II. Terminological Stage Setting

Before turning to the burdens of a successful prescriptive eliminativist view about moral responsibility, a bit of terminological housecleaning is in order. What I am here calling descriptive skepticism and prescriptive eliminativism calls for at least some clarification. By descriptive skepticism I mean any and all views that identify some kind of significant error regarding our standing concept or ordinary language attributions of “moral responsibility.” This characterization allows for the possibility of both conceptually motivated forms of skepticism, and ontologically motivated versions.^{Footnote 4} The former often involve identification of some significant error in our folk concept of moral responsibility.^{Footnote 5} Conceptually motivated skeptics claim that our concept is fragmented, naturalistically or normatively implausible, or downright incoherent or inconsistent.^{Footnote 6} I take most of the skeptical views about moral responsibility currently on offer to be conceptually motivated. But explicitly ontologically motivated versions of skepticism about moral responsibility might also be possible. Rather than pointing to some error in our concept, one might argue that the term “moral responsibility” fails to refer, and do so in a way that is largely silent on our widespread conceptual commitments. This path to skepticism is more common in regard to empirically verifiable kind terms,^{Footnote 7} and so it is not clear precisely how an ontologically motivated skeptical view of moral responsibility might go.^{Footnote 8} Rather than assess the plausibility of this path to eliminativism I wish here only to note it as a point in conceptual space. In what follows I will set it aside and focus exclusively on conceptually motivated descriptive skepticism about moral responsibility. Regardless of the path to skepticism, what both conceptually and ontologically motivated varieties have in common is that they identify some fatal error in either our folk concept of moral responsibility, or our discourse involving the term “moral responsibility.” The precise nature of the error may be a point of disagreement, but where skeptics converge is on the claim that there is such an error.

By prescriptive eliminativism I mean to pick out views that draw a substantive conclusion about what we should do in light of the purported error identified by some variety of skepticism. Namely, the error in question is sufficient to block a “business as usual” stance concerning our responsibility-related attitudes, discourse, and normative practices. I discuss the forms that such preservationist prescriptive alternatives to eliminativism might take in further detail below. The defining feature of eliminativism is that we should reject such attempts to sustain our responsibility-related attitudes and practices in light of the purported error(s) identified by some form of descriptive skepticism. Eliminativists claim instead that we should abandon these attitudes and practices, at least to the extent that we are able, rather than attempt to fix, justify, or otherwise sustain them.

So, prescriptive eliminativist views presuppose some variety of descriptive skepticism. But this dependence does not run in both directions. Mere descriptive skepticism, without an additional eliminativist prescriptive conclusion, is a coherent position. I turn now to a closer look at why this is so for descriptive skepticism about moral responsibility in particular.

III. The Limits of Descriptive Skepticism: Fixing the Skeptical Spotlight

Descriptive skepticism about moral responsibility fails on its own to entail prescriptive eliminativism. It is possible to carve out a coherent position such that our concept of responsibility suffers some significant error, while at the same time denying that we should eliminate responsibility from our discourse, attitudes, and practices insofar as we are able. The diagnostic portion of Manuel Vargas’s revisionist program, for example, might be categorized as a form of mere descriptive skepticism.^{Footnote 9} On this reading, Vargas offers a conceptually motivated brand of descriptive skepticism in particular—our folk concept of moral responsibility is significantly libertarian and these libertarian strands render the concept naturalistically and normatively implausible.^{Footnote 10} Despite this problematic feature, Vargas argues that we need not be moved to eliminativism. Instead, we might either perform a kind of conceptual surgery (connotational revision^{Footnote 11}), or a kind of conceptual transplant (denotational revision^{Footnote 12}). Regarding the former, connotational revision, we might excise the problematic libertarian features from our concept in a way that still preserves the reference of “moral responsibility” and the work of this concept to sustain our responsibility-related discourse, attitudes, and practices. Regarding the latter, denotational revision, we might abandon our standing concept of moral responsibility but “re-anchor” reference with a distinct but similar nearby concept capable of doing the same work as the original to sustain our responsibility-related discourse, attitudes, and practices. According to Vargas,^{Footnote 13} either strategy represents a live preservationist prescriptive alternative to eliminativism despite the significantly problematic features our folk concept identified. And so if Vargas is right, even if we embrace a descriptive skeptical position about moral responsibility, we have at least two live preservationist alternatives to prescriptive eliminativism.

The move from mere skepticism to full-blown eliminativism requires at least some further work. A closer look at the logical gap between these two positions can also serve to explicate one essential step in bridging it. However, at this point it will be helpful to move from general talk of skepticism and eliminativism to a more concrete instance of the former. Here I will use Ishtiyaque Haji’s luck-based view as an instructive example. Haji motivates a kind of dual skepticism about obligation and blameworthiness by identifying a conflict between the pervasiveness of luck and three independently plausible commitments: that motivation requires ability, that a version of ought-implies-can is correct, and that blameworthiness requires impermissibility. While Haji concludes that impermissibility and blameworthiness are threatened most by this conflict, here I will argue that this kind of skeptical conclusion requires a further comparative defense of the commitments that generate it. Given that Haji himself takes each of these commitments to be independently plausible, why think that it is the connection between blameworthiness and impermissibility in particular that cannot be salvaged?^{Footnote 14}

Haji’s skeptical argument proceeds as follows^{Footnote 15}:

1. 1. Regardless of the truth of determinism or indeterminism, luck is pervasive; the way that we are, our actions, and the circumstances in which our actions occur are often out of our control.

2. 2. If luck is pervasive, then we frequently lack alternatives.

3. 3. Obligation requires alternatives.

4. 4. So, the range of obligation in our world is, at best, significantly less than we take it to be.

5. 5. But, there is a deontic requirement for moral responsibility.

6. 6. There are two plausible ways of understanding this deontic requirement: (a) blameworthiness requires impermissibility, or (b) blameworthiness requires nonculpable belief in impermissibility.

7. 7. If the deontic requirement is understood in terms of (a), then luck’s restriction of the range of our obligations also restricts the range of our responsibility.

8. 8. If the deontic requirement is understood in terms of (b), then blameworthiness could be sustained only by our being mistaken about the primary moral statuses (obligatory, permissible, impermissible, and so on) of our actions.

9. 9. And so, at best blameworthiness “hangs on a thread.”

Haji’s argument here is a kind of dual skepticism, in that he claims that luck threatens obligation and, via this threat to obligation and some additional principles, so too does it threaten blameworthiness and moral responsibility. For my purposes the key premises of the argument are (2), (3), and (5), which are each supported by what Haji takes to be a relatively uncontroversial principle. In regard to (2), we have Motivation/Ability:

Haji assumes that the desires we have and can acquire at a time are often a matter of luck (regardless of the truth of determinism or indeterminism). If one cannot intentionally perform an action that one has no desire (and cannot acquire a desire) to perform, and whether or not we have (or can acquire) such a desire is often itself a matter of luck, then the pervasiveness of luck significantly restricts our alternatives.

In support of premise (3) Haji appeals to Kant’s Law:

If Kant’s Law is correct, then obligation requires alternatives. In conjunction with Motivation/Ability and the pervasiveness of luck, Kant’s Law then delivers the intermediate conclusion that obligation “hangs on a thread.” In other words, if we assume that some version of Kant’s Law is correct and obligation requires alternatives, and luck rules out alternatives, then luck also undermines the range of our obligations. At best, the scope of our obligations in the actual world is significantly less than we take it to be.

But, if blameworthiness requires impermissibility then this restriction on the range of our obligation also restricts the range of our responsibility. And Haji argues that there is such a deontic requirement on blameworthiness. The initial formulation of this requirement is the following:

Haji considers, but ultimately rejects an objective interpretation (premise 6(a) in the argument above) of Blameworthiness/Impermissibility and instead defends a complex subjectivist interpretation:

But even if this subjective interpretation of Blameworthiness/Impermissibility (premise 6(b) in the argument above) is correct, a significant restriction on the scope of our obligation would in turn significantly restrict the scope of our blameworthiness. If, as Haji argues, the scope of our obligation is significantly restricted, then we are often mistaken when we (even nonculpably) believe that many of our actions are impermissible. And so while this kind of belief would leave open the possibility of meeting the deontic condition on blameworthiness, it would still leave blameworthiness precariously dependent on a kind of widespread “deontic irrationality.”^{Footnote 20} Thus, blameworthiness, for Haji, like obligation, is seriously threatened by the pervasiveness of luck.

One of the many merits of Haji’s argument is that it explicitly identifies the problematic feature of moral responsibility that motivates a concrete variety of skepticism: the commitment to some version of Blameworthiness/Impermissibility. Further, Haji’s argument is also explicit about why this feature is problematic. It is in direct tension with the pervasiveness of luck and at least two other independently plausible commitments: Motivation/Ability and Kant’s Law. However, Haji notably says very little about the prescriptive upshot of this argument. In fact, he leaves us only with the following remarks:

Haji himself seems aware of the gap between the skeptical position he argues for and further prescriptive conclusions. While the final questions he poses suggest the possibility of prescriptive eliminativism, he does not explicitly endorse such a move. And the details of the argument for skepticism sketched above allow us to see clearly why doing so requires further work. As things stand, Haji has identified an inconsistent set of conceptual commitments. A particular feature of our concept of responsibility—some version of Blameworthiness/Impermissibility—is in direct tension with the conjunction of two additional independently plausible commitments: Motivation/Ability and Kant’s Law. If we are to maintain our commitment to these two principles, and Blameworthiness/Impermissibility is a constitutive feature of our folk concept of moral responsibility, then indeed we have a problem. Then, that is, Haji has successfully identified a significant error in our folk concept and thus motivated a descriptive skeptical position. Moral responsibility cannot be what we thought it was, at least if we hold fixed other important background commitments such as our commitment to Motivation/Ability and Kant’s Law.^{Footnote 22}

But should we hold fixed these other background commitments? Motivating the move from skepticism to full-blown eliminativism requires, as a first step, an affirmative answer to this question. The fact that a set of our standing—even widespread and independently plausible—commitments are inconsistent tells us little about how we should resolve the relevant inconsistency. Perhaps in this case the best way to do so is instead to abandon Motivation/Ability or Kant’s Law.^{Footnote 23} Upon further consideration we might have better reason to give up one of these two commitments than to give up Blameworthiness/Impermissibility. If so, then we have no clear reason to think that this commitment, and along with it moral responsibility, ought to be abandoned.

This possibility highlights an interesting feature of skeptical arguments more generally: they are vulnerable to what I will call the shifting skeptical spotlight. One shifts the spotlight when one argues that we should resolve the inconsistency identified by the skeptic by rejecting a different plank of the inconsistency. This is of course akin to G. E. Moore’s famous response to the skeptic about knowledge.^{Footnote 24} When faced with an apparent inconsistency between a plausible definition of knowledge and the belief that I know anything at all, the skeptic claims that I should resolve the inconsistency by giving up the latter. But, as Moore points out, this is not the only possibility. When we examine the inconsistent set more closely (and see that the belief that I have two hands entails the belief that I know anything at all) perhaps we should conclude that we have better reason to shift the skeptical spotlight to a different member of the set. If one must choose between giving up a plausible definition of knowledge and giving up the belief that one has two hands, Moore makes the persuasive case that perhaps we should take a closer look at the former. It is open to us to shift the skeptical spotlight to what we previously took to be an independently plausible definition of knowledge, and away from the seemingly commonsensical belief with which we now see that it stands in tension.

In Haji’s case, we might grant the plausibility of his overall skeptical argument, but resist a further prescriptive move to eliminativism about blameworthiness and moral responsibility. Rather than abandoning a deontic requirement on blameworthiness, perhaps we have better reason to shift the skeptical spotlight to one or both of the independent commitments it is at odds with: Motivation/Ability and Kant’s Law.^{Footnote 25}

This discussion serves to highlight one important feature of the eliminativist’s burden: given the identification of some significantly problematic feature of our concept of moral responsibility, eliminativists must show further that this feature, and not one or more of the commitments it is in tension with, is the one that we should abandon in order to resolve the tension identified. In other words, eliminativists must fix the skeptical spotlight.

Is this a reasonable burden to hold eliminativists accountable to? Haji, of course, might argue that he has already met it by defending Motivation/Ability and Kant’s Law at length independently.^{Footnote 26} Surely we do not expect eliminativists to prove that the other principles that give rise to the inconsistency motivating a skeptical position are true. Such a burden would indeed be far too heavy. It is thus important to emphasize that this is not what fixing the skeptical spotlight requires. Instead, what a more charitable interpretation of fixing the spotlight requires is a comparative defense of these commitments, one that is Moorean in spirit. Fixing the skeptical spotlight requires only that skeptics provide some further motivation for resolving the inconsistency in the way that they endorse. The burden for a skeptic on the path to eliminativism is to show that, once we have clearly identified the fact that one of these commitments must go, our reasons weigh most heavily in favor of abandoning the problematic feature of the thing to be eliminated rather than these other commitments. Regarding Haji’s skeptical position, we are in need of further argument that, in light of the tension between Motivation/Ability, Kant’s Law, and Blameworthiness/Impermissibility, it is Blameworthiness/Impermissibility that should go, regardless of the independent plausibility of each of these three principles considered on their own.

Here eliminativists might pursue a further line of objection. They might grant that fixing the skeptical spotlight is theoretically desirable, but deny that meeting this burden is in fact necessary.^{Footnote 27} Given that there is no initial reason to think that eliminativists’ account of the plausibility of these commitments will fare any worse than their competitors’, why can’t eliminativists simply help themselves to a presumed fixing of the spotlight and proceed conditionally on the path to eliminativism from there?

It seems clear that whether or not eliminativists can get by with a promissory note in regard to fixing the skeptical spotlight will depend heavily on what the ultimate practical implications of the particular eliminativist view on offer turn out to be. The higher the “stakes” are for an eliminativist view of moral responsibility (in other words, the more drastic the view is regarding the ways it rationally entails we must revise our responsibility-related attitudes and practices, and the potential costs that go along with such sweeping changes), the heavier the burden on eliminativists to proceed on firmer ground and fix the spotlight before proceeding. If the brand of eliminativism on offer turns out to be particularly “low stakes” in terms of its practical implications, then perhaps a promissory note in regard to fixing the skeptical spotlight will be sufficient. However, I take it to be a characteristic feature of any robust eliminativist view that it will have at least some significant practical implications. If not, then it is not at all clear what motivates tackling the methodological hurdles of framing the view as eliminativist in the first place. Varieties of eliminativism are, at least in paradigm cases, not ho hum affairs. And so while there may be room for some uncharacteristically low stakes versions of eliminativism to initially sidestep the skeptical spotlight with a promissory note, there is good reason to think that it is a genuine burden that most varieties of eliminativism must attend to. At the very least, those varieties of eliminativism that do attempt to circumvent fixing the skeptical spotlight are saddled with the related burden of explaining why they are licensed to do so. They ought to make explicit the low practical stakes of the view on offer, and explain why eliminativism in this case is still a worthwhile pursuit.

I take fixing the skeptical spotlight to be a first step on the path from skepticism to full blown eliminativism. But while this step may be necessary, it is not sufficient for meeting the eliminativist burden. Even if the skeptic can show that the purportedly problematic feature of our concept of responsibility (for example, Blameworthiness/Impermissibility) is the commitment that should go, they must further motivate the claim that our responsibility-related attitudes and practices cannot or should not be retained without this feature. I turn now to two possible strategies for meeting this second feature of the eliminativist burden.

IV. Meeting the Motivational Challenge: Assessing Gains and Losses

Let’s assume that the skeptic is able to fix the skeptical spotlight. Some feature of our concept of moral responsibility (perhaps Blameworthiness/Impermissibility) is incoherent, implausible, or gives rise to an inconsistency with other independently plausible background commitments, and the skeptic is able to persuasively argue that the relevant feature of our concept of moral responsibility is the most vulnerable. In order to move from skepticism to eliminativism there is a further challenge: eliminativists must meet what I will hereafter call the motivational challenge. They must motivate eliminativism over preservationist prescriptive alternatives like revisionism. They must argue that in light of the error identified by the skeptical argument they endorse we cannot or should not retain our responsibility-related attitudes, and practices. In what follows I take each of these possible paths to elimination in turn. I begin with familiar arguments that we should not retain these attitudes and practices, arguments that appeal to the gains and losses of doing so. In the next section I turn to an alternative path to elimination, direct arguments that we cannot retain our responsibility-related attitudes and practices in light of acceptance of some brand of descriptive skepticism.

Perhaps most notably, Derk Pereboom has argued extensively that we would not suffer significant losses if we were to give up those responsibility-related attitudes and practices that presuppose basic desert of praise and blame, and in fact we might be better off if we did so. Pereboom argues that many of the relevant attitudes that we value (forgiveness, gratitude, and feelings of achievement and self-worth among them) can either be rationally retained, even if no one is ever deserving of praise and blame in the basic sense, or they can be replaced with satisfactory analogues.^{Footnote 28} Furthermore, those attitudes (foremost among them moral anger and resentment) and practices (foremost among them retributivist forms of punishment) that we would be rationally forced to give up are, on the whole, features of human life that do more harm than good. And the same can be said for our responsibility-related practices. Those practices that we value (for example, moral education) can be rationally retained even if no one is ever deserving of praise and blame in the basic sense, and those that we would be rationally forced to give up (for example, retributive punishment) do more harm than good.

If these arguments are correct, then they can be used to construct a General Eliminativist Argument (GEA) that we should not retain our responsibility-related attitudes and practices:

1. 1. If no one is ever deserving of praise and blame in the basic sense, and the attitudes and practices that depend on basic desert of praise and blame do not make us better off, then we should not retain them.

2. 2. No one is ever genuinely deserving of praise and blame in the basic sense.

3. 3. The attitudes and practices that do depend on basic desert of praise and blame do not make us better off.

4. 4. So, we should not retain these attitudes and practices.

Though Pereboom himself does not make it explicit, something much like GEA seems to be what motivates the prescriptive element of his hard incompatibilism.^{Footnote 29} Premise (2) is motivated by a series of familiar arguments against a variety of success-theories^{Footnote 30} about moral responsibility, including the luck objection to libertarian incompatibilism, Pereboom’s own Four Case Manipulation Argument (FCMA) against compatibilism, and his wild coincidences objection to agent-causal libertarianism.^{Footnote 31} If these arguments are successful, and these positions represent an exhaustive picture of the only plausible success theories of moral responsibility, then no one is ever deserving of moral praise and blame in the basic sense. Premise (3) is motivated by Pereboom’s own assessment of the gains and losses of abandoning the responsibility-related attitudes and practices that rationally presuppose basic desert of praise and blame, and I will not rehearse these arguments here.^{Footnote 32} Finally, premise (1) appears to be fairly obvious. If in fact no one is ever deserving of praise and blame in the basic sense, and furthermore the attitudes and practices that depend on basic desert of praise and blame have no consequentialist, contractualist, or pragmatic justification either, then it is difficult to see why we would have any reason to retain them. Especially if, as Pereboom suggests, doing so might actually make us worse off in that at least some of these undeserved attitudes (for example resentment) and unfair practices (for example retributive punishment) often result in harm that would be unfair or undeserved.

On the face of things, GEA looks like a plausible argument. However, appeal to an argument of this form will not allow eliminativists to meet the motivational challenge. Upon closer inspection, arguments like GEA offered in service of meeting the motivation challenge are not likely to be persuasive. In order to meet the challenge eliminativists must show that, in light of a plausible version of skepticism, our responsibility-related attitudes and practices cannot or should not be retained. GEA represents one possible argument that we should not retain them via appeal to the claim that doing so would not make us better off, and in fact might make us worse off. But this claim will only be plausible to those who have already rejected eliminativism’s preservationist prescriptive competitors.

To see why, consider again revisionism as a possible preservationist alternative to eliminativism. When faced with GEA, revisionists can and should point out that the plausibility of (3) depends on antecedent assumptions regarding what, if anything, might play the role of sustaining and justifying our responsibility-related attitudes and practices. In Pereboom’s case, it is because these attitudes and practices give rise to widespread unfairness and undeserved harm that we are no worse off (and perhaps even better off) without them. But his arguments in favor of this claim depend crucially on the following assumptions^{Footnote 33}:

• (AP1): Being the ultimate source of one’s action is the only kind of responsible agency capable of sustaining and justifying basic desert of praise and blame.

• (AP2): Basic desert is the only kind of fittingness relation capable of sustaining and justifying differential treatment of one another in regard to moral responsibility that is fair.

Is a preservationist of any stripe, or a relatively neutral party without prior eliminativist leanings, likely to accept these assumptions? It seems clear that the preservationist will not. Returning to revisionism as an instructive example, a rejection of either (AP1) or (AP2) is precisely what motivates the claim that some form of revision is possible despite the problematic feature of our concept of moral responsibility identified by revisionists’ initial descriptive project. According to revisionists, even if our standing concept is fatally flawed we might still be able to sustain and justify our responsibility-related attitudes and practices with a significantly revised or replacement concept. For connotational revisionists it is possible to revise our concept in such a way that still leaves us with a view of responsible agency capable of sustaining and justifying basic desert of praise and blame. And so the connotational revisionist will reject (AP1) out of hand, and with it premise (3) of GEA above. For the denotational revisionist, the standing concept is in fact too flawed to meet the task. But the denotational revisionist will still claim either that the kind of agency picked out by the flawed concept is not the only one capable of doing this work (and thus reject (AP1)), or that basic desert is not the only relation of fit or appropriateness between our attributions of moral responsibility and a form of agency sufficient to ground fair differential treatment of one another (and thus reject (AP2)).^{Footnote 34} If this is correct, then an argument like GEA will have no motivational force for revisionists, or for those who take revisionism as a live option. Premise (3) will have little to no plausibility for those who already reject (AP1) or (AP2).

But this is precisely the task at hand for eliminativists hoping to meet the motivational challenge. They need not convince those who already agree that our concept of moral responsibility is in error, and that our responsibility-related attitudes cannot be sustained and justified any other way. What eliminativists need in order to meet the motivational challenge is an argument that those who reject the latter claim might find plausible, and GEA cannot meet this task. This is a serious problem for eliminativism, and one that is likely to generalize for any attempt to meet the motivational challenge by way of an argument that appeals only to the gains and losses of abandoning our responsibility-related attitudes and practices. What these arguments fail to show, and it seems must show, is that these attitudes and practices do in fact depend on the problematic concept of moral responsibility at issue in the relevant way.

This discussion highlights the fact that eliminativists’ prospects for meeting the motivational challenge via arguments like GEA alone look grim. However, perhaps a better strategy is available. In particular, eliminativists might offer direct arguments in support of something like (AP1) or (AP2).

Before turning to these direct arguments, it is worth noting at least two further considerations that might bear on the way relatively neutral parties (those without prior eliminativist or revisionist leanings) might assess the plausibility of (AP1) and (AP2), and thus the adequacy of eliminativist attempts to meet the motivational challenge via arguments like GEA. The first involves the nature of moral responsibility facts, and the second a return to questions about the practical implications of the variety of eliminativism on offer. First, one might argue that whether or not revisionists can and should reject (AP1) and (AP2) in the way I discuss depends on a broadly realist assumption about the nature of moral responsibility facts. What is really going on at the heart of this discussion is that preservationists see rejecting (AP1) and (AP2) as live options only because they assume that there could be some broadly realist facts about agency and fittingness capable of grounding basic desert of praise and blame and the overall fairness of our responsibility-related attitudes and practices—facts that are distinct from those that eliminativists claim should do this grounding work, given our normative framework as it currently stands. But, eliminativists might argue further that there are no such facts, reject this broadly realist assumption, and argue that there is reason to think the facts that ground (AP1) and (AP2) cannot be disconnected from our normative framework in this way. Perhaps instead the truth of these claims is more heavily grounded in the framework itself. Here it is well beyond the scope of my current purposes to adjudicate this potential disagreement. I wish only to note it as a potential place eliminativists might plausibly push back on the claim that arguments like GEA are incapable of meeting the motivational challenge, and as a larger issue that might bear on the plausibility of (AP1) and (AP2) for relatively neutral parties.

Second, the adequacy of GEA for meeting the motivational challenge might also depend in important ways on the practical implications of the variety of eliminativism on offer. In particular, the above discussion proceeds under the assumption that Pereboom’s hard incompatibilism is sufficiently high stakes—it recommends at least some significant revision to our responsibility related attitudes and practices, at least those that pertain to blame, moral anger, and resentment. But Pereboom himself has recently defended a largely Scanlonian view of blame,^{Footnote 35} one that, if correct, might still be preserved even by hard incompatibilists. In light of this one might argue that the practical implications of Pereboom’s particular brand of eliminativism are relatively low stakes, and as such there are alternative routes to meeting the motivational challenge that appeal to gains and losses, but need not have the form of GEA. On an especially low stakes reading of Pereboom, for example, perhaps all that is needed is appeal to the fact that giving up the attitudes and practices that depend on basic desert of praise and blame itself costs us very little. And so why not err on the side of caution with low stakes eliminativists and give them up?

Considerations of space here also constrain my ability to fully explore the way practical implications might influence the adequacy of eliminativist attempts to meet the motivational challenge. However, it is worth echoing my remarks on the practical implications of varieties of eliminativism at the end of the previous section. It seems low stakes varieties of eliminativism are, to an important extent, nonstandard varieties of eliminativism. Here I am concerned primarily with the potential for more robust or paradigmatic varieties of eliminativism to meet the motivational challenge. How we should classify Pereboom’s own view, and whether there are alternative strategies for meeting the motivational challenge available to low stakes eliminativists are both important questions that merit separate treatment elsewhere.

V. Meeting the Motivational Challenge: Direct Arguments

In order to successfully move from descriptive skepticism to prescriptive eliminativism, eliminativists must meet the motivational challenge. They must motivate eliminativism over preservationist competitors like revisionism in one of two ways: (1) by arguing that we should not retain our responsibility-related attitudes and practices, or (2) arguing that we cannot retain our responsibility-related attitudes and practices. If the above discussion is correct, then a persuasive eliminativist argument that we should not retain our responsibility-related attitudes and practices itself depends on further arguments for the claim that we cannot do so. It seems, then, that successfully motivating eliminativism about moral responsibility requires an argument for thinking that the work of our standing concept of moral responsibility—to sustain and justify these attitudes and practices such that they render our differential treatment of one another regarding praise and, most importantly, blame fair—is work that cannot be done by a significantly revised concept of responsibility (free from the purported error at issue) or by some replacement concept that is distinct but capable of playing the same theoretical role.^{Footnote 36}

How might such an argument proceed? Galen Strawson’s impossibilism suggests one possibility. Strawson’s brand of skepticism is motivated by his Basic Argument, which can be stated as follows:

1. 1. Nothing can be causa sui.

2. 2. In order to be truly morally responsible for one’s actions, one would have to be causa sui, at least in crucial mental respects.

3. 3. Therefore nothing can be truly morally responsible.^{Footnote 37}

Premise (2) is a claim about a necessary or constitutive feature of our concept of moral responsibility. According to Strawson, true^{Footnote 38} moral responsibility requires being a self-caused cause, at least in crucial mental respects. However, acting as a self-caused cause requires control over these crucial mental aspects “all the way back,” a notion that has notoriously faced charges of incoherence or logical impossibility. If it is, and premise (2) is true, then it follows that being truly morally responsible is logically impossible. No one can ever be truly responsible for what they do.

Strawson takes the kind of moral responsibility at issue, true moral responsibility, to be just the one presupposed by any responsibility-related attitudes and practices intended to deliver differential treatment of one another that is fair. As he himself puts the point, it is the only kind of responsibility that would allow us to make sense of our notions of heaven and hell, the religious presupposition that it could be just to punish individuals for what they do with eternal torment and reward them with eternal salvation.^{Footnote 39} Given this understanding of the kind of responsibility at issue, if (2) is correct then we have good reason to think that any responsibility-related attitudes and practices intended to deliver fair differential treatment cannot be sustained and justified. A necessary or constitutive feature of our concept of moral responsibility—and furthermore any revised or replacement version capable of doing the required work of the concept—is logically impossible to instantiate, and so the attitudes and practices that presuppose it cannot be fair.

Here it will be helpful to note that I do not take this last line of argument to be Strawson’s own. Strawson, after all, has himself suggested that despite the skeptical conclusion of impossibilism it may not be in our best interest prescriptively to abandon our responsibility-related attitudes and practices.^{Footnote 40} Rather, here I wish to point out only that something like premise (2) is precisely what prescriptive eliminativists need in order to meet the motivational challenge directly.

This discussion also serves to highlight how hard incompatibilists might offer direct support for (AP1) and (AP2) above. The problematic feature of basic desert-entailing moral responsibility that Pereboom identifies is ultimate sourcehood. It is this feature that Pereboom claims compatibilists, event-causal libertarians, and agent-causal libertarians cannot plausibly deliver. But what hard incompatibilists need in order to move from this skeptical claim to full blown eliminativism is a direct argument for the claim that ultimate sourcehood is in fact necessary to or constitutive of a system of genuinely deserved, fair, fitting, or appropriate responsibility-related attitudes and practices. This is precisely where hard incompatibilists and revisionists (and possible preservationists of other stripes) reach an impasse. Revisionists, though, offer arguments for why we should reject the claim that ultimate sourcehood (and perhaps other sorts of libertarian agency) does in fact play this necessary or constitutive role. And so it seems that if eliminativists hope to motivate the prescriptive element of their view they are in need of more direct supporting arguments of their own. Without such arguments, it is not at all clear how the motivational challenge might be met in any satisfying way.

Here I anticipate an objection similar in spirit to the one already discussed in regard to fixing the skeptical spotlight: Isn’t the argumentative burden that I have here claimed eliminativists are saddled with unreasonably demanding? How might eliminativists meet this burden, and successfully argue that a particular feature of our concept of moral responsibility (perhaps ultimate sourcehood, perhaps being causa sui) is a necessary or constitutive feature of responsibility-related attitudes and practices that are deserved, fair, fitting, or appropriate? If nothing else, the list of notions appealed to here (desert, fairness, fit, and appropriateness) suggests that we are not even clear on what it is eliminativists must deliver. Furthermore, at least some of these notions appear to be basic in such a way that they may not be analyzable in any further terms.

First, while I grant that the further concepts appealed to here (desert, fairness, fit, and appropriateness) are relatively basic, I deny that they are brute and admit of no further, fruitful analysis. Michael McKenna, for example, has made a similar point about basic desert of praise and blame,^{Footnote 41} and I have argued elsewhere for just such an analysis. According to this analysis, we can further explicate basic desert of blame in terms of fittingness, where an agent deserves blame in the basic sense only in virtue of the right kind of reasons, and such reasons are in turn restricted to features of the agent and her action.^{Footnote 42} Given that such analyses do not appear to be obvious nonstarters, charges that the motivational challenge is too demanding grounded in appeals to the basicness of these notions at best requires further argument.

Second, it is not the case that we simply have no idea how an argument in favor of (AP1) or (AP2) might go. There are at least two options available to eliminativists. First, they might adopt an empirical approach and appeal to recent work in experimental philosophy to support an argument along the following lines: there is good empirical evidence that ultimate sourcehood or being causa sui are constitutive features of our folk concept of moral responsibility, and that the fairness of our practices depends on this feature of the folk concept.^{Footnote 43} The latter will, of course, take us somewhat astray from the empirical realm, but I see no reason to deny that an argument can be made that, given how these practices arose and came to be systematized in the first place it is reasonable to think that they cannot deliver their intended role—fair differential treatment of one another—without a particular constitutive feature of the folk concept. Alternatively, eliminativists might explicitly offer support for (AP1) or (AP2) via the more familiar route of conceptual analysis. In lieu of a reason to believe that these notions are in fact brute (as I have denied above), this kind of strategy might also prove fruitful.

Third, my characterization of the motivational challenge provides no argumentative advantages to preservationists. If the motivational challenge is too demanding, then it is equally so for preservationists as they face precisely the same burden in motivating preservation over elimination.^{Footnote 44} And in light of the fact that preservationists have explicitly attempted to meet this challenge (via arguments rehearsed above that we can and should retain the relevant attitudes and practices despite the problematic features of the concept identified), claims that it is too demanding would be deeply puzzling. In fact, one common charge one often hears levied against revisionists is that they do not adequately motivate the plausibility of genuine revision over elimination. Given that revisionists like Vargas have devoted a great deal of energy to taking on the motivational challenge directly, it seems eliminativists can only support this kind of demandingness objection if they are happy to absolve their opponents of the very same burden.

VI. Taking Stock

So where does this leave eliminativism? In light of the arguments I’ve provided here, I think that eliminativists do well to meet the burdens identified head on and offer explicit arguments for how we can fix the skeptical spotlight and meet the motivational challenge. While I have admittedly done little to help them on this front here, it is my hope that in identifying these burdens eliminativists will at least be better equipped to do so.

In a purely diagnostic spirit, I take part of what makes assessing prescriptive views like eliminativism and revisionism based on a form of skepticism so difficult is that little work has been done to articulate precisely what a successful defense might look like. On the preservationist side of things, Vargas’s work has been a notable exception. But on the eliminativist side of things, it is surprising that so little attention has been paid to the motivational challenge in particular.

Finally, in a purely speculative spirit, I suspect that much of the disagreement between preservationists and eliminativists has to do with methodological disagreements that run deep. On the preservationist side of things, eliminativists seem to rightly identify (as many of P. F. Strawson’s critics did) something unsatisfying about preservationist appeals to the value of and priority given to our responsibility-related attitudes and practices as they stand.^{Footnote 45} On the eliminativist side of things, preservationists seem to rightly identify something unsatisfying about eliminativist appeals to some problematic feature of our concept of moral responsibility being constitutive or necessary. What we are in need of, I think, is a better way to adjudicate this disagreement, and I take a necessary step in doing so to be a clearer characterization of what burdens are shared by those on each side of the debate. This in turn will require more explicit attention to arguments that might be offered in favor of privileging features of our concept more heavily than features of the practice, and vice versa. Preservationists have, thus far, attempted to meet this challenge head on. Regardless of what one thinks of the success of these attempts, I suggest eliminativists might benefit from more explicit attempts to meet their opponents on the same ground.