least until the relevant empirical sciences have more conclusive results and a more unified picture of how they relate to one another. I am sympathetic both to the worry that philosophers are, to some degree, trading in inconclusive empirical findings and to the implied suggestion that we ought not to do so. I offer my own empirically supported defense of abolitionism with these worries in mind.
Before turning to my criticisms of Nichols’ anti-abolitionist position, I also want to note an ironic feature of Nichols’ Strawsonian project. Nichols defends Strawson using the same tactics by which Strawson’s opponent defends holding people responsible.
Strawson’s naïve optimist opponent defends the practice of holding people responsible on the grounds that a set of practices is useful, perhaps even optimal. The naïve optimist thinks that she can justify the practice of holding responsible in terms of social benefits like public safety, prevention, and deterrence (Strawson 1962, 61-2). Similarly, Nichols thinks he can undermine the proposal to employ analogue attitudes and justify the practice of taking reactive attitudes in terms of their respective social harms and benefits
(e.g., relative motivational force). Strawson finds this form of compatibilism inadequate—and calls its proponents “one-eyed utilitarians”—because it ignores the issue 110 of whether blame or praise is deserved (1962, 79).61 It would seem that Nichols is open to a similar charge. Of course, the irony ultimately attaches to Strawson himself, insofar as it is he who implies that if it is psychologically possible to abandon the reactive attitudes wholesale, then we would have to assess the rationality of doing so in terms of the “gains and losses to human life” as Nichols and others have done (1962, 70).
4.4.1 The Empirical Problem
I begin by questioning the strength and relevance of Nichols’ empirical claims.62
Nichols has not assessed the full variety of non-reactive analogue attitudes and, while his discussion of moral anger is provocative, it does not give decisive support for the claim that non-reactive attitudes are less effective at producing any given benefit than reactive attitudes. Assessing the efficacy of a framework of attitudes requires more than merely determining the relative effectiveness of a few individual attitudes. Admittedly, we cannot reasonably expect an exhaustive assessment, but more must be done to show that a framework of analogues cannot replace the reactive framework.
61 One might rightly note that this criticism of the naïve optimist is actually leveled by the pessimist and that Strawson does not explicitly endorse the criticism himself. However, it is the main objection to the naïve optimist and, insofar as Strawson views that position as inadequate, I think he must endorse the objection. If he does not, then his own optimist position remains in competition with naïve optimism. For these reasons, I believe it is fair to read Strawson as endorsing the pessimist critique. 62 As I said above, in this section, I follow Nichols into a debate that is based largely, and in some cases entirely, on empirical evidence from social psychology, experimental economics, and related fields. As a result, our conflicting conclusions are only as strong as the studies from whose results we generalize and only as fitting as the generalizations themselves. For example, the empirical literature on regret is, I think, compelling and relevant, but it, like the experimental economics described by Nichols, is a relatively shaky foundation for broader claims about the roles of particular attitudes in our many and varied daily experiences. 111
I raise four empirical problems for Nichols’ anti-abolitionist position. The first two concern his claims about the benefits of the reactive attitudes and latter two concern the efficacy of analogue attitudes.
Moral Anger. As Pereboom rightly notes, moral anger is likely to be suboptimal in virtue of the negative effects that follow from the fact that it cannot be deployed precisely (2009, 176). Even if anger-based reactive attitudes have the benefits that
Nichols’ claims, they will sometimes (perhaps often) be excessively punitive. If this is true, then there is still good reason to abandon them. This is not to say that we cannot learn to express (and feel) anger in proportion with the magnitude of the harm. However, it might be that the practice of expressing only optimal moral anger—the minimum moral anger sufficient to produce the necessary motivations—requires as substantial or as difficult a revision of our existing framework of attitudes as the project of abandoning that framework completely. That is, it may be just as difficult to temper existing attitudes as to replace them, to govern them as to exclude them. Both abolitionism and Nichols’ anti-abolitionist position are revisionary in this sense. This argument loses some of its force if we acknowledge that non-reactive analogues also can and likely would be disproportionate to their causes. However, as I noted above, disproportionate anger is more harmful than disproportionate sadness and therefore more morally objectionable.
Pereboom and Sommers consider replacing moral anger with non-anger-based emotions with similar functions, including moral sadness. Such emotions are held to be viable replacements insofar as they perform the same functions of demonstrating self- respect and communicating moral protest. However, given the reasons for abolitionism, one might just as well suggest replacing attitudes like moral anger—i.e., anger that 112 presupposes moral responsibility—with non-moralized anger—i.e., anger that does not presuppose moral responsibility. For, while the fact (if it is one) that we are neither free nor responsible requires, according to the abolitionist, that we abandon the reactive emotions, it does not necessarily require that we abandon the feelings associated with those emotions. For example, while abolitionism finds fault with blaming anger, non- blaming anger is acceptable. Resentment both mistakenly attributes freedom and responsibility to its target and also involves wrongful blaming. But anger itself does not necessarily do so. For example, I may be angry at my computer for crashing or at the weather for raining during a race. Or, one may experience anger driven solely by recognition of an act’s wrongness rather than by its blameworthiness. The appraisal component of this emotion is not a judgment of responsibility, but merely a judgment of wrongness. I may experience anger or frustration at a person who does something nice, but insufficiently nice given his ability or opportunity. For example, I may be angry at a billionaire who sends a few hundred malaria nets to Malawi and then rests on her charitable laurels.
The idea of replacing moral anger with non-moralized anger is plausible, but is potentially objectionable for some of the same reasons that Pereboom criticizes moral anger. For example, it may be just as blunt an instrument for social management and interpersonal engagement. It may also be the case that, while possible, it is often difficult to separate anger at a person for her actions from the non-moralized dimension or element of that anger. Finally, like moral anger, non-moralized anger is liable to harm its target by causing fear or anxiety. This is a serious downside to anger that is independent of its presumption of responsibility. To the extent that this is true, replacing moral anger 113 with distinct emotional analogues like sadness or disappointment rather than emotional analogues of the same form, like non-moralized anger, may be the preferable alternative.
A more serious worry about abandoning moral anger concerns our typically different responses to responsible and non-responsible wrongdoers, toward whom different attitudes are rightly held to be appropriate.63 Whether we replace moral anger with non-moralized anger or some other analogue like sadness or disappointment, the anti-abolitionist might reasonably ask whether and how our reactions to different kinds of wrongdoers should differ—e.g., our reactions to insane versus sane (but non-responsible) wrongdoers. However, the abolitionist retains the resources for justifying and explaining the appropriateness of different reactions in different cases of this sort. At the very least, the abolitionist, like the anti-abolitionist, has variety of ways of dealing with wrongdoers who are incapacitated in different ways and to different degrees. Just as we do not typically respond to (or engage with) those who have volitional incapacities in the same way we respond to (or engage with) those who have cognitive incapacities, so we can respond in different ways and with different emotions to those who are non-responsible and acutely psychotic versus those who are non-responsible but cogent and on the whole rational. For example, an acutely psychotic person might be beyond the reach of any effective treatment other than detention and medication, while other non-responsible agents might be congenitally blind to certain sorts of reasons as result of their upbringing.
None of the abolitionist’s commitments require that she treat (or respond to) such people the same way
63 Thanks to David Brink for raising this potential problem for the abolitionist. 114
Positive Attitudes. Nichols defends moral anger in the form of resentment and indignation, but does not give an argument in defense of other reactive attitudes. In particular, he does not argue for retaining the positive reactive attitudes. Likewise, his criticisms are restricted to the negative analogue attitudes. Nichols seems to view his conclusions about the anger-based reactive attitudes as sufficient to ground his general anti-abolitionist position. If, as I have argued, none of Nichols empirical evidence is decisive, then it is incumbent upon him, and upon anti-abolitionists generally, to consider the variety of reactive attitudes and the corresponding variety of analogues.64
Considering the positive reactive attitudes is especially important because positive emotions in general are important for our social lives. There is evidence that the expression of positive emotions towards children is both positively correlated with healthy emotional development and expression by those children and also negatively correlated with problem behavior (Eisenberg et al. 2001). Such results, as well as common sense, suggest that it may be the positive reactive attitudes rather than the negative attitudes that are most important to healthy social interaction—and, therefore, to the general richness or impoverishment of human life. If this is the case, then it would seem that both the abolitionist and the anti-abolitionist should be concerned primarily
64 The imperative to assess the variety of non-reactive attitudes becomes especially clear when we consider how broad this class is. Emotions are combinations of cognitive (belief) and affective (feeling) components and the effectiveness of an emotion, its tendency to motivate one to act or refrain from acting, depends upon these two elements. While the space of probable emotions, understood as appraisal-feeling combinations, is much smaller than the space of possible emotions, it is still large enough that there might be combinations that have the motive power of anger, but not the negative side effects. An assumption of this sort is exactly what lies behind the abolitionist’s confidence that effective analogues to the reactive attitudes can be found. 115 with the relative effectiveness of the positive reactive attitudes and their analogues. We should wonder, first, do positive reactive attitudes have identifiable functions, as negative attitudes do? And, second, can non-reactive positive emotions perform the same, or equally beneficial, functions as their reactive counterparts?65
Consider forgiveness. While it is controversial whether forgiveness itself is an emotion, most would agree that some sort of forgiving emotion is involved in forgiveness and that this emotion presupposes that its target, the offender, is responsible for her wrongdoing. That is, on most accounts, the forgiving emotion is a reactive attitude in the standard sense (Strawson 1962).66
Suppose that reconciliation is one function of forgiveness. This is a common, though by no means ubiquitous, view amongst philosophers and it is supported by empirical evidence. Studies have shown that forgiving is almost always followed by some form of reconciliation between victim and offender (Darby et al. unpublished manuscript, Fincham 2010) and while this does not imply that reconciliation was the aim of forgiving in these cases, it is plausible to suppose that it was the aim in most cases.
Forgiving also has benefits that are indirectly related to its reconciling function. For example, some mental and physiological benefits accrue to forgivers more so than to those that do not forgive (Witvliet et al. 2001; Larsen et al. 2012). The question, then, is
65 One might question the relevance of discussing positive reactive attitudes on the grounds that the abolitionist argument for abandoning the positive attitudes is significantly weaker than the argument against the negative attitudes, since the former do not involve wrongful blaming. Nonetheless, a version of both the epistemological and the moral reason for abolition still apply. 66 Henceforth, I will simply use the term ‘forgiveness’ to refer to the forgiving emotion, keeping mind that it is debatable whether the folk concept of forgiveness is adequately captured by our conception of the emotion. 116 whether the same benefits accrue (or would accrue) to those who abandon the reactive attitudes and/or replace them with analogue attitudes.
There is reason to think that the same benefits would indeed accrue. Just as forgiving involves the emotion of forgiveness, so similar processes involve similar emotions that perform similar functions. For example, suppose that Sally believes that
Jesse has wronged her. Sally might reconcile with Jesse because it has become apparent that Jesse did not actually wrong her. The process here is condoning; the function is reconciliation; and the emotion is an analogue of forgiveness whose appraisal component does not presuppose a judgment of wrongdoing. Call this the condoning emotion.
Pereboom (2001) contends that reconciliation of the sort prompted by forgiveness is compatible with analogues that do not presuppose moral responsibility. Consider the case of condoning again. Suppose that Sally believes that Jesse has been spreading lies about her and that she has decided not to be friends with Jesse any more. If Sally comes to believe that Jesse did not really lie, that his words were twisted into a vicious rumor by someone else, Sally may condone the action she thought was wrong. Her condoning attitude may remove an obstacle to their continued friendship and enable her to reconcile with Jesse.
The psychiatric literature on PTSD and survivor guilt—i.e., pathological guilt felt by those who survive natural disasters, genocides, or wars in which their loved ones perish—provides examples of the reconciling power of condoning, specifically self- condoning (Ehlers and Clark 2000). Patients suffering from post-traumatic stress disorder often have inaccurate or incomplete memories of the event. These inaccurate or incomplete memories make it difficult for patients to work through their traumas with 117 their doctors. The psychiatric techniques of cognitive restructuring and reliving attempt to integrate blame-relevant facts known by the doctors and learned by the patients into their conception of the traumatic events. For example, “a bus driver who had run over an elderly lady and felt very guilty became increasingly aware during reliving that the lady had intended to commit suicide by stepping in front of the bus and his intrusions
[intrusive memories] of seeing the lady look at him shortly before the impact decreased”
(2000, 340). By integrating these previously unrecognized elements of the event into his memory of the trauma, the bus driver is (more) able to work through his self-blame and condone his actions.67
I suggest that the condoning emotion is a plausible analogue for forgiveness.
Moreover, it seems to serve the same basic function as forgiveness, minus the latter’s function of holding responsible. Of course, more precise assessment of the relative effectiveness of this analogue is needed, but the evidence for this claim is as strong as the evidence that forgiveness performs its function well. In these emotions, we have the foundation of a case for replacing forgiveness and other positive reactive attitudes with non-reactive analogues.
67 Of course, the empirical evidence does not unequivocally support the abolitionist claim that analogues can perform the reconciling function. According to one study, those who excuse instead of forgiving—e.g., those who deny that the offender is responsible rather than responding to her change in quality of will from ill will or objectionable disregard to basic respect—are significantly less likely to reconcile than forgivers and are less likely to experience similar decreases in negative feeling and increases in positive feeling (Darby et. al. unpublished manuscript). However, it is possible that the smaller correlation between excuse and reconciliation in Darby et al. is an artifact of the study design, which compared the experiences of self-identified forgivers and non-forgivers rather than forgivers and excusers. 118
We might gain similar insight into the relative effectiveness of positive emotions by considering gratitude. Consider the following case.
Gratitude. Jesse meets a visiting scholar, Raphael, while working in the philosophy department. She invites him to a dinner she is having with other friends from the department. He happily accepts and is grateful for the offer and the opportunity to make new friends.
Now suppose Raphael has been convinced by free will skeptics that no one is responsible and has managed to cleanse his attitudes of their responsibility appraisals. Very little seems to be lost if Raphael expresses gratitude-sans-praise rather than gratitude-with- praise to his gracious host (Pereboom 2001, 201-202). He is still able to express appreciation, delight, a belief that the act was supererogatory, and perhaps an intention to repay the favor in the future (Sommers 2007, 329-330). Both Pereboom and Sommers see many of the positive reactive attitudes—including gratitude, pride, and forgiveness— as having strong analogues.
Regret. Regret is not as poorly understood as Nichols suggests. There is a substantial body of empirical research on regret, though much of this research has been on non-moral regret.68 Regret is distinguished both from guilt (Zeelenberg and
Breugelmans 2008) and from disappointment (Zeelenberg et al. 2000). It has identifiable behaviors associated with it and a motivational effect antecedent to decision-making.
Conceptually, regret is associated with tendencies both to “kick oneself” regarding one’s decision or action and also with the tendency to try to change the situation created by that decision or action or avoid similar situations in the future (Zeelenberg et al. 2000, 525).
68 See, for example, Zeelenberg and Beattie (1997), Zeelenberg et al. (2000), Zeelenberg et al. (2002), and Zeelenberg and Breugelmans (2008). 119
In one study participants played an ultimatum game, in which one participant (the proposer) was given control of $60 and allowed to propose a split between herself and her partner (the responder) (Zeelenberg and Beattie 1997).69 If the responder accepted the split, then each would receive the proposed allotment. If the responder refused the split, neither participant would receive anything. Unknown to the participants, the responders were controlled by the researchers and all offers were accepted. After the game was over, each proposer was informed that his or her offer was accepted. They were then randomly divided into two feedback conditions, both of which were told that they offered more than the minimum their responder was willing to accept: participants in the first group were told that they had offered $1.20 too much (the average offer for this group was $21.20); those in the second were told that they had offered $5.90 too much
(the average offer for this group was $18.08). Participants were then asked to report the level of their regret about the game. Predictably, those who substantially overpaid reported significantly more regret than those who minimally overpaid. Participants were then told that they would play the game again with new responders. They were also informed that the average minimum acceptable offer in the first round was $13.07. All participants then played the game again. Analysis of the second round of offers showed that participants’ second offers were significantly affected by which feedback condition they were in, the minimally overpaid or the substantially overpaid condition. Moreover, the significance of the feedback condition depended upon the level of experienced regret.
69 The original experiment was done in the Netherlands in 1996 and used Dutch guilders as the in-game currency. I have converted guilders into current dollars in order to make the relevant amounts more intuitive. 120
That is, the feedback they received mattered in virtue of the regret it elicited rather than, for example, because it allowed them to play the game better.
This study suggests that experiencing regret can and does affect future actions in a way that benefits the regretful agent. Regret, like guilt, seems to have the function of disincentivizing imprudent or harmful actions. The negative emotional experience of regret is itself a disincentive. The fact that we feel bad after making regrettable choices makes us less likely to make such choices in the future. Moreover, anticipated regret70, again like guilt, has a similar function. Anticipating future negative emotions, whether consciously or not, produces the same sorts of benefits as experienced regret.
Anticipating regret leads one to behave prudently and morally, to act so as to avoid harming oneself or others. Moreover, a review of similar studies provides substantial evidence that anticipating regret prior to decision significantly affects choices—e.g., those who anticipate regretting their actions are less likely to make risky investments or practice unsafe sex (Zeelenberg et al. 2000, 534-5). Finally, the effectiveness of regret has been demonstrated in non-student populations of lottery participants (Zeelenberg and
Pieters 2004).
However, he empirical evidence just presented about the motivational effectiveness of regret is lacking in two important respects. First, most (but not all) of the research reported here concerns non-moral regret—e.g., regret about non-optimal decisions to do A rather than B, spend $X rather than $Y, etc. As such, it is not clear
70 “According to regret theory, people can anticipate emotions such as regret, because they compare possible outcomes of a choice with what outcomes would have been, had a different choice been made” (Zeelenberg and Pieters 2004, 156). 121 whether the results can be generalized to cases of moral regret—e.g., regret about having harmed or wronged oneself or another.71 Second, the benefits of regret demonstrated by these studies are primarily, though not exclusively, benefits to self (i.e., to the regretful agent. 72 However, despite these two admitted limitations, I take this evidence to be quite suggestive. And it is unreasonable, I think, in light of these results, to write off regret.
Further research is necessary, but the empirical question is open, and the evidence about guilt and regret seems to point in favor of the abolitionist position.
However, Pereboom does face an empirical problem concerning his account of regret. Regret is poorly understood in the sense that it is not clear whether common conceptual distinctions between, for example, regret and guilt correspond to differences in experienced emotion. It is not clear whether or in what ways feeling guilt (as commonly understood) differs from feeling regret (as commonly understood) (Sabini and
Silver 2005). To some extent, the obstacle to an empirical understanding of regret posed by the mismatch between naming and experiencing an emotion is irrelevant to
Pereboom’s defense. His claim is simply that there exists a self-directed, negative, but non-blaming attitude that can replace guilt and which he calls regret (Pereboom 2009,
182-3). However, the obstacle is relevant insofar as any available evidence concerning the efficacy of regret will not necessarily be evidence of the efficacy of Pereboom’s
71 However, there is evidence that people who have been harmed respond in a similar way to offenders who express guilt and to those who express moral regret (Van Kleef, De Dreu, and Manstead 2006). In these studies guilt has a somewhat stronger effect than regret, but the behavior associated with both appears to be the same. 72 Indeed, Van Kleef, De Dreu, and Manstead speculate that expressions of regret produce a weaker effect than expressions of guilt precisely because regret is ambiguous about whether the person regrets the harm to herself or the harm done to the other (2006, 130). 122 analogue. That is, Pereboom’s conception of regret may not correspond to the conception of regret about which we have empirical evidence. Luckily, this is a worry for Pereboom, but not for the abolitionist generally. The abolitionist can ignore
Pereboom’s conception of regret—i.e., Bok’s account of guilt—and accept the conception employed by empirical investigators—assuming, of course, that it is not conceptually mistaken.
Resolve. Neither Nichols nor Pereboom gives an adequate account of the efficacy of moral resolve. They give anecdotal evidence of its inadequacy and power, respectively. Nichols notes the uselessness of resolve among abstinent teens and
Pereboom notes instances of the success of collective moral resolve at the level of institutional punishment. Taking seriously the possibility of moral resolve of the sort
Pereboom describes and relies on requires answering two questions. First, does moral resolve provide motive force that is independent of our emotions and attitudes? Second, if so, does moral resolve provide sufficient independent motive force that, as a supplement to analogue attitudes, it can help replace reactive attitudes? I claim that there is empirical evidence to support affirmative answers to both questions. I will consider one line of empirical investigation that provides evidence for the claim that resolve has motive force independent of the emotion or attitude it is supplementing. The studies I consider may show that willpower is a resource that can be depleted and restored, independently of any particular emotions.73
73 The aim of ego-depletion studies is not to demonstrate the claim that there is a source of resolve or motivation that is independent of a particular emotion of attitude. However, I contend that their results support this claim. 123
The phenomenon of ego-depletion or depletion of willpower may provide evidence for the abolitionist claim that moral resolve can supplement the motivational inadequacies of non-reactive attitudes. A number of studies demonstrate that willpower can be depleted by exercising self-control. This depletion is sometimes understood via the metaphor of an overworked muscle. The more you use the muscle, the weaker it becomes until it is allowed to recover. In these studies, participants are required either to use self-control in a neutral task, like writing a short essay without using the letters N or
A (Gino et al. 2011) or completing a Stroop task74 (Job, Dweck, and Walton 2010), or to perform a task with no self-control element. Those who perform the self-control task are more likely than undepleted controls to cheat on a later task for which they get a monetary reward (Gino et al. 2011)). Assuming that the neutral self-control tasks used to deplete participants do not illicitly deploy reactive attitudes, we can infer from these studies that willpower is a motivational resource that can be accessed (and depleted) independently of reactive attitudes. Moreover, if we assume that the reactive attitude of guilt functions as a disincentive to wrongful behavior, a motivation against acting wrongly, then we can explain the failure of guilt to disincentivize cheating behavior in these studies in terms of independently depleted willpower. That is, one description of what is demonstrated by these studies is that both neutral self-control tasks and reactive attitudes take from a common store of willpower. When this store is depleted by self-
74 The Stroop task is a reaction time task, in which participants are asked to name a word presented to them. When the name of a color (e.g., "blue," "green," or "red") is printed in a color not denoted by the name (e.g., the word "red" printed in blue ink), naming the color of the word takes longer and is more prone to errors than when the color of the ink matches the name of the color (Stroop 1935). In such cases, naming the word requires self-control. 124 control tasks, less it available for the reactive attitudes and they are therefore less effective.
But if the common store can be depleted independently of reactive attitudes, then it can perhaps also be independently strengthened, recharged, or shielded from depletion, by the exercise of moral resolve. It follows from the fact that it can be depleted, but that depletion is not permanent, that willpower can be recharged. Moreover, more recent studies show that not only can willpower be depleted and recharged, but it can also be shielded from depletion. Schmeichel and Vohs (2009) demonstrate that self- affirmation—i.e., affirmation of core values or commitments prior to self-control tasks— mediates or entirely counterbalances the depleting effects of these tasks. In one experiment, participants performed a self-control task (in which they were required to write an essay without use the letters A or N), followed either by a neutral task or a self- affirming task (writing about a time when they acted on a value they ascribe to themselves and take to be important in their lives). All participants then performed a physical self-control task in which they were required to submerge their hand in icy water for as long as they could. Participants who performed the self-affirming task were able to keep their hand in the icy water significantly longer than those who were depleted but not self-affirmed. Moreover, this effect remained statistically significant even after controlling for effects of mood and arousal caused by having to perform difficult or annoying tasks (2009, 773).75
75 A meta-analysis of the ego-depletion literature is more hesitant to rule out the effects of negative affect (Hagger et al. 2010). They find that few studies report the effects of 125
One might worry that, even if we accept that the willpower expended in ego depleting tasks is independent of the reactive or analogue attitudes, the shielding effect of self-affirmation is dependent upon the attitudes prompted by self-affirming task. For example, one explanation of the results is that writing about a value one takes to be important would evoke pride in a person and that such pride might have the effect of motivating actions that one can be proud of. However, if this were the case, then we would expect all participants in the self-affirming condition to experience this effect, but, in fact, only those who were first depleted and then affirmed perform better that the non- affirmed on the second self-control task (2009, 772-3). The other interpretation is that pride (or another emotion) only supplements motivation when a person has been depleted and that non-depleted participants were therefore unable to benefit from the pride elicited by their self-affirmation tasks. The evidence does not to my knowledge rule out this possibility. Nonetheless, even if we think the motivational force of self-affirmation is dependent on the emotions, the studies about ego-depletion alone appear to demonstrate that willpower—and resolve—is a resource that is independent of both reactive and non- reactive emotions.
These studies seem to show that willpower can not only be depleted independently of reactive attitudes, but also that it that it can be shielded from depletion and that the non-depleted resource can be accessed for future tasks. Finally, and most importantly, it is not unreasonable to suggest that self-affirmation of core values is itself a form of moral resolve. Admittedly, self-affirmation was not explicitly used by negative affect on task performance and suggest that future research investigate the possibility of such a mediating effect (2010, 517-8). 126 participants to augment or shield willpower reserves, but it plausibly could be. We can imagine an experiment in which the final self-control task was a moral task and in which the self-affirmation task was directly relevant to the upcoming self-control task. For example, the final self-control task might involve a cheating opportunity and the self- affirmation task might require the participant to reaffirm her (presumed) standing commitment to refrain from dishonest behavior. If such an experiment produced results similar to those in existing work on ego-depletion and self-affirmation, we would have evidence of a source of motivation that can be accessed independently of reactive attitudes and which can be used either on its own or to augment motivation internal to other emotional responses. And this is precisely what Pereboom and other abolitionists are suggesting in proposing that analogue attitudes like moral sadness can, if they are found to be motivationally deficient, be supplemented with moral resolve.
The challenge for the abolitionist who wants to invoke moral resolve as an independent source of motivation with which to supplement analogue attitudes is to identify cases in which a person experiences the attitudes, but the motivation is absent— or vice versa. This is very difficult to do. However, she can adequately meet this challenge by identifying cases in which the emotions are not the true or most relevant source of motive force. For example, the abolitionist can identify cases in which we would expect someone to be motivated X amount by pride, but whose behavior clearly demonstrates that he is motivated more or less than X amount. This is exactly what I take the results about ego-depletion and self-affirmation to show, though admittedly the evidence is less than decisive. 127
On the abolitionist account, the framework of attitudes that shape and enrich our interpersonal experience will indeed change, but the implications of these changes is not necessarily an impoverished or an incoherent world. It is true that resentment and indignation are useful retributive tools as well as forms of moral communication. We learn what others find offensive, what they find laudable, and to what degree by observing their behavior and reactive attitudes. In the same way, we learn what people are grateful for, what they forgive, what makes them ashamed or proud. We use these lessons to build relationships with others by demonstrating that we understand and care for them. Much of this learning is still possible with the analogues that I have described and much of what makes our actions meaningful to one another and to ourselves remains when we apply this knowledge. It is touching to learn that a friend values me enough to reconcile with me despite my bad behavior; and it is evidence of love, trust, and commitment when my partner proposes that we get married. All of this is true regardless of whether one assumes that the other is responsible.
4.4.2 The Contingency Problem
Pereboom and Nichols disagree both about the strength of their empirical evidence and about the gains and losses to human life of abandoning or retaining the reactive attitudes. However, both appear to accept that the proper context of evaluation is the world as it is, with its present social structures. I deny this and, as a result, pursue a different strategy in defense of abolitionism. I too assess the gains and losses of abandoning the reactive attitudes, but I calculate the opportunity costs of doing so by reference to the society that abandoning them will create rather than the one that currently 128 exists. Thus, in this section, I argue that even if Nichols is right that reactive attitudes are more effective in their attitudinal function than their non-reactive counterparts, this fact is contingent upon historical, cultural, and developmental circumstances. If things had been different or become different in the future, the effectiveness of these attitudes may have been less or will become diminished.
Even if Nichols is right that reactive attitudes are currently more effective than their non-reactive counterparts, it does not necessarily follow that we should not try to change the framework in which they are effective. The effectiveness of one emotion relative to another for any given function will inevitably be due, in large part, to contingent biological, cultural, and historical facts. The emotions we experience have evolved biologically, emerged in particular cultural contexts over the course of human history, and, for particular individuals, developed in response to such biological, environmental, and cultural pressures. All of these processes are contingent. Our emotional repertoire could have been different but equally functional.76
76 At the biological level, bonobos and chimpanzees illustrate the contingent evolution of different emotional repertoires. Bonobo groups have significantly less male aggression than do chimpanzee groups (Hare, Wobber, and Wrangham 2012, 576-7). This seems to have resulted from geographical isolation and, thus, different selection pressures acting on the two populations. Despite the evolution of different behavioral repertoires, both species have existed successfully for millions of years. Thus, insofar as we think that ape behavior reflects emotions as it does in humans, it seems fair to describe chimpanzees and bonobos as having evolved different, but equally functional, emotional repertoires. At the cultural level, Richard Nisbett and Dov Cohen note that herding economies are more likely to produce so-called cultures of honor, in which anger-based emotional expression and physical violence are significantly more common (1996, 5-7). Here we have a significant difference in the prevalence, though admittedly not the existence, of anger-based emotions between herding and farming economies, each of which seems to contain individuals with functional emotional repertoires. Finally, at the individual level, psychologists who study Buddhist meditative practices appear to agree that practitioners 129
It seems clear that the human emotional repertoire could have been and could someday be different. Admittedly, it is not as obvious a) whether these differences would be equally functional or b) whether the actions necessary to radically change our present emotional repertoire are within our control. Nonetheless, I think we have sufficient reason to assess the relative effectiveness of reactive and analogue attitudes by reference to the social context that would likely exist if we tried, or began, to abandon and replace the reactive attitudes rather than, as Nichols does, by reference to the status quo. And this is all I need in order to make my argument.
In a society in which physical force is the most effective means of deterring undesirable behavior, attitudes that both express protest and motivate physical retaliation against wrongdoers may be the most effective deterrents. In a different society, different attitudes may be (or have been) more useful. To say that only attitudes that meet current standards for effectiveness can perform the function currently performed by those attitudes is to deny that the larger set of practices in which our current attitudes are effective can or will change.77 But this sort of change is exactly what advocates for abandoning the reactive attitudes think can and should happen. Nor is it the kind of change that Strawson believed was psychologically impossible, namely, wholesale
are attempting and accomplishing a sort of emotional regulation. In this Buddhists, Ekman et al. not only identify a group of people who practice the elimination of various “afflictive” emotions from their emotional repertoire, they also describes how the practitioners do so, namely, by careful introspection and monitoring (2005, 60). This, too, seems to fall within the range of comparable effectiveness. 77 Nichols argues that our practices would not change just because we come to accept determinism and incompatibilism (2007, 407-9). However, this does not mean that our practices would not change if, in addition to accepting these (or similar) theses, we tried to change our practices. 130 psychological divestment from our “natural human commitment to ordinary interpersonal attitudes” (1962, 70).
But this is similar to the kind of error made by those who opposed women entering the labor force. They evaluated this trend within the context of traditional gender roles, which presupposed that women would retain domestic duties, and within which the cost of their absence from the domestic sphere appeared, on balance, destructive. Both assessments mistakenly hold fixed what is, in fact, a contingent feature of the world—i.e., traditional domestic roles and existing social practices—and thereby misrepresent the social costs of change. It is true that some beliefs, attitudes, and actions are ineffective or counterproductive within the current social structure. For example, a woman who chooses not to dress or make herself up in accordance with current norms of professionalism and feminine fashion may put herself at a disadvantage in the job market or amongst her social peers (Kyle and Mahler 1996). Similarly, at the social level, it may be that the early movement encouraging women to work outside the home, led them to take worse jobs for less pay than men rather than stay home.78 And it may be that the influx of women into the workplace contributed to the assumption that neither men nor women needed a full living wage (Okin 1989, 155-6).
However, these facts do not justify the claim that women ought to conform to current standards of beauty or that our mothers’ (and grandmothers’) generation ought to
78 Susan Moller Okin quotes Amy Gutman’s statistic that 87% of elementary school teachers are female, while 99% of school superintendents are male (Okin 1989, 177). While it is not the case that being a teacher is worse than being a superintendent, the point being made is that women, as of 1989, remained in lower paying jobs further down the ladder of promotion. See also Okin’s discussion of wives and wage work (155-6). 131 have refrained from working. Some of these changes in belief, attitude, and action are necessary for social improvement and progress toward basic justice. It is not a good argument against abandoning our patriarchal beliefs, attitudes, and actions that doing so will disadvantage us in our current patriarchal society. (At least not unless the immediate disadvantage is so great that it outweighs even long term benefits of systemic change, which seems unlikely.) Nor is it a good argument against abandoning our racist beliefs, attitudes, and actions that doing so will disadvantage us in a racist society. These arguments calculate advantage on the assumption that the existing system of oppression will persist instead of on the assumption that a system is responsive to individual and group changes. They discount or fail to consider that initial changes can alter the underlying framework and that relative value of possible future actions will be determined by that altered framework and not the present one. For example, they fail to consider that children (boys and girls) who see women in positions of power as they grow up may respond differently to women of their own generation when they enter the workforce (Lockwood 2006). A society whose citizens have had these role models will be more likely to view such positions as opportunities to which they have a right rather than a privilege.
Nichols makes a similar mistake when he calculates the relative effectiveness of particular attitudes on the assumption that the current attitudinal framework will persist instead of on the assumption that it will be responsive to individual and group changes in the attitudes they take. It is not a good argument against abandoning our retributive and blaming beliefs, attitudes, and actions that those who initiate these changes will be at a 132 disadvantage in our present society—that is, a society in which those who take “natural” blaming attitudes maintain an oppressive status quo.
Nichols may respond to this objection in one of two ways, each of which disputes the analogy drawn with systemic sexism. First, he may deny that systemic change in our practices is possible, that our practice of taking reactive attitudes is not contingent in the ways the abolitionist suggests. I have already given reasons for thinking that our attitude- taking practices are contingent across individuals, across groups, and across time.
Moreover, this contingency is evident in the fact that these practices seem to differ and change as the moralities of individuals and groups differ and change.79
Second, he may argue that his claim that certain non-reactive attitudes are ineffective, or that reactive attitudes do the same job more effectively, would hold even if the world changed in the ways predicted and desired by abolitionists—i.e., even if individuals and groups began to inhibit or renounce individual reactive attitudes or purposefully cultivate non-reactive dispositions. The claim here is that the relative effectiveness of the reactive attitudes would not differ across nearby possible worlds, in which abolition is attempted. However, to my knowledge, neither Nichols nor anyone
79 Barbara Rosenwein’s (1998) collection Anger’s Past: The Social Uses of an Emotion in the Middle Ages contains a number of studies of anger that illustrate differences both in how people at different times and in different cultures felt anger, used it to influence others, and judged those who showed anger. Of particular interest are Stephen D. White’s study “The Politics of Anger” and Paul Freedman’s “Peasant Anger in the Late Middle Ages.” White shows, for example, that many anger-driven behaviors that we would immediately denounce as excessive were held to be justified (139-140). Moreover, while he explicitly refrains from explaining the role of medieval emotions in functional terms (151), one might plausibly give such an explanation using just the sort of model Nichols describes for moral anger in the contemporary context. 133 else has made this argument, and it is a much more ambitious empirical project than the argument currently on offer.
4.4.3 The Generality Problem
The previous two subsections attacked the empirical foundation of Nichols’ brand of anti-abolitionism, first on the grounds that the available evidence does not establish that reactive attitudes are more effective than analogues, and then on the grounds that the purported superiority of the reactive attitudes may depend upon a foundation that is itself systemically unjust and ought to be abolished.
In this subsection, I consider one last objection to Nichols’ position. I contend that his defense of the reactive attitudes fails because it proves too much. That is, his argument about the relative effectiveness of reactive attitudes over non-reactive analogues applies more broadly than he suspects or suggests. Fundamentally, Nichols’ argument is that taking a prima facie inappropriate reactive attitude can be justified in virtue of how effective it is at performing its beneficial function. However, such an argument could plausibly be made not only about reactive attitudes like resentment and guilt, which are inappropriate in virtue of the fact (if it is one) that no one is free or responsible, but also about emotions like hatred and rage that are inappropriate in their extreme forms. It seems that Nichols’ argument could justify excessive hatred or rage if it could be shown that such emotions are necessary to motivate beneficial individual and group practices. If his argument does generalize in this way, then we must either reject it or accept that excessive hatred, rage, or other emotions can be justified in the same way. 134
Of course, Nichols might cheerfully accept this implication. He might note that the generalization depends upon a very unlikely empirical premise, namely, that seemingly excessive hatred or rage might most effectively fulfill some psychological function. Our worry about the generalizability of his argument should be proportional to the probability that its empirical premise is true, which is very small. Moreover, the more likely it is that rage is optimal, and thus defensible according to Nichols’ argument, the less likely it is that such rage is (or appears) excessive. Nichols can point out that if this unlikely empirical premise turns out to be true, then we have good reason to reconsider whether the intuitive distastefulness of strong hatred is actually unmotivated or misguided. (Consequentialists make similar arguments in defense of apparently counterintuitive implications of their theory (Kagan 1997). For example, some attack consequentialism because it appears to imply that it is permissible to kill a person in order to use his organs to save five people. However, one consequentialist response is that, if such an action really is optimal and we see how and why it is optimal, then the claim that it is permissible to kill one to save five becomes more intuitively plausible.)
I am sympathetic to these responses. I agree that it is unlikely that extreme rage or hatred will be an optimal emotion along any dimension of assessment. And I agree that a fleshed out story about how it could be optimal does diminish how counterintuitive it seems at first glance. However, some may be less sympathetic to this strategy. Some abolitionists will respond that what is objectionable is the implication that the effectiveness of an attitude at performing its role constitutes a decisive reason to take that attitude rather than another. (Indeed, this anti-abolitionist defense of the reactive attitudes fundamentally parts ways with Strawson’s own project.) The reason we take 135 extreme rage to be objectionable is that it is unfair, independently of its effectiveness.
The anti-abolitionist escape from the generalizability objection precludes giving serious moral weight to questions of fairness or desert. This puts the Strawsonian in a bind, since he must defend retaining the reactive attitudes at the cost of undercutting the motivation for finding them relevant in the first place, namely, the perceived weakness of naïve optimist arguments grounded in the mere usefulness of some emotional response.
4.5 Conclusion
On the basis of these three arguments—concerning the Empirical Problem, the
Contingency Problem, and the Generality Problem—I conclude that abolitionism is not threatened by the claim that reactive attitudes, especially moral anger, are an essential part of our moral psychology and that the benefits of taking them and of being disposed to take them outweigh the abolitionist reasons for abandoning them. To the contrary, I take this chapter to have shown that the empirical evidence does not support such a claim and, more importantly, that even if it did, it does not follow from such facts that we ought not pursue abolitionism. The fact that some attitudes and actions are ineffective (or even counterproductive) within the current social structure does not imply that they ought not be promoted. Moral and social progress often begins by exposing a system ill suited to supporting the necessary reforms. Thus, while Nichols and Pereboom both treat this objection (about the replaceability of the reactive attitudes) as primarily an empirical matter, I contend that it is not. Empirical evidence is relevant, but at bottom it is a moral question.
Chapter 5
Between Humanity and Intelligence: Abolitionism and Personal
Relationships
“But what is above all interesting is the tension there is, in us, between the participant attitude and the objective attitude. One is tempted to say: between our humanity and our
intelligence.”
– P.F. Strawson, “Freedom and Resentment”
Most would agree that if no one is responsible, then there are many ways in which we must act toward and conceive of one another differently, including adapting existing social practices to fit this reality. One such practice is taking reactive attitudes—i.e. praising and blaming attitudes like gratitude and resentment—toward one another.
Abolitionism is the view that if no one is responsible, then we ought to abandon the reactive attitudes. There are two powerful reasons, one moral and one epistemological, for this conditional obligation. Just as it is wrong to punish those who are not blameworthy, so it is wrong to blame those who are not blameworthy. Moreover, if we are not responsible, then taking reactive attitudes toward one another gets the facts wrong. When we resent one another, we ascribe properties, abilities, and capacities that the other does not actually have.
In this chapter, I defend both abolitionism and a position that I call Strong
Abolitionism, which states that if no one is responsible, then we ought to both abandon
136 137 the reactive attitudes and adopt the objective stance—i.e., we ought to view one another as non-responsible parts of the natural world. In particular, I defend both abolitionism and strong abolitionism against what I take to be their strongest objection. This objection states that abandoning the reactive attitudes, in particular by adopting the objective stance, precludes the formation and maintenance of valuable personal relationships. The importance of these relationships gives us good reason not to abandon the reactive attitudes or adopt the objective stance, even if doing so has other benefits. Proponents of this argument claim (a) that a person who takes the objective stance is unable to take personally others’ attitudes and actions regarding her and (b) that doing so is necessary for certain valuable relationships. I dispute both claims. In particular, I argue that this objection exaggerates the role of reactive attitudes in relationships and underestimates the role of non-reactive emotions, many of which are consistent with the objective stance, and which play an important role in human social life.
In section 1, I explain what it means to take the objective stance. In section 2, I introduce the Personal Relationships Argument (PRA), which concludes that taking the objective stance precludes the formation of valuable personal relationships. I focus on
Seth Shabo’s (2012a) formulation of this argument and explicate its key premises. In section 3, I evaluate this formulation of the PRA, conclude that it unsound, and identify what I take to be the faulty premises. In section 4, I identify three tasks that an adequate personal relationships argument must accomplish and argue that none of them is accomplished by Shabo and that each is unlikely to accomplished by any other PRA.
Further, I clarify and defend Tamler Sommers’ conception of the objective stance and argue that taking it does not preclude the formation and maintenance of personal 138 relationships. Finally, in section 5, I raise a methodological worry about this type of anti- abolitionist argument.
5.1 The Objective Stance
In order to understand the additional requirement proposed by the strong abolitionist, and thus the implications of this position, we must understand what is means to take the objective stance. To take the objective stance is to adopt a universal and exclusively objective attitude toward the world—i.e., toward all agents and their attitudes and actions. According to Tamler Sommers (2007)—perhaps the only philosopher who has argued directly for what I call strong abolitionism—taking the objective stance
“requires us to view all people as natural objects who cannot be [responsible] for their character or behavior” (2007, 324). On Sommers’ conception, viewing a person as a natural object means understanding her as, in Thomas Nagel’s words, “nothing but a portion of the larger sequence of events” (Nagel 1979, 184 quoted in Sommers 2007,
323). Thus, taking the objective stance differs from abandoning the reactive attitudes in two main ways. First, the former requires denying that anyone is responsible in addition to refraining from taking attitudes, like resentment and guilt, that presuppose responsibility. Second, it requires viewing oneself and others as nodes in the causal web and not as somehow separate from it. However, taking the objective stance does not require denying that anyone is ever responsive to reasons or refraining from non-reactive 139 participant attitudes—i.e. participant attitudes that do not presuppose that their targets are responsible.80
Abolitionists appear to disagree both with anti-abolitionists and with one another about what is entailed by taking the objective stance toward a person. The strong abolitionist argues that if no one is responsible, then we ought to both abandon the reactive attitudes and adopt the objective stance. Anti-abolitionists, like Susan Wolf
(1981) and Seth Shabo (2012a), and other abolitionists, like Derk Pereboom (2001 and forthcoming), disagree and argue, to the contrary, that we are not obligated to take the objective stance because doing so precludes the formation and maintenance of valuable relationships. As we saw at the end of chapter 3, the source of the disagreement depends, in part, on how these philosophers conceive of the objective stance. (Though Shabo explicitly takes for granted much of Sommers’ conception (2012a, 137-8).) Part of my task in section 4 is to determine the extent and nature of the disagreement between different abolitionists and between strong abolitionists and their anti-abolitionist opponents.
5.2 The Personal Relationships Argument
I contend that the fact (if it is one) that no one is either free or responsible requires that we revise our judgments and adapt our practices in such a way that they are
80 Recall that Peter Strawson describes the reactive attitudes as, “the non-detached attitudes and reactions of people directly involved in transactions with each other” (1962, 62) or as attitudes of “involvement or participation” (66). Non-reactive participant attitudes are attitudes of this sort that do not presuppose responsibility. 140 consistent with this reality. I further contend that two required changes are the abandonment of the reactive attitudes and the adoption of the objective stance.
Anti-abolitionists disagree with these claims and argue that we are not required to abandon the reactive attitudes, much less by adopting objective stance. One reason for their disagreement is the belief that we cannot abandon the reactive attitudes without losing much more. Some philosophers have suggested that radically changing our behavior towards and conceptions of others would preclude the formation, or at least diminish the value, of the relationships we have with one another. For example, Peter
Strawson says, “A sustained objectivity of inter-personal attitude, and the human isolation which that would entail, does not seem to be something of which human beings would be capable, even if some general truth were a theoretical ground for it” (1962, 68).
And further, “[I]n the absence of any forms of these [reactive] attitudes it is doubtful whether we should have anything that we could find intelligible as a system of human relationships, as human society” (80). Derk Pereboom, an abolitionist, concedes that,
“Strawson is right to believe that objectivity of attitude would destroy interpersonal relationships” and that “the attitudes and analogs that would survive [my proposal to abandon the reactive attitudes] do not amount to Strawson’s objectivity of attitude, and are sufficient to sustain good interpersonal relationships” (2001, 200).
Susan Wolf (1981) eloquently and provocatively describes a world in which we have abandoned the reactive attitudes.
We can see that the abandonment of all the reactive attitudes would make a very great difference indeed. To replace our reactive attitudes with the objective attitude completely is to change drastically—or, as most would say, reduce—the quality of our involvement or participation in all our human relationships (390). 141
A world in which human relationships are restricted to those that can be formed and supported in the absence of the reactive attitudes is a world of human isolation so cold and dreary that any but the most cynical must shudder at the idea of it (391).
The picture she paints connects the absence of reactive attitudes to an absence of emotional warmth, depth, and richness that Wolf takes to be necessary for valuable relationships. One feature of her impoverished world, then, is the fact that the individuals who inhabit it are unable to form personal relationships.
Seth Shabo develops the descriptions given by Strawson and Wolf into an account of the psychology of close personal relationships that supports their shared worry about the strong abolitionist position advocated by Sommers. That is, Shabo argues that,
“suspending these attitudes by consistently viewing others in this way [from the objective stance] would compromise our ability to form and maintain meaningful personal relationships” (Shabo 2012a, 137) and that, therefore, we are not required to abandon the reactive attitudes, even if no one is either responsible. He argues that taking an exclusively objective stance toward a person precludes having a personal relationship with that person; and he argues that this phenomenon generalizes such that if everyone took a wholly objective attitude, no one (or almost no one) would be able to have personal relationships.81 Assuming that personal relationships are a particularly valuable
81 Shabo acknowledges that Sommers’ conception of the objective stance is less obviously incompatible with personal relationships than, for example, Strawson’s conception. He takes the argument described below to be adequate even if taking the objective stance does not involve, “assimilating ordinary adults to mere natural systems” as Strawson might have thought it did (Shabo 2012a, 138). 142 part of human life, the impossibility of such relationships undermines the abolitionist case for abandoning the reactive attitudes and for taking the objective stance.
Shabo argues that two important types of relationship are impossible if we take the objective stance.82 Specifically, he argues that taking the objective stance renders
“personal relationships,” which set includes the broad subsets of “mature friendship and reciprocal love,” impossible (2012a, 133). The reason that taking the objective stance purportedly renders such relationships impossible is that taking the objective attitude toward a person precludes taking her actions personally, which in turn precludes one’s
“caring in an essentially personal way” about the person’s actions or her attitudes toward one. And a relationship in which one party does not care in an essentially personal way about the attitudes and actions of the other party is an emotionally impoverished relationship and cannot be a mature friendship or relationship of reciprocal love.
On Shabo’s view, personal relationships require that each party care what the other does, and that they care in a way that is not required by more superficial relationships. For, while it may not matter to one teammate that another would rather spend time with a third acquaintance than with him, it will presumably matter to one romantic partner that his partner prefers to spend time with friends rather than with him.
A romantic relationship may require that it matters to Sally whether Jesse finds her
82 Shabo recognizes that taking the objective stance does not make personal relationships strictly impossible. Rather, he says that they are impossible in the sense that a “kind of interpersonal caring…[is]…excluded by the objective attitude” (Shabo 2012a, 131). He follows Strawson in thinking that such relationships are “practically inconceivable” for one who takes the objective stance (Strawson 1962, 68; Shabo 2012a, 132). I will often describe them as impossible simply for convenience, but, unless otherwise stated, this locution should be understood in the weaker sense meant by Strawson and Shabo. 143 sexually attractive. A mature friendship may require that it matters to Jesse whether
Raphael is a source of comfort in hard times. Whatever is in fact required, Shabo argues that a person who takes the objective stance toward her purported friend cannot take personally her failure to feel and behave as a friend should. He asks, “To what extent can we share in the struggles, joys, and sorrows of someone about whose thoughts and feelings towards us we don’t care in this way” (2012a, 140)? On his view, Jesse cannot care—she cannot take personally—that Raphael is not a source of comfort if she takes an objective stance toward his callousness and indifference.
Shabo’s argument that taking the objective stance precludes the formation and maintenance of personal relationships can be formalized in the following way.83
1. If one takes the objective stance toward an agent, then one cannot take her actions
personally.
2. If one does not take an agent’s actions personally, then one cannot care about her
actions or attitudes in an essentially personal way.
3. If one does not care in an essentially personal way about an agent’s actions or
attitudes toward one, then one’s relationship with her cannot qualify as a personal
relationship.
4. If one takes the objective stance toward an agent, then one cannot care about her
actions or attitudes in an essentially personal way [from 1 and 2].
5. Therefore, if one takes the objective stance toward an agent, then one’s
relationship with her cannot qualify as a personal relationship [from 3 and 4].
83 I have rewritten Shabo’s three main premises in order to make the form clearer and the content more explicit (2012a, 140). 144
In the next section, I argue that this argument is unsound.
5.3 A Defense of Abolitionism
In this section, I defend the strong abolitionist position against the Personal
Relationships Argument presented above.84 I argue that premises 1 and 3 are false.
However, before addressing the argument directly, let me briefly clarify exactly what is at issue in the argument and what is not. First, I do not deny that taking the objective stance is sometimes associated with a failure to form or maintain valuable relationships.
It is possible that I might be unable to form a close relationship with a victim of a severe stroke because typical forms of social interaction have become impossible. And I might be unable to relate to him because of his diminished language ability. His personality may be affected in a way that makes him unpleasant to me. Short-term memory problems may prevent us from advancing any bonds we form from day to day. I may take the objective stance toward him for all these reasons. However, I contend that taking the objective stance is not the reason why I am unable to form or maintain a close relationship in such a case. His disabilities are the reason.
Second, the anti-abolitionist can accept that some relationships are not precluded by taking the objective stance. The cognitive and emotional repertoire consistent with taking the objective stance is sufficient for some interpersonal relationships, namely, those defined entirely or primarily by what we might call “organizational relationships”
84 My defense will focus on the argument against taking the objective stance. However, because taking the objective stance entails abandoning the reactive attitudes, I take my objections to this argument to constitute a defense of both abolitionism and strong abolitionism. 145 between the members. For example, in order to be a teammate, coworker, or parishioner, one need only be a member of a team, office, or congregation, respectively. Nor is it just organizational relationships that are possible for those who take the objective stance.
Taking the objective stance is consistent with being a lover, a companion, or a caregiver.
For example, Sally can have a fulfilling sexual relationship with Jesse; Jesse can be an enjoyable vacation companion to Raphael; and Raphael can be an attentive and compassionate caregiver to Sally. Each of these relationships is possible even if the parties take the objective stance toward one another.
The significance of these points should not be underestimated. Such relationships—occupational, recreational, accidental, and occasional—constitute the majority of the relationships that many people will form in their lives. Moreover, unless one is lucky enough to work with one’s close friends or romantic partners, such relationships will likely constitute a majority of one’s social experience over the course of a lifetime—i.e., most of the hours spent socially will be with those who are not close friends or romantic partners.
Let me now return to the PRA and consider the first three premises in turn.
Premise 1 is false if one can take the objective stance toward an agent and still take personally her actions and attitudes. For example, it is false if the following scenario is likely or even possible. Sally takes the objective stance toward Jesse. When Jesse insults
Sally, Sally understands that Jesse is not responsible for choosing to do this. However,
Sally nonetheless recognizes that it was Jesse’s intention to insult her and her hope that the insult would cause Sally to feel sad. Moreover, Sally knows that Jesse would never 146 speak to her other friends like this. These two facts lead Sally to take Jesse’s insult personally. Such a case strikes me as quite likely.
However, premise 1 does have some apparent intuitive force. Consider two paradigmatic reasons for taking the objective stance toward a person, namely, perceived deficiency and a need for an impersonal relationship. When one takes the objective stance in response to a perceived deficiency relevant to the action, doing so does seem to preclude taking the action personally. For example, a father who takes the objective stance toward his young daughter may not take personally her claim to love her mother more than him. Similarly, when one takes the objective stance in virtue of the type of relation between oneself and the agent, it does seem like doing so precludes taking the other’s action personally. Indeed, Strawson’s conception of the objective stance as a
“refuge from the strains of involvement” seems to presuppose that, in taking it, one is not taking personally the other’s attitudes or actions (1962, 67). For example, a psychiatrist who takes the objective stance toward a patient may not take what the patient says personally (Shabo 2012a, 139).
In these cases taking the objective stance appears to preclude taking the offender’s actions personally. However, this appearance is deceiving. I suggest that in these cases it is the offender’s incapacity, not taking the objective stance, that explains the fact that one party does not take the other’s actions personally. To see this, consider another version of the hurtful daughter case.85 Suppose that, in addition to recognizing that his daughter
85 Derk Pereboom describes a similar case, involving a teenager and a parent, to dispute Shabo’s premise that abandoning the reactive attitudes precludes taking an action personally (forthcoming ms, 15). I agree with Pereboom that this is a good case. 147 has undeveloped social skills and is not responsible for the hurtful things she says, the father also recognizes that, in this case, his daughter means to hurt his feelings. Perhaps she feels slighted by some punishment and says something she knows will make him sad in order to retaliate. This is not inconceivable. Children are able to recognize how to hurt others before they have all the capacities typically associated with full moral responsibility. Nor is it inconceivable, or even unlikely, that the father could be hurt in a personal way by his daughter’s words. In both father-daughter cases the father takes the objective stance toward his daughter. However, only in the first case does he not take her comment personally. In the second case, he does because he recognizes in her the capacities relevant to taking her actions personally, though not the capacities necessary for holding her fully responsible.
At this point one might worry that the abolitionist and anti-abolitionist are speaking past each other and using the term “taking personally” differently. This is indeed possible. However, it is not a technical term with a precisely delineated meaning.
It is a commonsense concept about which the anti-abolitionist is reporting his intuition.
Shabo explicates caring in an essentially personal way—which, as we shall see, he takes to entail and be entailed by taking personally—by means of the following example:
When we care about someone’s attitudes toward us in an essentially personal way, those attitudes matter to us in their own right, quite apart from what they portend for our (and others’) interests. Thus, someone who hasn’t received an invitation to a social event might wonder at the apparent snub, even if he or she has little desire to attend and realizes that the situation has no bearing on his or her social or professional prospects” (2012b, 112).
However, I would go further and suggest that such a scenario is consistent with taking the objective stance. 148
That is, when we take an action personally, we care about the attitude that the action betrays in addition to the inconvenience or hurt it may cause. (Compare the way that some non-consequentialists take the intentions behind an action to matter over and above its harmful consequences.) My objection to premise 1 is based on the intuition that the father described above can take the objective stance toward his daughter, but also take personally the hurtful things she says. He takes her insult personally insofar as he believes it was targeted at him personally, insofar as he believes he has special reason (as her father) not to be hurt in this way, and insofar as he believes her intent was purely malicious (though not responsibly so). In short, he takes it personally insofar as he thinks it betrays some ill will. The above case demonstrates that we can judge a person to have a responsibility undermining incapacity, but nonetheless take her actions personally.
In general, premise 1 appears to rely on the assumption that taking the objective stance entails detachment from the person toward whom one is taking it. Strawson, for example, conceives of the reactive stance as involving attitudes “of involvement or participation” and contrasts this with the objective stance, which involves isolating oneself in various ways (1962, 62). One who takes the objective stance is thought to be detached in the sense that she deploys a diminished or restrained emotional repertoire and, as a result, is incapable of forming deep or rich relationships.86
I agree that taking the objective stance involves deploying a diminished emotional repertoire. Reactive emotions like resentment, indignation, and guilt must be inhibited or abandoned. However, I deny that this diminished emotional repertoire constitutes a
86 Susan Wolf conceives of the objective attitude as detached in this way (1981, 390- 391). 149 sufficient level of detachment to preclude personal relationships. Recall the father whose daughter insults him. He may take the objective stance toward her and maintain a degree of detachment from her. However, the degree of detachment he is likely to maintain is not sufficient to preclude a personal relationship. Perhaps such extreme detachment is possible, but it is not entailed by taking the objective stance.
Moreover, while taking the objective stance entails abandoning the reactive attitudes, it does not entail abandoning the affective repertoire involved in taking reactive attitudes. Reactive attitudes are emotions, and emotions have a cognitive (appraisal) component and an affective (feeling) component (Kalat and Shiota 2007). For example, resentment is typically characterized as involving both the belief that a person responsibly wronged me and a feeling of anger at the perceived wrong. But abandoning resentment does not require abandoning anger. Anger may be the affective component of non-reactive emotions like frustration. While taking the objective stance does diminish one’s emotional repertoire, it does not diminish the affective repertoire from which emotions are partially composed. There is thus a range of non-detached emotions available to the person who takes the objective stance, including at least some of those emotions that have both affective and cognitive components that are about the target’s actions or attitudes, but which do not presuppose responsibility.87 For example, we can express sadness, disappointment, or frustration toward a child who fails to share her toys.
For all of these reasons, I believe that premise 1 of the PRA is false.
87 Strawson himself acknowledges that the objective attitude can be “emotionally toned” in many ways (1962, 66). 150
Premise 2 states that if one does not take an agent’s actions personally, then one cannot care about her actions or attitudes in an essentially personal way. These two phrases strike me as two ways of saying the same thing.88 I do not think it even makes sense to say that Sally cares in an essentially personal way about Jesse’s insult but that she does not take the insult personally. (Importantly, though, my suggestion is that taking personally is the same as caring in an essentially personal way. I do not suggest that taking personally is the same as caring about, period. We care about many things that we do not take personally. I may care deeply about our treatment of the environment, the distant needy, or animals, but not take personally any of the harms that befall them.) Henceforth, I will assume that premise 2 is true and that taking personally entails caring in an essentially personal way.
Premise 3 is false if one does not care in an essentially personal way about another’s attitudes or actions but is still able to have a personal relationship with that person. For example, it is false if the following scenario is possible. Jesse verbally abuses Sally. Her abuse appears to Sally to be malicious. However, Sally does not really care in an essentially personal way about Jesse’s apparent malice or the insults that
88 Shabo’s own account implies, if not synonymy, then at least a conceptual connection between taking personally and caring in an essentially personal way. He says, “If we are not disposed to taking anything someone says or does personally (or if we subdue this disposition to the greatest psychologically possible degree at all times), we cannot be said [my emphasis] to care in an essentially personal way about that person’s actions and attitudes toward us” (2012a, 140). Elsewhere he is even more explicit. “I take this to be conceptually true: if one person is immune to taking another’s behavior personally (where this isn’t an aberration), the first person does not care about the second’s behavior and attitudes toward him or her in this essentially personal, emotionally engaged way” (Shabo 2012b, 113). This second statement endorses an interpretation of the conditional as stating a conceptual truth. The first indicates that Shabo imagines the interpretation to apply biconditionally. 151 express that malice. Nonetheless, she is still very close with Jesse because they have known each other since they were very young and their shared history means a lot to
Sally. This case seems to me like one that might well exist between many people. After all, there are many things besides bad attitudes and blameworthy actions that one might care about and that are just as relevant to the formation and maintenance of close relationships. The PRA says that if Sally does not care in an essentially personal way about Jesse’s actions toward her, then Sally is not a “mature friend” of Jesse. But it is compatible with the fact that Sally does not care in an essentially personal way about
Jesse’s attitudes or actions that other facts about their relationship are true, including the following: Sally wants to be in a personal relationship with Jesse. Sally cares about whether Jesse does wrong by her. Sally cares whether Jesse does what Sally reasonably expects of her. Sally cares for Jesse—she is fond of her, has affection for her, usually enjoys her company, etc. Sally feels safe when Jesse is around and is comforted by her presence.
I suggest that various combinations of these facts are sufficient for the formation and maintenance of a personal relationship, like a mature friendship. Consider a more elaborate version of the scenario. Sally has known Jesse for many years. She knows that
Jesse is, in some ways, a mean and wrathful person. She also believes that Jesse is not responsible for being this way. Her upbringing was such that it was overwhelmingly likely that she would turn out this way. However, the same upbringing that has made
Jesse angry and mean has also made her fiercely loyal to those whom she trusts enough to get close. Moreover, she is no more responsible for these positive traits than for her failings. Because they grew up together, Sally and Jesse have always been an important 152 part of each other’s lives. They share their trials and tribulations as well as their hopes and successes. Sally is fond of Jesse; she cares what happens to her; she wishes her well.
She also generally likes Jesse’s company, at least when Jesse is not being mean. They have many similar interests and opinions, though Sally disapproves of Jesse’s maliciousness, and the time they spend together is, for the most part, fun for both of them. For all this, Sally is not immune to Jesse’s paranoia, anger, and malice. Jesse often says hurtful things to Sally, sometimes with the intent to make her sad or angry.
Sally has grown used to this, though, and does not care about these attacks in an essentially personal way. She does not take her friend’s anger personally.
I suggest that this is an accurate description of a possible relationship. I suspect that it is an accurate description of many actual relationships. Moreover, these relationships, whether actual or merely possible, qualify as personal relationships on any reasonable delineation of Shabo’s intuitive class. The question then is whether they are valuable relationships and, if so, how valuable they are relative to other possible relationships. While the relationship between Sally and Jesse is undesirable in many ways, neither the fact that Sally does not take personally Jesse’s malicious acts nor the fact that Jesse is not responsible for being malicious seems to explain why it is undesirable. The explanation lies in Jesse’s maliciousness itself, not in her non- responsibility.
Finally, even if one thinks that a valuable personal relationship could never form so long as one party does not care in an essentially personal way about the other’s actions and attitudes, I would contend that if such a relationship already existed, it could be maintained even if one party entirely ceases to care in an essentially personal way about 153 the other’s actions and attitudes. Relationships like the one between Sally and Jesse do exist in the actual world and it seems quite plausible that they could persist despite one party (or both) deciding to take the objective stance.89
On the basis of these arguments against premises 1 and 3, I conclude that Shabo’s formulation of the PRA is unsound. And, insofar as Shabo’s formulation was intended to defend the as yet undefended claims about the impoverishment of human life by taking the objective stance made by Strawson and Wolf, I conclude that a sound formulation of the PRA has not yet been offered.
5.4 Value, Scope, and Incompatibility
In the last section, I argued that Shabo’s formulation of the PRA is unsound. The aim of this section is more general, namely, to evaluate personal relationship arguments as a type. There are three claims that I think any formulation of the PRA must establish.
In this section, I consider each of them and argue both that they are not adequately established by Shabo’s formulation and, more importantly, that they are not likely to be established by any PRA. First, any adequate PRA must establish that some type of relationship, in this case personal relationships, is impossible if one takes the objective
89 Of course, this is not to deny that there are relationships in which one party cares deeply for the other and whose actions demonstrate his caring, but which nonetheless do not constitute a personal relationship of the sort Shabo has in mind. For example, suppose a wealthy man comes to care deeply for a poor orphan and becomes her secret benefactor. His care may be genuine and it may be clearly manifested by his beneficent actions, but the two parties cannot be said to have a personal relationship in Shabo’s sense. A relationship of this sort exists in the novel Daddy-Long-Legs by Jean Webster. Thanks to Dana Nelkin for drawing my attention to this point. 154 stance.90 Call this the Incompatibility Problem. Second, it must correctly delineate the class of impossible relationships and explain which relationships are made impossible when one takes the objective stance and why it is only those. For it is surely not the case that all relationships are impossible for a person who takes the objective stance. Call this the Scope Problem. Finally, it must demonstrate that these relationships are indeed valuable and that their loss is a significant sacrifice. For it is not obvious that a relationship built on reactive attitudes has the same value it is typically thought to have if no one is responsible. Call this the Value Problem. Shabo, for his part, makes his value claim as modest as possible by delineating the class of important relationships in advance and by choosing types of relationships—mature friendship and reciprocal love—that are intuitively the most valuable. I argue that all three problems undermine Shabo’s argument and pose a serious obstacle to the establishment any formulation of the PRA.
5.4.1 The Value Problem
I begin with the assumption that personal relationships are necessarily valuable and that their loss is a significant sacrifice. Shabo argues that adopting the objective stance makes it impossible for a person to form or maintain a mature friendship or
90 A different formulation of the PRA might focus instead, or additionally, on those relationships that require mutual respect. One could plausibly argue that at least some forms of respect are precluded by taking the objective stance. For example, in the Groundwork of the Metaphysics of Morals, Kant reserves the word respect for assessments of autonomous action and the practical law itself (4:400-401 and 4:436). However, there are also many forms of and grounds for respect (and self-respect) that are not precluded by the objective stance. For example, I may respect my partner in virtue of her compassion, intelligence, open-mindedness, or benevolence, none of which seems to require the reactive attitudes. 155 reciprocal love relationship with another. Let us grant this claim for now. Lacking personal relationships certainly seems bad, but recall the reasons why we are considering the objective stance in the first place. If no one is responsible, then taking reactive attitudes is unfair! Negative reactive attitudes are blaming attitudes and blaming a person for a non-blameworthy action is unfair. It is unfair because it is an unwarranted and unbalanced harm. Insofar as overt blame is a form of punishment and punishment is harmful, a blaming attitude is a harm. It is also degrading. Blaming someone puts that person in the category of people to be disapproved of. In so doing, it declares that they are less worthy of certain forms of respect than others. Moreover, blaming the non- blameworthy may be unjust independently of whether it results in a concrete degradation.
The disrespect of being inappropriately blamed is also unfair in the way that slander, libel, or false promising are unfair. Even when such actions do not result in concrete personal or financial harms, they are unfair in that they communicate a lack of appropriate basic respect that moral agents are expected to show one another. Such actions communicate the judgment that the victim is not owed such basic respect.
Relationships built on unfairness are bad, or at least worse than those that are not, all else equal. And the more pervasive or oppressive the unfairness, the worse the relationship. Relationships whose existence is premised on unfairness are not often the subject of spirited defense. We criticize the patriarchal commonplace of women performing all or the large majority of housework in relationships where both partners work full time. We criticize the distressingly common experience of women who face the difficult and unfair choice of remaining with an abusive partner or abandoning their children to an abusive father (Dressler 2006). Indeed, we criticize any relationship in 156 which one partner does not recognize and treat the other as his moral equal, whether this involves division of labor, physical or psychological abuse, coercion, neglect, or simple disrespect. Most relevant to my current point, whether or not we are willing to criticize every form of unfairness we find in relationships, we do criticize relationships in which either party is subject to unwarranted or excessive blame from the other.
The proponent of the PRA must defend one of three claims about personal relationships. It is bad that taking the objective stance makes such relationships impossible because they are not unfair. It is bad that taking the objective stance makes relationships impossible because there is no better alternative than this unfairness—i.e., the relationships are flawed, but are still the best source of such value that we have. Or, it is bad that we are not free and responsible because freely and responsibly formed and maintained personal relationships are good. Which of these claims does (or should) the anti-abolitionist accept? No one seems to disagree that taking a reactive attitude toward a person who is not responsible is unfair. Nor does Shabo challenge the unfairness claim.
Perhaps, then, the anti-abolitionist accepts that the endemic unfairness of reactive attitude-based relationships is the best we can do.
But the unfairness of a typical reactive attitude-based relationship is not the best we can do. No one denies that it is better for a parent to abandon her resentment when she learns that the person who broke her laptop was a child. Nor would anyone deny that resentment toward a spouse who failed to meet an unreasonable domestic demand ought to be abandoned (or inhibited). Nor would one think to defend resentment in such cases by suggesting that abandoning resentment would make the relationship between the subject and the target of the attitude—e.g., parent and child or husband and wife— 157 impossible. For the parent-child relationship is not impossible without resentment and patriarchal entitlement is not a significant loss.91 Similarly, we can at least sometimes do better in personal relationships by abandoning the reactive attitudes and taking the objective stance. And, to the degree that personal relationships require reactive attitudes, we can do better by abandoning these relationships.
Perhaps we think that there is some difference in how the two types of relationships—patriarchal entitlement vs. reactive attitude-based friendship—are unfair.
Taking reactive attitudes is at least a practice or standard that is applied equally to everyone, while the unfairness of patriarchal entitlement arises from the application of different norms to men and women. This is true, but does not justify taking reactive attitudes. Resenting a person who is not responsible for her actions is morally objectionable and resenting all non-responsible people is worse, not better, regardless of the equality of the practice. Distributing bads equally is not better than not doing bads things.
The anti-abolitionist may object that I have not yet argued that abandoning the reactive attitudes makes unfair personal relationships better. However, I would suggest that the burden of proof is on him. He has not argued that abandoning the reactive attitudes makes unfair personal relationships worse. Rather, he has argued that it makes them impossible. Moreover, I have argued that abandoning inappropriate reactive attitudes makes some unfair relationships better—e.g., those in which one partner resents
91 Admittedly, neither of these cases involves a relationship that Shabo or others would say requires reactive attitudes. However, my aim here is merely to demonstrate that relationships that involve unfair reactive attitudes can be made better by refraining from or inhibiting those reactive attitudes. 158 another for failing to perform some non-obligatory task or those in which one friend causes another to feel undeserved shame. We have been given no reason to think that such cases do not generalize to personal relationships. Moreover, any argument that personal relationships based on reactive attitudes are valuable enough to outweigh their unfairness faces the additional burden of weighing the fact that most relationships are not personal relationships and most interpersonal encounters are with people other than our friends and romantic partners. And those relationships will benefit from abandoning the reactive attitudes.
The anti-abolitionist may instead object that what he is defending is not the value of any individual instance of taking a reactive attitude, but rather the value of the disposition to take reactive attitudes. Abandoning the disposition is too great a sacrifice because it is the disposition to respond from the reactive stance that is valuable.
However, while she need not do so, the abolitionist can accept this. It is consistent with the abolitionist position that an individual have a morally appropriate disposition to take reactive attitudes. A disposition can be more or less sensitive and we typically think of a disposition to take reactive attitudes as appropriate only if it is within an appropriate range of sensitivity. For example, even assuming that we are all free and responsible, a person who is disposed to resent anyone who questions the truth of his claims is inappropriately disposed to resentment. One might reasonably say that a person is appropriately disposed to take the reactive attitudes only if she inhibits or renounces such attitudes toward those she believes to be neither free nor responsible. And the abolitionist can accept that this disposition is appropriate. The abolitionist view is that one may be disposed to take reactive attitudes, but never take them because an 159 appropriate situation never arises—perhaps in the way that a person can possess a virtuous disposition to martial courage, but never have the opportunity to act on that disposition because of a lack of martial conflicts.92
The proponent of the PRA has one remaining option. He may argue that what is bad is not that taking the objective stance makes valuable relationships impossible but rather that the fact that no one is free or responsible means that fair, reactive attitude- based relationships are impossible. However, the abolitionist can happily agree to this claim. The abolitionist can accept that it would be better to be free and responsible and to have the sort of relationships these facts would allow. She merely claims that, if we are not, we ought to abandon the reactive attitudes and adopt the objective stance.
5.4.2 The Scope Problem
I turn now to the assumption that the PRA applies only to the class of personal relationships and does not generalize beyond this class. Shabo suggests that there is something special about personal relationships. He says that, “we can distinguish—at least in favourable cases—relationships of mature friendship and reciprocal love from associations based chiefly on mutual convenience or attraction, shared personal or professional interests, general collegiality or amiability, and so on” (2012a, 133 n.5).
And he suggests that this distinction matches the distinction between relationships that
92 I realize that it is controversial whether a person could have a virtue or psychological disposition upon which she never has, and could never have, the opportunity to act. However, I cannot enter this debate in this chapter. Instead, I rely solely on the intuitive sense that a person can be disposed to respond in a particular way to a type of situation that may never or can never arise. 160 are made impossible by taking the objective stance and those that are not. I contend, to the contrary, that the argument to the effect that the objective stance makes personal relationships impossible also applies to other relationships, if it applies at all.
Of course, the anti-abolitionist could simply define personal relationships as those made impossible by taking the objective stance. The Scope Problem could be solved by defining personal relationships as those which require whatever it is that the objective stance precludes. On Shabo’s view, this is the ability of parties in the relationship to take one another’s actions personally.93 The anti-abolitionist might state the definition thus: a relationship is a personal relationship if and only if it requires the parties to take personally one another’s actions to an extent rendered impossible by taking the objective stance. This is one way to avoid the Scope Problem. However, this is not a promising strategy for the anti-abolitionist. He is not making a purely conceptual claim; it is partly empirical.94 He is claiming that taking the objective stance has an effect on how human beings are able to interact with one another and that the limitations it places on these
93 Inability to “take personally” is not necessarily the only feature of taking the objective stance that is relevant to forming or maintaining personal relationships. However, it is a broad concept that seems to capture what, if anything, intuitively goes wrong relationship-wise when one takes the objective stance. I cannot think of other problems with taking the objective stance that would not fall under the admittedly broad umbrella of “taking personally”. 94 As a matter of fact, Shabo does seem to understand some of his claims as conceptual rather than empirical. For example, he says that, “We could not form or maintain personal relationships if we lost or relinquished our ordinary susceptibility to a significant range of personally engaged attitudes” (2012b, 101). Shabo holds that ‘could not’ should be taken as a conceptual impossibility claim. Here Shabo is claiming that being in a personal relationship just is to be susceptible to a range of personally engaged attitudes. However, this is not to claim, and it does not follow as a conceptual matter, that to be in a personal relationship is to be susceptible to the reactive attitudes. The connection between the personally engaged attitudes and reactive attitudes like resentment is, he admits, psychological (2012b, 114). 161 interactions preclude certain relationships. It renders a person unable to interact in the necessary ways. Nor would an alternative anti-abolitionist position be more plausible if it made the purely conceptual claim rather than the partly empirical one. The above definition is ad hoc and unmotivated by any aim other than fitting the view.
A more promising way to avoid the Scope Problem is to say that the objective stance precludes some feature necessary to personal relationships and not necessary to any other relationships. Let us assume that the necessary feature that the objective stance precludes is the ability to take actions and attitudes personally. The question is whether the class of relationships that do require taking personally, assuming that class has any members, is limited to mature friendships and reciprocal love relationships. I think it is not so limited. For example, if reciprocal love relationships require the ability to take personally, then non-reciprocal ones plausibly require it, too. Romeo’s unrequited love for Rosaline requires that he take her indifference personally. If he were indifferent rather than despondent in response to her indifference, it would be reasonable to say he was not in love with her. At least it would be as reasonable as it would be to say that he was not in love with Juliet if he did not take her love for him personally.
Similarly, if mature friendships require the parties to take one another’s actions personally, so do many (and perhaps all) friendships. Indeed, the initial formation of a friendship might require that the parties do this. Shabo notes the importance of vulnerability to the formation of close relationships and it is as likely to be important at the initial stages of a friendship as at the transition from immature to mature friendship. I suspect that there may indeed be a point (or period) in a friendship where taking personally (in some way or to some degree) indicates or inaugurates a deeper, richer 162 friendship. However, a similar change in taking personally may also indicate or inaugurate the beginning of a friendship—that is, it may mark the difference between being friends and not being friends.
In this subsection, I have argued that it is difficult to limit the scope of application of the PRA. The best candidate for providing a principled distinction between relationships precluded by taking the objective stance and those not precluded is that some relationships require parties to take personally one another’s actions and attitudes.
However, while this is a principled distinction, the line it draws is not between “mature friendships and romantic relationships,” on the one hand, and every other type of relationship, on the other. This is important because it forces the anti-abolitionist to make a stronger claim, namely, that many types of relationship are precluded by taking the objective stance. Of course, the proponent of the PRA may simply agree and claim that his argument establishes this stronger conclusion. However, in the absence of a clear delineation of the class of supposedly precluded relationships, it seems more likely that taking the objective stance does not preclude any type of relationship than that it precludes all types, or all but “organizational relationships” (see section 3).
5.4.3 The Incompatibility Problem
In this subsection, I take for granted that taking the objective stance does not preclude taking personally and instead consider the plausibility of developing a PRA that does not employ claims about detachment, taking personally, or caring in an essentially personal way. So, why else might taking the objective stance make personal relationships impossible? On Sommers’ view, taking the objective stance entails 163 abandoning the reactive attitudes, denying that anyone is responsible, and viewing others as “natural objects”. One might worry, though, that viewing people as natural objects does preclude personal relationships, even if the degree of emotional detachment entailed by the objective stance does not. However, I will show that this worry is unfounded.
Sommers’ conception of the objective stance does not support the claim that taking the objective stance precludes personal relationships.95
In order for the practice of viewing one another as natural objects to explain why taking the objective stance precludes personal relationships, it must be the case both that, in virtue of viewing people in this way, a) taking the objective stance entails X, and that b) X is sufficient to preclude a personal relationship with that person. I will argue that taking the objective stance, understood as viewing people as natural objects, entails no X such that both (a) and (b) true.
In order to establish the incompatibility claim stated above, the PRA must show that, insofar as one views a person as a natural object, one cannot form a personal relationship with her.96 As a matter of fact, there are number of ways in which viewing
95 One might argue that the following description of what happens when a person takes the objective stance describes the same phenomenon that Shabo calls taking personally, or failing to take personally. If so, then the same arguments and examples used against Shabo apply here. I see the following as a distinct proposal, though I do think that they are subject to similar objections and counterexamples. 96 R.S. Downie gives one interpretation of what it means to view someone as a natural object. “[I]t is tempting to say that what promotes the objective attitudes is the belief that the operation of these factors renders the behavior open to sufficient explanation in causal terms…Indeed, it is not clear that an attitude can be identified as objective in Strawson’s sense unless we assume that it is that which fittingly arises with the awareness that behavior can be sufficiently explained in causal terms” (1966, 34). However, contrary to what Downie suggests, it is simply not true that taking the objective stance toward a person entails believing that her behavior is sufficiently explained in causal rather than 164 someone as a natural object could preclude forming a personal relationship with that person—e.g., if doing so meant viewing them as not a person, not sentient, or not alive.
However, if this is what it means to view someone as a natural object, then it is not the case that taking the objective stance toward someone entails viewing her as a natural object. Taking the objective stance toward someone does not entail viewing her as not alive, not sentient, or not a person. In this case (b) is true, but (a) is not. Moreover, while it is indeed true that we do not, and perhaps cannot, form personal relationships with some natural objects, it is not clear that the objective stance is what explains this fact.
The fact that I cannot form a mature friendship with my Chia Pet is not explained by the fact that I take the objective stance toward it. It is explained, if by anything, by my inability to relate to it socially, which is not precluded by the objective stance.
I suspect that there is some confusion on both sides about Sommers’ conception of the objective stance and of what it means to view something or someone as a natural object. There is both a descriptive and a normative claim to make here. The descriptive claim has two parts. First, Sommers misleadingly implies that we take the objective stance toward all natural objects, but it is doubtful that we do this. After all that would mean that we are constantly taking the objective stance toward, for example, every locked door, angry cat, and cloudy sky. Second, even if we do take the objective stance toward most natural objects, the objective stance we take toward falling branches and
purposive terms. After all, I may take the objective stance because I believe that her action is unexplained or unexplainable. Indeed, we might think that this is what the psychiatrist does. Her patient’s behavior is insufficiently explained by his explicit intentions and identifiable values, so she treats him as a natural object whose functioning she does not yet understand. 165 inconveniently placed coffee tables is plausibly different than that we take toward persons’ intentional actions. The attitude I take toward a tree that sheds a branch on my car is, just as a matter of fact, very different than the attitude I take toward an insane criminal, despite the fact that neither is responsible and both are natural objects. The normative claim is that we are justified in taking these different forms of the objective stance toward different natural objects. For, while neither the tree nor the insane person is responsible for their behavior, they are relevantly different. Our attitudes, whether reactive or objective, are not merely responses to the perceived responsibility of entities for their behavior; they are responses to other properties of those entities, such as their desires and intentions. So long as her condition is not too severe, the mental states antecedent to the insane person’s actions, her desires and intentions, are morally relevant and evaluable. The causal antecedents of the branch falling from the tree are not. Insofar as it makes sense to condemn a person for her intention to harm me, while refraining from blaming her, my attitude toward her will be different from my attitude toward the tree, about whom neither blame nor any sort of moral appraisal makes sense.
Having considered, albeit briefly, this explanation of why taking the objective stance precludes the formation and maintenance of personal relationships, I conclude that it does not adequately explain the incompatibility of personal relationships and taking the objective stance. And I conclude that, given my earlier conclusion about detachment and taking personally, on no plausible account of the nature of the objective stance is it practically inconceivable that we could take the objective stance and still have personal relationships. These conclusions leave the abolitionist position on solid ground with respect to the PRA. 166
In this section, I identified three claims that an adequate PRA must establish and argued that none of them has been or promises to be established. First, I have suggested that the relationships we would purportedly be sacrificing by taking the objective stance are not valuable enough that the cost of sacrificing them gives us reason to deny abolitionism (the Value Problem). Second, it appears that a PRA generalizes beyond personal relationships and is less plausible because it makes this broader and stronger claim (the Scope Problem). Third, I have argued that on no plausible account of the nature of the objective stance is it practically inconceivable that we could take the objective stance and still have personal relationships (the Incompatibility Problem).
These conclusions leave the abolitionist position on solid ground with respect to the PRA.
4.5 Methodological Epilogue
The debate about whether universal adoption of the objective stance precludes the possibility of valuable personal relationships rests on an assumption that deserves closer scrutiny. I have left consideration of this assumption to the end, choosing instead to engage the anti-abolitionist on his own terms and dispute the content of his objection.
However, in this final section, I consider a possible methodological limitation of the PRA and, by extension, of my own defense. In particular, I consider the assumption that the compatibility of the objective stance with personal relationships can be assessed without reference to empirical evidence about friendship and romantic relationships.
The claim that taking the objective stance precludes the formation and maintenance of valuable personal relationships is, at least in part, an empirical claim. It is a claim about what human beings are like and what certain kinds of relationships 167 between them require in order to form and flourish. And the PRA is supported by an argument two of whose premises have an empirical dimension (premises 1 and 3). As such, it is a considerable assumption that such a claim can be established without substantial empirical evidence. Indeed, one could go so far as to say that it is a claim for which armchair speculation can provide at best a suggestive and anecdotal account.
Adequately establishing claims about how friendships form and under what conditions might reasonably be thought to require anthropological, sociological, and psychological evidence, none of which has been adduced in support of the anti-abolitionist position.97
97 Empirical evidence of this sort is available. See, inter alia, Fehr (2004), Hall (2011), and Zarbatany, Conley, and Pepper (2004). Beverly Fehr aims to identify expectations for intimate same-sex friendships by assessing which patterns of relating to one another we take to be constitutive of friendships. The results of her prototype analysis suggest that the patterns of relating most commonly identified by subjects themselves—e.g., “if I need to talk, my friend will listen” or “if I need support my friend will provide it”—do not appear to require the experience or expression of reactive attitudes. Rather the vast majority of subject-identified expectations for intimate same-sex friends concern emotional, physical, or social support (2004, 270). Nonetheless, more evidence is needed if we are to conclude that no reactive attitudes are necessary. Some of the expectations listed do mention or imply reactive attitudes—e.g., “If I do something wrong, my friend will forgive me” (2004, 270). Hall’s meta-analysis of the data about friendship expectations identifies four central friendship expectations: symmetrical reciprocity, communion, solidarity, and agency. Reactive attitudes are not mentioned, though they may play a role in meeting these expectations. For example, trust and loyalty are forms of symmetrical reciprocity, but it is an open empirical question whether these traits require, say, resentment in cases of disloyalty or broken trust (2011, 725). Finally, Zarbatany, Conley, and Pepper conclude on the basis of their results that, “close same-sex friendship is not a homogenous experience but is shaped by the dispositions of its constituent members” (2004, 308). Thus, while it is important for a friendship to meet common expectations, the individual personalities of the parties also determine what the expectations are and how important each is to a particular friendship. This brief empirical sketch is at most suggestive. However, it does point to both the need for and the possibility of further empirical investigation by abolitionists and anti-abolitionists alike into the role of the reactive attitudes in close personal relationships. 168
At this point, the anti-abolitionist may claim that concepts like ‘reactive attitude’ or ‘objective stance’ are philosophical terms of art and do not directly map onto whatever empirical concepts are used to analyze actual human relationships. For example, one might argue that any empirically established conclusions about what kind or degree of emotional connection is necessary for a mature friendship to form will not translate into claims about which reactive attitudes are necessary or to what degree we can take the objective stance without precluding or damaging such friendships. However, this response is unsatisfactory. The fact remains that anti-abolitionists are committed to empirical claims in addition to their conceptual claims about what counts as friendship.
If the philosophical vocabulary in which these claims are couched does not map onto the vocabulary employed by those who study relationships empirically, then it is incumbent upon philosophers to provide such a mapping. And, to the extent that doing so is impossible, we must accept that the philosophical claims are not necessarily about the empirical phenomena we care about. But if the PIC is instead a purely conceptual claim, then it takes on a much stronger meaning. Understood as a purely conceptual claim, the view that taking the objective stance precludes personal relationships requires a deeper analysis of each of the terms, which analysis has not yet been offered either by abolitionists or their opponents.98
98 If it is a purely conceptual claim that the objective stance precludes personal relationships, then it looks much more like, for example, the claim that being a bachelor precludes being married. It is true that being a bachelor precludes being in a married relationship, but it is not the case that a bachelor is unable to be married. Rather, as a conceptual matter, a bachelor cannot both get married and remain a bachelor. It would be implausible for the abolitionist to make a parallel claim about one who takes the objective stance, especially without a deeper analysis of the terms. It is much more plausible to 169
Of course, the PRA is not entirely an empirical matter. The nature of friendship is, at least in part, a conceptual question and the objective stance is a philosophical construct.99 However, neither is impervious to empirical evidence. Whether or not our accepted conceptions of ‘personal relationship’ and ‘objective stance’ are conceptually compatible, it is relevant to the plausibility of abolitionism which actual relationships are instances of personal relationships. In both cases, we must ask whether the category of personal relationships includes the types of actual relationships that we value. And in both cases, answering this question will require empirical evidence.
I do not take this methodological objection to be decisive against any formulation of the PRA. Nor do I take the empirical literature on close relationships to point univocally toward the abolitionist conclusion. However, at the very least, both the assumption that we can determine the compatibility of the objective stance with personal relationships without reference to the empirical evidence about friendship and romantic relationships and the evidence itself deserve more attention than they have received thus far from either side.
5.6 Conclusion
In this chapter, I have defended strong abolitionism against the claim that taking the objective stance precludes the formation and maintenance of valuable personal relationships—and, by extension, against the claim that not taking reactive attitudes
take the anti-abolitionist as claiming that a person who takes the objective stance is, as a result, unable to have personal relationships. 99 Thanks to Dana Nelkin for emphasizing this important point. 170 precludes such relationships. I have argued that Seth Shabo’s (2012a) formulation of the
Personal Relationships Argument is unsound. Further, I have argued that concerns about viewing one another as natural objects are unfounded and that doing so also does not preclude the formation, maintenance, or flourishing of personal relationships. Finally, I have suggested that any plausible argument that taking the objective stance precludes the formation and maintenance of personal relationships must provide at least some empirical support for this claim.
These conclusions should prompt us to reflect on the relationships we most prize—our mature friendships, romantic partnerships, and close family relationships.
These relationships, it seems to me, are characterized, in different measures, by desire, trust, loyalty, kindness, empathy, shared history, healthy communication, enjoyment of the other’s company, and, seemingly most important, the many ways in which we support one another. All of these things remain possible, as do the various forms of love that depend upon them, whether we abandon the reactive attitudes or not, and whether we take the objective stance or not.
Chapter 6
Conclusion
In this dissertation, I have defended abolitionism—the view that if no one is responsible, then we ought to abandon the reactive attitudes. I have argued that we have good reasons, both moral and epistemological, to think that abolitionism is true and that these reasons are stronger than any of the countervailing reasons identified by anti- abolitionists. In the preceding chapters, I have explicated the abolitionist view and some of its variants and I have responded on behalf of this view to the main anti-abolitionist arguments, namely, what I have called the Impossibility Argument, Peritrope Argument,
Social Harmony Argument, and Personal Relationships Argument. However, while my central task in these chapters has been negative, repelling attacks and shoring up defenses, I have also tried to make a positive case for abolitionism. In this brief conclusion, I wish to draw together these positive points and paint a picture of what a world without the reactive attitudes would be like.
Let me begin with a brief summary of four main arguments against abolitionism and my conclusions about their implications for the abolitionist conditional. The first two are impossibility arguments and the latter two are outcome arguments.
The Impossibility Argument: we are not able to abandon the reactive attitudes, so,
because ought implies can, we are not obligated to abandon them, even if no one
is responsible; therefore abolitionism is false.
171 172
The Social Harmony Argument: the losses of abandoning the reactive attitudes outweigh the gains of doing so because the reactive attitudes are the most effective means to securing social benefits like trust, cooperation, and fair play; therefore, even if no one is responsible, it is not the case that we ought to abandon the reactive attitudes; so abolitionism is false.
The Personal Relationships Argument: the losses of abandoning the reactive attitudes outweigh the gains because abandoning the reactive attitudes—and especially taking the objective stance—precludes the formation and maintenance of valuable personal relationships; therefore, even if no one is responsible, it is not the case that we ought to abandon the reactive attitudes; so abolitionism—especially strong abolitionism—is false.
I have argued that abolitionism does not succumb to these threats. The mere existence of these arguments illustrates that, while heterodox and radical, abolitionism cannot simply be written off. It is a well-founded philosophical position that deserves the time and attention it has received. Moreover, understanding where the arguments succeed and where they fail can help us understand what abolition truly entails.
I have argued that neither of the impossibility arguments succeeds in undermining abolitionism, though the Peritrope Argument may establish a limit on strong abolitionism. We can, and some people do, eliminate reactive attitudes and dispositions from our emotional repertoire. Complete abolition will probably be very difficult and perhaps no individual has ever achieved it, but there is little reason to think we are psychologically incapable of doing so, either as individuals or as a species. However, the success of one variation of the Peritrope Argument does establish a more limited impossibility claim. If we are all not only non-responsible, but non-reasons-responsive, then taking the objective stance in response to the epistemological reason for abolitionism involves a performative contradiction. As a result, strong abolitionism may be impossible in some sense and that there is a limit on abolitionism. 173
I have also argued that neither of the outcome arguments succeeds. Neither the
Social Harmony Argument nor the Personal Relationships Argument establishes the evaluative claim that the world would be better all things considered if we retained the reactive attitudes. I suggest that both the descriptive and the normative elements of the two arguments are inadequate and that they fail to establish their central conclusion.
With Pereboom, I argued that non-reactive attitudes (e.g., regret) can effectively fill the roles left vacant by abolished reactive attitudes (e.g., guilt). Against Strawson and others
I argued that, on the most plausible interpretation of what the objective stance is and of what is required for a personal relationship, we can in fact form and maintain personal relationships even from the objective stance.
Perhaps more important is the fact that both outcome arguments rely on questionable normative claims. The Social Harmony Argument relies on the illicit normative assumption that the value of a reactive attitude is properly assessed by considering the effect of its absence on our current social practices, rather the effect of its absence on a set of practices revised in response to its obsolescence and governed by a revised framework of non-reactive attitudes. Likewise, the Personal Relationships
Argument fails to consider that the value of close relationships is typically assessed on the assumption that both parties are responsible, which, by hypothesis, is false and which, as a result, should change our evaluation of the relationships. It seems, then, that sussing out the gains and losses to human life of the abolitionist proposal is harder than it appears at first glance and that a reassessment is unlikely to vindicate the Strawsonian picture of an existence impoverished by the absence of the reactive attitudes. 174
What lessons should we draw from these conclusions? First, impossibility claims and outcome claims are distinct, but often reinforce one another. The purported inconceivability of abandoning the reactive attitudes influences evaluative judgments about the relative benefits of retention versus divestment. If we cannot imagine life without them, then we cannot imagine what might take their place and we cannot begin to assess the value of these replacements. Likewise, our assessments of the value of the reactive attitudes influence the “practical conceivability” of abandoning them. Their centrality to our emotional lives is mistaken for inevitability. Distinguishing impossibility claims from outcome claims allows us to evaluate them independently using the tools appropriate to each.
Second, the reactive/objective distinction is a false and misleading dichotomy.
The class of non-reactive participant attitudes—including Pereboom’s analogue attitudes—is a real and crucial part of our emotional repertoires. Recognizing this class of attitudes undermines much of the force of Strawson’s Practical Inconceivability Claim.
While a world without participant attitudes—or social emotions, where that label includes interpersonal, self-directed, and vicarious emotions—may be inconceivable, a world without reactive attitudes is not. Abandoning the reactive/objective dichotomy also has implications for the outcome arguments. If we reject the dichotomy Strawson proposes, then we must reassess outcome claims not only in light of the third, and much larger, category of non-reactive attitudes, but also using new conceptions of the original two categories. In particular, we must reassess what it means to take the objective stance and whether doing so is consistent with taking non-reactive participant attitudes. 175
Third, abolitionism can be more or less accommodating or revisionary. The abolitionist proposals made by hard incompatibilists like Derk Pereboom and Tamler
Sommers have tended toward accommodating our existing practices, whether of cooperation, friendship, or self-assessment. Pereboom argues against strong abolitionism and for analogue attitudes upon which supervene the same, or very similar, relationships and self-conceptions. Sommers proposes a stronger form of abolitionism, according to which we ought to adopt the objective stance in addition to abandoning the reactive attitudes. However, his actual proposals and justifications do not diverge substantially from those advocated by Pereboom. Like Pereboom and Sommers, I argued in chapter 4 that many of the criticisms Nichols makes against non-reactive analogues are misguided.
For example, I developed accounts of regret and moral resolve that show how
Pereboom’s proposed analogues are more robust than Nichols’ empirical objections would have us believe. Alternatively, however, one might pursue a more revisionary and less accommodating abolitionism.
I took two small steps in this direction in this dissertation. In chapters 4 and 5
(sections 4.2 and 5.1), I argued that mistaken assumptions have led us to inflate the value and importance of reactive attitudes and personal relationships, respectively. We take their value for granted when, in fact, they are tarnished by wrongful blaming and mistaken attributions of psychological and/or metaphysical properties. The implication of these mistaken value judgments is that, while abandoning the reactive attitudes might allow us to accommodate our existing social practices, a more revisionary alternative program is also plausible. The failure (if it is one) of the non-reactive analogues to fill the roles left vacant by abandoned reactive attitudes is only an objection to abolitionism 176 insofar as we think these roles must be filled. However, if the abolitionist simply disagrees with Nichols that the behaviors associated with retaliatory emotions are important or valuable, then she need not attempt to accommodate them in her abolitionist emotional repertoire or the pattern of behavior and social practices that it supports.
Finally, if abolitionism is true, then skeptics about free will and responsibility have a job to do. If it is indeed possible to abandon the reactive attitudes, and if the balance of gains and losses for human existence does favor abolition, then the skeptic must address the concrete moral problem of the injustice of retaining the reactive attitudes. If these things are true, then our natural disposition to take reactive attitudes is a moral problem in the same way that institutional punishment is a moral problem.
Neither is justified in its traditional form and the existence of each requires personal, social, and/or political action.
What does abolitionism envision? What does a world without the reactive attitudes look like in the light of the abolitionist’s conclusions? A world in which we accept the abolitionist conditional is one in which we recognize the importance of reevaluating our framework of reactive attitudes and the social practices that supervene upon it. A world in we accept both the abolitionist conditional and skepticism about free will and responsibility is one in which we pursue the abolition of the reactive attitudes.
Such a world is possible. With effort, societies and individuals can abolish the reactive attitudes. It will take more than the personal projects of a few committed Stoics and Buddhists, but we can divest ourselves of our traditional reactive habits and adopt new ones. 177
This imagined world will not revert to a state of nature. Cooperation amongst self-interested individuals and groups will not disintegrate at either the global or the local level. Familiar forms of human interaction will remain. Members of the moral community will retain concepts of right and wrong and trade in these concepts as they do now. They will condemn most of the same harms and promote most of the same goods.
And the emotional framework upon which all of these practices supervene will function in essentially the same way. Emotions will continue to express both our appraisals of the world and of people as well as our feelings about them.
It is true that some of the institutions that have developed on the backs of, and which thereby reflect the structure of, our interpersonal emotional practices will have to be reformed. In this, abolitionism unsurprisingly has similar implications to its namesake in the punishment literature (Duff 2001, 30-34). However, as Pereboom and others have noted, retributivist theories do not have exclusive claim to either moral or practical plausibility—i.e., to either moral justification or adequate deterrent and protective policies consistent with those non-retributive theories (Pereboom 2001; Duff 2001). It is also true that interpersonal emotional practices will themselves have to be reformed. For example, we will have to actually replace guilt with some emotion that does not presuppose responsibility (like regret), as I discussed in chapter 4. However, the striking similarities in content and function between the reactive and non-reactive attitudes in many such cases are promising. Also promising are recent extensions of the very studies that drive the Social Harmony Argument. Nichols argued that our deployment of moral anger in ultimatum games is evidence that it is an important basis for trust and cooperation and that, therefore, abandoning the anger-based attitudes is unwise. 178
However, recent studies have identified other variables, notably serotonin levels, that affect how individuals in these games respond to perceived unfairness (Crockett et al.
2008). Because serotonin levels can fluctuate naturally in response to everyday circumstances, it is not at all clear whether one or the other response to unfairness (i.e., anger-based retaliation or no retaliation) is either natural or beneficial in the ways
Nichols suggests. The complex physical processes upon which our social practices of trust and cooperation supervene remain to be explored. In the meantime, there is good reason to believe that there are many routes to these benefits, only some of which involve reactive attitudes.
Susan Wolf suggested that life without the reactive attitudes would be, “a world of human isolation so cold and dreary that any but the most cynical must shudder at the idea of it” (1981, 391). I have argued, building upon Sommers own response to Wolf, that this is simply false. Life without the reactive attitudes remains rich and various. My awe at the Nærøyfjord remains; my sense of accomplishment upon completing a marathon is undiminished by the absence of self-praising pride; and the experience of seeing the water lilies or hearing “Kind of Blue” is just as captivating without an attribution of responsibility. It is true that contemplation of some wonders, whether of music, writing, or human ingenuity, brings us to a point at which we might think to praise and ask, “Is it not laudable that she accomplished this feat?” But it is also true, I think, that often this inclination to praise is one thought too many and that it is more fitting to respond as we do to the beauty and sophistication of the natural world and to wonder, “Is this not simply astounding?” It is this perspective on the world that Nagel and Sommers 179 nicely capture, whether purposefully or not, with their intuition that to take the objective stance is to view human beings as “natural objects”. We are extraordinary things.
Nor is life without the reactive attitudes isolated. Far from it. Compassion for those who are suffering, benevolence towards the needy, admiration for those who sacrifice, and disappointment with those who do not; all these emotions remain and their expression—whether toward strangers, loved ones, or ourselves—connects us to one another and puts us in conversation with one another, just as do resentment and gratitude.
And the same is true for the emotions that make possible our personal relationships, whether friendships, romantic relationships, or any of the other “occasional relationships,” as I have called them, that shape our days and lives. I mentioned in the previous chapter, and want to emphasize again now, that unless one is lucky enough to work with one’s close friends or romantic partners, it is our occasional relationships that occupy most of our time. And these relationships are even less defined by reactive attitudes than personal relationships.
The relationships we value most—our close friendships, romantic partnerships, and family relationships—are characterized, in different measures, by desire, trust, loyalty, kindness, empathy, shared history, healthy communication, enjoyment of the other’s company, and, seemingly most important, the many ways in which we support one another. All of these things remain possible, as do the various forms of love that depend upon them, whether we abandon the reactive attitudes or not.
When we see the world in this way, we can see that Strawson got his description of optimism and pessimism the wrong way round. He rightly saw that true pessimists fear not that determinism will undermine responsibility, but rather that it will undermine 180 those valuable things that require responsibility. However, Strawson and his followers are pessimists of this sort; they see a world without freedom and responsibility, whatever that may mean, as one in which we cannot help but wrongfully blame and mistakenly attribute properties and capacities to one another that we do not have. Abolitionist skeptics are true optimists. They see a world without freedom or responsibility as a promising candidate for reform. The picture I paint of the world is an attractive one, in which our ascriptions of praise and blame match the circumstances of our responsibility.
It is a world in which we can be touched, as Abraham Lincoln said, by the better angels of our nature and, in doing so, can augment the bonds of our mutual affection.
References
Arneson, Richard J. 2003. “The Smart Theory of Moral Responsibility and Desert” in Desert and Justice. Ed. Serena Olsaretti. New York: Clarendon Press. 233-258.
Aune, Bruce. 1967. “Hypotheticals and ‘Can’: Another Look” in Free Will. Ed. Gary Watson. 1982. New York: Oxford UP. 36-40.
Bok, Hilary. 1998. Freedom and Responsibility. Princeton: Princeton UP.
Brink, David O. 2013. “Responsibility, Incompetence, and Psychopathy.” The Lindley Lectures.
Brink, David O. and Dana K. Nelkin. 2013. “Fairness and the Architecture of Responsibility” in Oxford Studies in Agency and Responsibility. Ed. David Shoemaker. New York: Oxford UP. 284-313.
Burnyeat, M.F. 1976. “Protagoras and Self-Refutation in Later Greek Philosophy.” The Philosophical Review 85.1. 44-69.
Butler, Joseph. 1900. “Sermon IX: Upon Forgiveness of Injuries” in The Works of Bishop Butler. Ed. J.H. Bernard. London: MacMillan & Co., Limited. 112-24.
Crockett, Molly J., Luke Clark, Golnaz Tabibnia, Matthew D. Lieberman, and Trevor W. Robbins. 2008. “Seratonin Modulates Behavioral Reactions to Unfairness.” Science 320. 1739
Darby, Ryan S., Britta A. Larsen, Per-Erik Milam, Dana K. Nelkin, Christine R. Harris, and Nicholas J.S. Christenfeld. (unpublished manuscript). “Forgive or Forget: Is Forgiveness More than Letting Go or Excusing?”
Dennett, Daniel C. 1973. “Mechanism and Responsibility” in Free Will. Ed. Gary Watson. 1982. New York: Oxford UP. 150-173.
Downie, R.S. 1966. "Objective and Reactive Attitudes." Analysis 27.2. 33-39.
Dressler, Joshua. 2006. “Battered Women and Sleeping Abusers: Some Reflections.” Ohio State Journal of Criminal Law 3. 457-472.
Duff, R.A. 2001. Punishment, Communication, and Community. New York: Oxford UP.
Ehlers, Anke and David M. Clark. 2000. “A cognitive model of post-traumatic stress disorder.” Behavioral Research and Therapy 38. 319-45.
181 182
Eisenberg, Nancy, Sandra Losoya, Richard A. Fabes, Ivanna K. Guthrie, Mark Reiser, Bridget Murphy, Stephanie A. Shepard, Rick Poulin, and Sarah J. Padgett. 2001. “Parental Socialization of Children's Dysregulated Expression of Emotion and Externalizing Problems.” Journal of Family Psychology 15.2. 183-205.
Ekman, Paul, Richard J. Davidson, Matthieu Ricard, and B. Alan Wallace. 2005. “Buddhist and psychological perspective on emotion and well-being.” Current Directions in Psychological Science 14.2. 59-63.
Epictetus. 1991. Enchiridion. Trans. George Long. Amherst, New York: Prometheus Books.
Epicurus. 1994. Letter to Menoeceus in The Epicurus Reader. Trans. Brad Inwood and L.P. Gerson. Indianapolis: Hackett Publishing Company, Inc.
Epicurus. 1994. On Nature book 25 in The Epicurus Reader. Trans. Brad Inwood and L.P. Gerson. Indianapolis: Hackett Publishing Company, Inc.
Fehr, Beverley. "Intimacy expectations in same-sex friendships: A prototype interaction pattern model." Journal of Personality and Social Psychology 86.2 (2004): 265- 284.
Fehr, E. and Fischbacher, U. 2004. “Third party punishment and social norms.” Evolution and Human Behavior 25. 63-87.
Fehr, E. and Gachter, S. 2002. “Altruistic Punishment in Humans.” Nature 415. 137- 140.
Fincham, Frank D. 2010. “Forgiveness: Integral to a Science of Close Relationships” in Prosocial Motives, Emotions, and Behavior. Eds. Mario Mikulincer and Phillip R. Shaver. Washington DC: American Psychological Association.
Fischer, John Martin and Mark Ravizza. 1998. Responsibility and Control. New York: Cambridge UP.
Freedman, Paul. 1998. “Peasant Anger in the Late Middle Ages” in Anger’s Past: The Social Uses of an Emotion in the Middle Ages. Ed. Barbara H. Rosenwein. Ithaca: Cornell UP, 1998. 171-188.
Gino, Francesca, Maurice E. Schweitzer, Nicole L. Mead, and Dan Ariely. 2011. "Unable to resist temptation: How self-control depletion promotes unethical behavior." Organizational Behavior and Human Decision Processes 115.2. 191-203. 183
Goldman, Alvin I. 1969. "The Compatibility of Mechanism and Purpose." The Philosophical Review 78.4. 468-482.
Hagger, Martin S. 2010. “Ego Depletion and the Strength Model of Self-Control: A Meta-Analysis.” Psychological Bulletin 136.4. 495-525.
Hall, Jeffrey A. 2011. "Sex differences in friendship expectations: A meta-analysis." Journal of Social and Personal Relationships 28.6. 723-747.
Hampton, Jean. 1993. “Selflessness and the Loss of Self.” Social Philosophy and Policy 10.1. 135-165.
Hare, Brian, Victoria Wobber, and Richard Wrangham. 2012. "The self-domestication hypothesis: evolution of bonobo psychology is due to selection against aggression." Animal Behaviour 83.3. 573-585.
Job, Veronika, Carol S. Dweck, and Gregory M. Walton. 2010. "Ego depletion—Is it all in your head? Implicit theories about willpower affect self-regulation." Psychological Science 21.11. 1686-1693.
Kalat, James W., and Michelle N. Shiota. 2007. Emotion. Belmont, CA: Thomson Wadsworth.
Kane, Robert. 1996. The Significance of Free Will. New York: Oxford UP.
Kant, Immanuel. 1997. Groundwork of the Metaphysics of Morals. Trans. Mary Gregor. New York: Cambridge UP.
Kim, Jaegwon. 1989. “Mechanism, Purpose, and Explanatory Exclusion.” Philosophical Perspectives 3. 77-108.
Kim, Jaegwon. 1996. Philosophy of Mind. Boulder, CO: Westview Press.
Kim, Jaegwon. 1998. Mind in a Physical World: An Essay on the Mind-Body Problem and Mental Causation. Boston: MIT Press.
Kyle, Diana J., and Heike I.M. Mahler. 1996. "The effects of hair color and cosmetic use on perceptions of a female's ability." Psychology of Women Quarterly 20.3. 447- 455.
Larsen, Britta A., Ryan S. Darby, Christine R. Harris, Dana K. Nelkin, Per-Erik Milam, Nicholas J.S. Christenfeld. 2012. "The immediate and delayed cardiovascular benefits of forgiving." Psychosomatic Medicine 74.7. 745-750. 184
Lockwood, Penelope. 2006. ''Someone Like Me can be Successful'': Do College Students Need Same-Gender Role Models?” Psychology of Women Quarterly 30. 36-46.
Machamer, Peter, Lindley Darden, and Carl F. Craver. 2000. “Thinking about Mechanisms.” Philosophy of Science 67.1. 1-25.
Malcolm, Norman. 1968. “The Conceivability of Mechanism” in Free Will. Ed. Gary Watson. New York: Oxford UP, 1982. 127-149.
McKenna, Michael. 2012. Conversation and Responsibility. New York: Oxford UP.
Mele, Alfred R. 2006. Free Will and Luck. New York: Oxford UP.
Moore, G.E. 1912. “Free Will” in Principia Ethica. Ed. Thomas Baldwin. 1993. New York: Cambridge UP.
Murphy, Jeffrie. 1988. Forgiveness and Mercy. New York: Cambridge UP.
Nagel, Thomas. 1979. “Moral Luck” in Free Will. Ed. Gary Watson. 1982. New York: Oxford UP. 174-186.
Nelkin, Dana K. 2008. “Responsibility and Rational Abilities: Defending an Asymmetrical View.” Pacific Philosophical Quarterly 89.4. 497-515.
Nelkin, Dana K. 2011. Making Sense of Freedom and Responsibility. New York: Oxford UP.
Nichols, Shawn. 2007. “After Incompatibilism: A Naturalistic Defense of the Reactive Attitudes.” Philosophical Perspectives 21.1. 405-428.
Nisbett, Richard E., and Dov Cohen. Culture of honor: The psychology of violence in the South. Westview Press, 1996.
Norcross, Alastair. 2013. “Consequentialism and Friendship” in Thinking About Friendship: Historical and Contemporary Philosophical Perspectives. Ed. Damien Caluori. New York: Palgrave Macmillan. 161-179.
Nozick, Robert. 1989. The Examined Life. New York: Simon and Schuster.
Okin, Susan Moller. 1989. Justice, Gender, and the Family. Basic Books.
Pereboom, Derk. 2001. Living Without Free Will. New York: Cambridge UP. 185
Pereboom, Derk. 2008. “A Hard-Line Reply to the Multiple-Case Manipulation Argument.” Philosophy and Phenomenological Research 77.1. 160-170.
Pereboom, Derk. 2009. “Free Will, Love, and Anger.” Ideas y Valores 58. 169-189.
Pereboom, Derk. (unpublished manuscript). Free Will Skepticism, Agency, and Meaning in Life. Oxford: Oxford UP.
Rosenwein, Barbara H. 1998. Anger’s Past: The Social Uses of an Emotion in the Middle Ages. Ithaca: Cornell UP.
Sabini, John, and Maury Silver. 2005. "Why emotion names and experiences don't neatly pair." Psychological Inquiry 16.1. 1-10.
Schmeichel, Brandon J., and Kathleen Vohs. 2009. "Self-affirmation and self-control: affirming core values counteracts ego depletion." Journal of Personality and Social Psychology 96.4. 770-782.
Sedley, David. 1983. “Epicurus’ Refutation of Determinism.” SUZETESIS. 11-51.
Shabo, Seth. 2012a. “Incompatibilism and Personal Relationships: Another Look at Strawson's Objective Attitude.” Australasian Journal of Philosophy 90.1.131-147.
Shabo, Seth. 2012b. “Where Love and Resentment Meet: Strawson's Intrapersonal Defense of Compatibilism.” Philosophical Review 121.1. 95-124.
Smart, J.J.C. 1961. “Free Will, Praise, and Blame” in Free Will 2nd ed. Ed. Gary Watson. New York: Oxford UP, 2003.
Smilansky, Saul. 2000. Free Will and Illusion. New York: Oxford UP.
Sommers, Tamler. 2007. "The Objective Attitude." The Philosophical Quarterly 57.228. 321-341.
Strawson, Galen. 1994. “The Impossibility of Moral Responsibility” in Free Will 2nd ed. Ed. Gary Watson. New York: Oxford UP, 2003. 212-228.
Strawson, P.F. 1962. “Freedom and Resentment” in Free Will. Ed. Gary Watson. 1982. New York: Oxford UP. 59-80.
Stroop, J.R. 1935. “Studies of interference in serial verbal reactions.” Journal of Experimental Psychology 18.6. 643-662. 186
Talbert, Matthew. 2012. “Moral Competence, Moral Blame, and Protest.” Journal of Ethics 16.1. 89-109.
Van Kleef, Gerben A., Carsten KW De Dreu, and Antony SR Manstead. 2006. "Supplication and appeasement in conflict and negotiation: The interpersonal effects of disappointment, worry, guilt, and regret." Journal of Personality and Social Psychology 91.1. 124-142.
Van Oyen Witvliet, Charlotte, Thomas E. Ludwig, and Kelly L. Vander Laan. 2001. "Granting forgiveness or harboring grudges: Implications for emotion, physiology, and health." Psychological Science 12.2. 117-123.
Vihvelin, Kadri. 2004. “Free Will Demystified: A Dispositional Account” in The Philosophy of Free Will: Essential Readings from the Contemporary Debates. Eds. Paul Russell and Oisín Deery. 2013. New York: Oxford UP. 166-191.
Wallace, R. Jay. 1994. Responsibility and the Moral Sentiments. Cambridge: Harvard UP.
White, Stephen D. 1998. “The Politics of Anger” in Anger’s Past: The Social Uses of an Emotion in the Middle Ages. Ed. Barbara H. Rosenwein. Ithaca: Cornell UP, 1998. 127-152.
Wolf, Susan. 1981. “The Importance of Free will.” Mind 90.3. 386-405.
Yaffe, Gideon. 2010. Attempts: in the philosophy of action and the criminal law. New York: Oxford UP.
Zarbatany, Lynne, Ryanne Conley, and Susan Pepper. 2004. "Personality and gender differences in friendship needs and experiences in preadolescence and young adulthood." International Journal of Behavioral Development 28.4. 299-310.
Zeelenberg, Marcel, and Jane Beattie. 1997. "Consequences of regret aversion 2: Additional evidence for effects of feedback on decision making." Organizational Behavior and Human Decision Processes 72.1. 63-78.
Zeelenberg, Marcel, and Seger M. Breugelmans. 2008. "The role of interpersonal harm in distinguishing regret from guilt." Emotion 8.5. 589-596.
Zeelenberg, Marcel, and Rik Pieters. 2004. "Consequences of regret aversion in real life: The case of the Dutch postcode lottery." Organizational Behavior and Human Decision Processes 93.2. 155-168. 187
Zeelenberg, Marcel, Wilco W. van Dijk, Antony S.R. Manstead, and Joop van der Pligt. 2000. "On bad decisions and disconfirmed expectancies: The psychology of regret and disappointment." Cognition & Emotion 14.4. 521-541.
Zeelenberg, Marcel, Kees van den Bos, Eric van Dijk, and Rik Pieters. 2002. "The inaction effect in the psychology of regret." Journal of Personality and Social Psychology 82.3. 314-327.