ArticlePDF Available

It’s (Almost) All About Desert: On the Source of Disagreements in Responsibility Studies

Article

It’s (Almost) All About Desert: On the Source of Disagreements in Responsibility Studies

Abstract

In this article I discuss David Shoemaker’s recently published piece “Responsibility: The State of the Question. Fault Lines in the Foundations.” While agreeing with Shoemaker on many points, I argue for a more unified diagnosis of the seemingly intractable debates that plague (what I call) “responsibility studies.” I claim that, of the five fault lines Shoemaker identifies, the most basic one is about the role that the notion of deserved harm should play in the theory of moral responsibility. I argue that the deep divide between those theorists who affirm and those who deny that moral responsibility is essentially about the justification of desert thus understood can be traced to the disagreement about whether the focus on the reactive attitudes by itself entails that moral responsibility has nothing to do with traditional questions about desert and free will. I then show that the seeming intractableness of the other four fault lines Shoemaker identifies is expectable and explicable in light of this more basic disagreement. After this diagnostic work, I conclude by suggesting a solution to the “morass” that has taken over responsibility studies: theorists working in the field should acknowledge that it has effectively bifurcated into two discrete subareas, which I suggest calling “retribution studies” and “interpersonal studies.”
It’s (almost) all about desert: on the source of disagreements in responsibility studies
Fernando Rudy-Hiller
This is a pre-peer reviewed version of the article published in The Southern Journal of
Philosophy
https://doi.org/10.1111/sjp.12405
Abstract
In this paper I discuss David Shoemaker’s recently published piece “Responsibility: the
state of the question. Fault lines in the foundations.” While agreeing with Shoemaker on
many points, I argue for a more unified diagnosis of the seemingly intractable debates that
plague (what I call) “responsibility studies”. I claim that, of the five fault lines Shoemaker
identifies, the most basic one is about the role that the notion of deserved harm should play
in the theory of moral responsibility. I argue that the deep divide between those theorists
who affirm, and those who deny, that moral responsibility is essentially about the
justification of desert thus understood can be traced to the disagreement about whether the
focus on the reactive attitudes by itself entails that moral responsibility has nothing to do
with traditional questions about desert and free will. I then show that the seeming
intractableness of the other four fault lines Shoemaker identifies is expectable and
explicable in light of this more basic disagreement. After this diagnostic work, I conclude
by suggesting a solution to the “morass” that has taken over responsibility studies: theorists
working in the field should acknowledge that it has effectively bifurcated into two discrete
subareas, which I suggest calling “retribution studies” and “interpersonal studies”.
1
1. Introduction
David Shoemaker’s “Responsibility: the state of the question. Fault lines in the
foundations” (2020), recently published in this journal,1 makes a valuable contribution in
attempting to make sense of the messy theoretical field of what we can call “responsibility
studies.”2 Shoemaker’s diagnosis is that the messiness emerges due to five different “fault
lines” about certain basic assumptions having to do both with the object of study of the
field and the appropriate theoretical framework that should be adopted to address it, fault
lines that can be traced to Strawson’s “Freedom and resentment” (1962/2003). These fault
lines concern: (1) the field’s data set (what kinds of responses count as holding someone
responsible); (2) the data set’s theoretical status (whether these responses provide evidence
of the nature of responsibility and responsible agency or rather constitute what
responsibility and responsible agency are); (3) the question whether desert is among the
appropriateness conditions of responsibility-characteristic responses; (4) the question
whether these responses are essentially (or even exclusively) backward-looking rather than
forward-looking; and (5) the question whether there is only one type of moral responsibility
or many.
While agreeing with Shoemaker on many points, my aim in this discussion will be
to introduce more unity in the diagnosis of the fault lines he identifies in the hope that, once
we clearly isolate the more basic source of disagreements, the resulting disputes will
become more tractable largely because it will become apparent that in many cases
disagreements result from the fact that different theorists are theorizing about different
things. In particular, I will argue that, of the five fault lines mentioned above, the third one
1 Subsequent references to this text will only give page numbers.
2 Like Shoemaker, Zimmerman (2015) also characterizes the field as a mess.
2
—concerning the role that desert should play in the theory of responsibility—is the most
basic and that the other four can be derived from it. If this is correct, the upshot is that
different theorists can focus their attention on a single question (or at least a cluster of
related questions) in order to find out whether they really are disagreeing or simply
theorizing about different things. My final suggestion will be that the field of responsibility
studies should (and most likely will) bifurcate into two discrete subareas, which we might
call “retribution studies” and “interpersonal studies”.
2. A core concept of moral responsibility
To begin with, it’s worthwhile to emphasize that there is a core concept of moral
responsibility that all theorists Shoemaker cite will accept and in virtue of which there is
such a thing as a field meriting the label “responsibility studies”. It’s this:
Moral responsibility. Moral responsibility is about the evaluations and reactions
that are appropriately triggered by how well or badly people’s conduct coheres with
certain norms and expectations, particularly (but not exclusively)3 those of morality.
This definition is sufficiently broad so as to encompass all the different approaches
to responsibility Shoemaker surveys in his article.4 For starters, “conduct” should be
understood very capaciously, since it must include things like intentional actions and
3 Other relevant norms and expectations include those of friendship, collegiality, marriage, etc., assuming that
they can’t be neatly subsumed under the norms and expectations of morality. This suggests that we will likely
face an explosion of the kinds of norms and expectations that are relevant for moral responsibility—which
may, at the limit, become coextensive with the norms and expectations that regulate all kinds of interpersonal
relations. This will be a natural result from the research agenda initiated by Strawson in “Freedom and
resentment.” I will return to this point below. Notice also that “morality” here isn’t restricted to
determinations of right and wrong, but include as well the notions of good, bad, virtue, and vice.
4 This definition doesn’t presume a “merited-consequences” conception of responsibility (Hieronymi 2019).
Reactions being appropriate isn’t the same as consequences being merited. See section 3 below.
3
omissions; unintentional actions and omissions; states of awareness (what one notices and
forgets); spontaneous emotional reactions; and the expression of character traits,5 all of
which have been thought to be fitting objects of responsibility assessments. Also, the terms
“evaluations” and “reactions” are so general that they can encompass things as different as
judgments of moral worth; judgments that one no longer stands in relations of mutual
regard with another person; aretaic appraisals; requests for justification and recognition of
fault; emotional responses; attempts at moral address; interpersonal dynamics that aren’t
sanctions such as withdrawals of trust and friendly attitudes; sanctions properly; and
punishment. In short, the whole panoply of things that constitutes the data set occupying
responsibility theorists. Thirdly, this definition makes room for investigating not only what
is involved in holding people responsible but also what is involved in people being
responsible, since almost everyone after Strawson accept that the evaluations and reactions
in question provide the key for establishing what the distinctive capacities of responsible
agents are—even if there is disagreement about whether responsible agency provides an
independent standard against which we can gauge the appropriateness of these evaluations
and reactions or not.
It should also be noted that the proposed definition strides over all five of
Shoemaker’s fault lines, since it’s silent on exactly which evaluations and reactions are at
stake; on whether these evaluations and reactions constitute what responsibility is or merely
provide evidence for the same; on what their appropriateness conditions are (including
whether they involve desert or not); on whether they are backward- or forward-looking; and
on whether there is a single type of responsibility or several.
5 Even if one thinks that responsibility-characteristic evaluations can concern appraisals of character traits
themselves (“aretaic appraisals”), one must concede that one appraises them by focusing on their outward
expression in conduct (how else could you know that someone is greedy or courageous except by witnessing
her perform greedy or courageous actions?).
4
The importance of isolating this core common concept of moral responsibility is
that in this way we defuse skepticism about the existence of a sufficiently unified enterprise
that merits the label “responsibility studies” at all, which isn’t an unreasonable conclusion
to draw after reading Shoemaker’s alphabet-long list of the diverse approaches to
responsibility (207). With this core concept at hand, let’s investigate how the five fault lines
Shoemaker identifies are generated.
3. Desert as the basic fault line
I will now argue that the fundamental disagreement among people working in responsibility
studies is about the place (if any) that the notion of deserved harm or, alternatively,
deserved suffering, ought to have in a theory of moral responsibility, which corresponds to
the third fault line Shoemaker identifies. As Shoemaker characterizes it, the disagreement
here is twofold (221-2): on the one hand, philosophers disagree on whether blame really
has a harmful aspect of the sort that requires moral justification; on the other, they disagree
about what moral justification can be offered for those instances of blame that are harmful
(assuming there are any). I characterize the dispute more succinctly in terms of the
fundamental divide between those who accept and those we deny the following claim:
Desert. Moral responsibility has something intrinsically to do with the justification
of the harm or suffering deserved by those who commit wrongdoing.
The thought here is simple enough: if you did something wrong or bad, now you
deserve (in some sense) to suffer (in some way) in response. This is the essence of
5
retributivism.6 It’s true that retributivists disagree on many points (Nelkin 2016), for
instance on whether the fact of wrongdoing makes it permissible or obligatory to inflict
harm or simply warrants a contrastive reason, i.e., if you need to harm anyone, harm the
blameworthy rather than the blameless (the “deontological question”). They also disagree
on whether it’s intrinsically good that the blameworthy suffer and on how to characterize
the goodness in question (“the axiological question”). In addition, retributivists disagree
massively on which kinds of harms or sufferings are at stake, going from the pain of guilt
(Rosen 2015; Carlsson 2017) all the way to eternal damnation in hell (Strawson 1994).
And, of course, as Shoemaker points out, they disagree on how the very notion of desert
should be understood (222). And yet, accepting or rejecting Desert is the fundamental
theoretical decision-point for those working in responsibility studies. If you think
something like Desert is right, this will color your choices at all the other theoretical
decision-points Shoemaker describes—and the same is true of you reject it. Going forward,
I will call those philosophers who accept Desert “retributivists” and, for reasons that will
become apparent in a moment, I will call those who reject it “Strawsonians.”
How was the divide between retributivists and Strawsonians generated? As with
almost everything else in this domain, it goes back to “Freedom and resentment.” As
Shoemaker nicely puts it, Strawson’s chief aim in that paper was to “free us from the quest
for freedom” (212), that is, to convince philosophers that questions about moral
responsibility can and should be divorced from questions about free will and determinism.
To accomplish this, Strawson focused on the kind of reactions that are the staple mark of
engaged interpersonal relationships—the reactive attitudes. The fundamental divide on the
post-Strawsonian era is, I will argue now, a divide over whether this focus on the reactive
6 Fischer (2007: 81-2) claims that the ordinary concept of moral responsibility is committed to retributivism.
6
attitudes by itself entails that moral responsibility has nothing to do with desert as
traditionally understood and, given the central place that desert plays in the free will debate,
that moral responsibility and free will are two distinct and possible unrelated topics.
On the one hand, many people—those I have called retributivists—tend to conceive
the reactive attitudes as an interpersonal analogue of institutionalized sanctions and
punishment and therefore assume we can continue asking the same questions that
traditionally have been asked in the free will debate by appealing to them (e.g., Wolf 1990:
20-1; Wallace 1994; Fischer and Ravizza 1998: 5-8; Watson 2004: 278; Pereboom 2014: 2;
Nelkin 2016). On the other hand, many people—those I have called Strawsonians—think
that, once we focus on the reactive attitudes, we thereby have changed the topic. The crucial
point for this second group of theorists is that the reactive attitudes are categorically
different from sanctions and punishment and therefore their appropriateness conditions—
including the agential capacities required for being aptly targeted by them—are
categorically different as well (Hieronymi 2004, 2019; Scanlon 2013a, 2015; Shoemaker
2013; Talbert 2016; Smith 2019).7
Let me elaborate on the reasons why Strawsonians think this is the case. First, some
have noted that many of the paradigmatic reactive attitudes, including resentment and
indignation, are non-voluntary in the sense that one can’t decide to adopt them at will,
much in the same way in which one can’t believe at will. Rather, one simply finds oneself
resenting or being indignant at someone after judging that they have shown disregard or ill
will to oneself or another, just as one finds oneself believing that p after judging the p is the
7 It’s worth noting that, according to the characterization I have offered, Strawson himself was not a
Strawsonian, since he explicitly held the view that the reactive attitudes are not categorically different from
punishment. He wrote: “So the preparedness to acquiesce in that infliction of suffering on the offender which
is an essential part of punishment is all of a piece with this whole range of attitudes of which I have been
speaking” (Strawson 1962/2003: 90, italics added). As it sometimes occurs, I think this is a case of a
philosopher not fully understanding the consequences of his own view.
7
case. Thus, the reactive attitudes are fundamentally unlike sanctions and punishment, which
are voluntary activities and, as such, are subject to the same standards of justification that
apply to any action that intentionally burdens others (Hieronymi 2004, 2019).
Second, others have argued that, even if it’s true that the reactive attitudes can be
burdensome to their targets, they aren’t a species of sanctions or punishment because they
aren’t adopted in order to make their targets suffer “but for reasons having to do with our
own concern with our relationships with them” (Scanlon 2015: 93). In other words, the
primary function of the reactive attitudes is “interpersonal calibration”, that is, they are
grounded on our concern with having attitudes toward people that are suitably attuned to
the treatment they dispense us and others we vicariously identify with.8
Third, even if, as some people have claimed, blaming reactive attitudes often
incorporate thoughts and desires that their targets suffer (McGeer 2013; Rosen 2015), this
entails neither that they are sanctions nor that they necessarily involve a disposition to
engage in sanctioning behavior (Wallace 1994; Watson 2004). For instance, even if upon
learning that a sadistic crime has been committed against a particularly vulnerable person (a
child, say) I find myself experiencing not only intense indignation but also a strong desire
that the perpetrator suffer, this doesn’t entail that my reactions are sanctions nor does it
entail that I would, even if the circumstances were propitious, act on that desire or even
think that doing so would be appropriate. Therefore, the preparedness to acquiesce in the
infliction of suffering (to use Strawson’s phrase) doesn’t incur the justificatory burdens of
actually inflicting suffering. Moreover, even if some people—plausibly, all of us at some
point or another—do employ blaming responses with the aim of causing pain, this doesn’t
show that we need a moral justification for the reactive attitudes of the sort that is needed
8 This applies to interpersonal dynamics that give expression to non-voluntary reactive attitudes, such as
giving someone the cold shoulder or refusing to help an uncooperative person with their projects.
8
for sanctions and punishment. As Shoemaker aptly puts it, “People can be fairly mean in
their retribution. But that is not what angry blame itself wants; it is instead what angry
people sometimes want” (224). That is, the presence of retributive impulses in people
doesn’t show that the reactive attitudes are by nature vehicles for the expression of those
impulses; therefore, they don’t incur the justificatory burdens of retributive practices.
Fourth and finally, given that accountability relations among adults are premised on
their status as equal members of the moral community, they possess mutual authority to
make demands on one another (Darwall 2006). At the same time, their equal status entails
that they lack authority over one another to inflict harm in response to disappointed
demands and expectations. So while I can certainly respond to your wrongdoing in ways
you have reason to dislike (like giving the cold shoulder to an untrusty friend), these
responses don’t have the status of sanctions because both formal and informal sanctions
(e.g., legal punishment administered by the state, a parent grounding their children)
presuppose a form of authority over another that is simply absent in moral relations among
equals (Shoemaker 2013).
I think this set of reasons constitutes a fairly weighty case in favor of the
Strawsonian position. However, we should be very careful in stating exactly what follows
from it. It does not follow that moral responsibility is categorically different from questions
about sanctions, punishment and desert, full stop. Rather, it only follows that, if moral
responsibility is understood in terms of the reactive attitudes, then moral responsibility is
categorically different from questions about sanctions, punishment and desert (where desert
is understood as whatever justifies harm or suffering in response to wrongdoing). And yet
this is a very substantive conclusion, since it serves both to largely explain the mess that
many people feel has taken over responsibility studies and to show how to clean it up. It
9
explains the mess because, in failing to recognize that this fundamental divide between
retributivists and Strawsonians actually entails a radical change of topic, it has become very
difficult (if not impossible) for philosophers to evaluate the mutually contradictory claims
that are made about the conditions for moral responsibility (Zimmerman 2015; see section
4 below). And it contributes to cleaning up the mess because it suggests the following
advice: if you are interested in the interpersonal reactions and dynamics that characterize
moral relations and, more broadly, engaged interpersonal relationships in general, then you
should stop worrying about issues of free will, control,9 and desert. If, on the other hand,
you are interested in free will, control and desert, i.e., on the agential condition that could
justify the infliction, or simply the presence, of harm or suffering in response to
wrongdoing, then you should not focus on the reactive attitudes as your preferred gloss on
what moral responsibility is. How, then, should you understand moral responsibility if you
are interested in these things? My suggestion: focus on either divine or legal punishment.10
In his final assessment of the dispute between those I have called retributivists and
Strawsonians, Shoemaker writes: “These are indeed radically different approaches, and it is
unclear how they may ultimately be brought together” (225). I agree with Shoemaker that
these are radically different approaches, but the crucial question is not whether they can be
brought together but rather whether defenders of each can be brought to see that they are
9 I admit that it’s more controversial to claim that once we adopt the Strawsonian position not only free will
but also control becomes dispensable. In work in progress I argue in detail that control understood in terms of
reasons-responsiveness isn’t a condition of moral responsibility when the latter is conceived along
Strawsonian lines.
10 Galen Strawson (1994) explicitly links moral responsibility to divine punishment, while Zimmerman
(2015: 58) explicitly links the kind of moral responsibility that requires control to legal punishment. See also
Watson’s (2004: 280) remark that moral accountability can be conceived as a “legal-like practice.” Brink and
Nelkin (2013) sketch a theory of moral responsibility based on the presumption that moral and criminal
responsibility share the same agential conditions and differ only on the norms and responses that characterize
each. Although they mistakenly assume that moral responsibility so conceived can be characterized in terms
of the reactive attitudes, they are on the right track in noticing the tight parallels between a retributivist
conception of moral responsibility and a retributivist conception of criminal responsibility.
10
actually theorizing about different things. In other words, these approaches are radically
different because they have different objects of study. Strawsonians are concerned with
explaining the relevance of the reactive attitudes for social beings like us for whom the
regard of others inescapably matters (Hieronymi 2020) and investigating those attitudes’
“normative felicity conditions” (Darwall 2006: 4-5, quoted by Shoemaker, p. 224), whereas
retributivists are interested in what kind of agents we must be for “truly” deserving harm or
suffering in response to wrongdoing (Pereboom 2014) or for deserving the “special
condemnatory force” of blame and its associated harms (Wolf 1990, 2011; Wallace 1994,
2011; McKenna 2019).
I am confident that substantial agreement can be reached on this point. That is, I
think that people can be led to agree that the reactive attitudes are just too different from
sanctions and punishment and therefore have categorically different appropriateness
conditions, and thus they can be led to agree that if moral responsibility is understood in
terms of the former then it doesn’t have anything to do with desert and free will/control.11
Before exploring in the final section the consequences this imagined agreement would have
for the future of responsibility studies, I will explain in the next one how the other four fault
lines Shoemaker identifies can be derived from this basic divide between retributivists and
Strawsonians.
11 Theorists like Clarke (2016), Carlsson (2017), and McKenna (2019), who think that desert is relevant for
the reactive attitudes not because the latter are a species of sanctions and punishment but because they
actually cause suffering to their targets (e.g., the suffering of not being able to interact with others
unimpededly or the suffering of guilt) regardless of the intentions of those who deploy them, will vigorously
disagree with this conclusion. In work in progress I defend the claim that the fact that the reactive attitudes
cause suffering isn’t enough for securing the relevance of free will/control for blame attributions. Following
Scanlon (2015), I argue there that people don’t have an unconditional right to expect good will from others;
rather, they can demand good will on the condition that they show it themselves. When they don’t, they lack
grounds for complaining about the withdrawal of good will the reactive attitudes entail (which includes the
suffering such withdrawal often causes to their targets), irrespectively of whether the wrongdoing in question
was freely willed or not.
11
4. The other fault lines
I will now argue that, once this fundamental divide between retributivists and Strawsonians
is in place, the other four major fault lines Shoemaker identifies are fully expectable and
explicable.12 I don’t mean the following to be an accurate reconstruction of how these other
fault lines were actually generated; rather, my point will be that what gives these other
disputes their air of intractableness is the foundational disagreement about the role that
desert should play in the theory of responsibility. That is, these disputes seem to embody
intractable disagreements because what drives them is a basic unclarity on the part of
theorists about what their object of study is supposed to be. By contrast, once we are clear
about this—once we distinguish clearly between retributivism and Strawsonianism—these
disputes will lose much of their apparent recalcitrancy (though not necessarily their
interest).
Composition of the data set
Shoemaker correctly argues that there is a major dispute “over the nature of the data set and
how exactly to determine its contents,” that is, a dispute about “[w]hat counts as the
relevant way of ‘holding responsible’” (217). This is exactly what we should expect in light
of the underlying disagreement between retributivists and Strawsonians.13 That is, given a
fundamental disagreement about whether responsibility theory is essentially about
explaining and justifying how people can come to deserve harsh treatment—if only under
the guise of the “opprobrium” of blame (Wallace 2011)—then we should expect that
apparently intractable disputes will emerge downstream about whether the reactions that
12 I accept for the sake of argument that these fault lines are indeed the most important extant disputes in the
field of responsibility studies.
13 Presumably Shoemaker will agree on this point, since one of his glosses of this fault line is in terms of the
question “what are we supposed to be theorizing about in the first place?” (217).
12
matter for theorizing about responsibility are those that necessarily involve (or presuppose
or license) a harmful aspect or not.14
If you fall in the retributivist side, you will give an affirmative answer to this
question. This is why people with retributivist leanings focus on a “harsh” construal of
blame (Wolf 2011; Wallace 2011; McGeer 2013; Rosen 2015) and, more specifically, on
resentment, because they assume that resentment is, by its nature, harmful and the
interpersonal analogue of sanctions and punishment—what gives blaming reactions is
characteristic force or depth (Wolf 1990: 20-1; Rosen 2015). This is also why they focus on
sanctioning activities (Wallace 1994; Watson 2004) and assume that there is a continuum
between reactive attitudes, sanctions, and punishment (Fischer and Ravizza 1993: 19;
Wallace 1994: 51; Vargas 2020: 413) and between moral and legal responsibility (Brink and
Nelkin 2013). By contrast, if you are a Strawsonian you will deny all of the foregoing and
will be progressively drawn to a position that welcomes into the theory pretty much all the
responses that are relevant for managing our interpersonal relationships irrespectively of
whether they can be categorized as species of praise and blame (Tognazzini 2015: 38-9;
Watson 2019: 224), excluding only those that indisputably are sanctions and punishment—
most clearly, negative consequences imposed by some person or body in a position of
authority in response to the violation of a norm (Scanlon 2013b; Hieronymi 2019).
The relevant point for present purposes is that in all likelihood the dispute about the
data set will become far less intractable once different theorists get clear about whether they
are interested in desert or not, since it will then be much clearer what responses are the ones
they should focus on.15 This doesn’t mean that all disputes in this regard will come to an
14 For a related diagnosis, see Vargas (2020: 413).
15 Don’t fall into the trap of thinking that the distinction between retributivists and Strawsonians is simply
another name for the dispute about the “faces” of responsibility. I elaborate on this point below when
discussing Shoemaker’s fifth fault line.
13
end; on the contrary, interesting internal debates will likely emerge within the retributivist
and Strawsonian camps. A case in point is Shoemaker’s contention that if we include hurt
feelings in the date set, Strawsonians may be forced to abandon their fixation with quality
of will, i.e., the view that all responsibility-relevant reactions are sensitive to the agent’s
moral regard or lack thereof (213-4). This is an interesting suggestion worth addressing, but
if we start from a basic agreement about what we are supposed to be theorizing about—e.g.,
the spontaneous responses that arise when interpersonal expectations are unmet—it should
be easier to ascertain whether the response in question (e.g., hurt feelings) should be
included in the data set or not.16
Response-dependent vs. response-independent theories
Another major dispute according to Shoemaker concerns the theoretical status of the data
set: do responsibility-relevant reactions play a constitutive or an evidential role regarding
moral responsibility? (218) In other words, is the property of being responsible for
something a response-dependent or a response-independent property? Once again, my
suggestion is that the apparent intractableness of this debate owes a lot to an upstream
disagreement about the role that desert should play in the theory of responsibility. If you are
a retributivist, you will think that a response-independent theory it’s the only viable option,
because it’s crucial for you to find out whether people really have the requisite agential
capacities (e.g., control, free will) for truly deserving harm in response to wrongdoing. A
response-dependent theory of desert makes no sense, since only those agents who are
16 Just for the record, I think Shoemaker’s (214) example of someone whose feelings are hurt because
someone she considered a friend threw away a gift of hers can be squarely understood as a case in which the
second person evinced lack of regard—understood, precisely, as a lack of concern for the other person’s
feelings. The first person could confront the second by asking, “how would you like if someone threw away a
gift of yours, eh?”
14
actually culpable truly deserve harm or suffering (on the retributivist’s view). By contrast,
if you are a Strawsonian the decision to adopt either methodology isn’t settled as a matter
of course; both may appear attractive for different reasons. A response-independent theory
is attractive because it allows us to preserve the view that we (not only as individuals, but
as community of appraisers) can make genuine mistakes in our attributions of
responsibility, for example mistakenly thinking that certain agents express morally relevant
ill will when they are actually incapable of such a thing (psychopaths, perhaps) (Fischer
and Ravizza 1993: 18). On the other hand, a response-dependent theory is attractive for
exactly the reason Shoemaker (220) gives: since (at least most of) responsibility-
characteristic responses are triggered by whatever us human beings regard as relevant for
evaluating each other’s quality of will, it’s plausible to think that what it is to be responsible
for something is simply to be targeted by one of these responses under appropriate
circumstances. These are genuine and interesting theoretical disputes within the
Strawsonian camp, but in the end the distinction between a response-dependent and a
response-independent approach won’t be as important as it is for retributivists. Once you
agree on the basics, i.e., that your object of study is the interpersonal reactions that
characterize engaged human relationships, it will be less consequential if different theorists
disagree on the exact theoretical status of these reactions.17 By contrast, if we haven’t
agreed on the basics—if we are discussing across the fault line separating retributivists and
Strawsonians—the disagreement between response-independent and response-dependent
views will naturally seem not only extremely important but intractable. The essential point
for current purposes is again that the apparent intractableness of this disagreement is really
17 I thus think, contra Shoemaker, that both response-dependent and response-independent Strawsonians will
have to engage in “down and dirty empirical work” (220) to figure out the nature of the responses they claim
to be investigating. My point, again, is that once you agree on the basics many of the theoretical disputes
Shoemaker describes will be less divisive than he thinks.
15
an artifact of the previous and more fundamental dispute about the role of desert in the
theory of responsibility.
Backward-looking vs. forward-looking views
I will suggest a very similar diagnosis of the fourth fault line Shoemaker identifies. If one is
a Strawsonian, the need to establish a sharp distinction between the backward-looking and
the forward-looking aspects of responsibility should be less urgent than if one is a
retributivist, since ordinary interpersonal responses are an inextricable mixture of both. So
while these responses are certainly triggered by what their targets did (or simply by what
they are) and, on the side of blaming responses, they surely have as one of their central
functions that of registering that an offense was suffered by oneself or another, they have all
sorts of other functions that don’t seem to be exclusively backward-looking in nature.
These include protesting wrongdoing (Smith 2013), bolstering the victim’s status as
someone who merits consideration (Talbert 2012), signaling commitment to norms
(Shoemaker and Vargas 2019), initiating a moral conversation (McKenna 2012),
scaffolding people’s moral sensitivities (Vargas 2013; McGeer 2019), recalibrating one’s
relationships with others (Scanlon 2008), etc. Thus, for Strawsonians the distinction
between the backward-looking and the forward-looking aspects of responsibility-
characteristic reactions will be blurry and mostly unimportant.18 On the other hand, for
retributivists it is crucial to emphasize that the deservingness conditions of these reactions
are exclusively backward-looking, since it makes no sense to say that someone deserves
harm or suffering in the “basic sense” on account of something that lies in the future
(Pereboom 2014), just as it makes no sense to say that an athlete deserves a medal because
18 I say “mostly” because Strawsonians will surely agree on Strawson’s point that the reactive attitudes aren’t
(exclusively) tools for social control.
16
it will bolster her athletic capabilities (Doris 2015).19 Once again, my suggestion is that the
apparent urgency of settling this debate is really an artifact of a failure to clearly distinguish
the very different objects of study with which retributivists and Strawsonians are
concerned.
Monism or pluralism about responsibility
We arrive finally at the one place in responsibility studies where the word “morass” is most
aptly applied: the dispute about the “faces” or “types” of responsibility (Shoemaker’s fifth
fault line). Once more, I contend that the seeming intractableness of this debate owes a
great deal to a confusion about the very different theoretical concerns of retributivists and
Strawsonians. If you are a retributivist, you will naturally think either that there is only one
legitimate type of responsibility (Levy 2005) or that, while there might be several, the truly
important one is the one linked to deserved harm or suffering (Levy 2011: 3; Pereboom
2014: 2; Nelkin 2016). Moreover, you will insist that the relevant type of responsibility is
the one that connects to, or is at issue in, the free will debate (Levy 2011; McKenna 2012;
Pereboom 2014; Nelkin 2016). Since Watson (2004) introduced the distinction between two
faces of responsibility, retributivists have taken it as a given that they are concerned with
the accountability face, since it has been assumed (again, since Watson) that accountability
is necessarily linked to deserved harm (see Nelkin 2016 for a very clear articulation of this
view). But here come the confusion: as I noted above, retributivists often describe
accountability in terms of the appropriateness of the reactive attitudes (Wallace 1994;
19 This isn’t to deny that interesting intramural disputes will remain in the retributivist camp about whether
the basic commitments of retributivism can be reconciled with some forward-looking elements, such as the
improvement of the wrongdoer through blaming and sanctioning activities (Vargas 2013; Pereboom 2014;
McGeer 2019). However, the presence of forward-looking functions such as this doesn’t impugn the basic
commitment of retributivism, viz., that the desert basis of responsibility-characteristic reactions only admits
backward-looking elements (Pereboom 2014). Thanks to an anonymous referee for pressing for clarification.
17
Fischer and Ravizza 1998: 5-8; Nelkin 2016), assuming that the latter embody their central
concern about deserved harm. This sows confusion because, obviously, Strawsonians insist
that they are also concerned with the appropriateness conditions of the reactive attitudes but
deny pretty much all that retributivists assert about what these conditions involve (Scanlon
2013a; Talbert 2016). Given this set up, retributivists often dismiss Strawsonians accounts
on the grounds that they fail to engage with the traditional problem of free will (Levy 2011:
209; Pereboom 2014: 98-9) and, by contrast, Strawsonians accuse retributivists of
misunderstanding the nature of these attitudes and, by extension, of responsibility itself
(Scanlon 2013b: 108; Hieronymi 2019).
Thus, I disagree with Shoemaker that the fault line here is generated by an
ambiguity in the notion of “quality of will” (228). To be clear, what I deny is that the
apparent intractableness of the debate about the types of responsibility is due to a confusion
about the different ways in which “quality of will” can be interpreted. I claim, rather, that
it’s due to a confusion about the different objects of study of retributivists and
Strawsonians. This doesn’t deny that there are different ways of construing what quality of
will involves nor does it deny that they generate important disagreements among
Strawsonians, but it does entail that the source of the seeming intractableness of the debate
about the faces of responsibility lies elsewhere.
If you are a Strawsonian you should be more open to plurality about responsibility,
especially if you accept that your object of study involves an enormous swath (maybe all!)
of the responses that characterize engaged human relationships. It would be astonishing if
all these responses shared the same set of appropriateness conditions, including the same
set of agential conditions. Interesting debates will remain about whether there is one
organizing set of conditions or not (Smith 2012, 2015; Shoemaker 2015), but this shouldn’t
18
make Strawsonians go to the barricades, as they will have to if they continue to be engaged
with retributivists in the debate about what the “true” or “most important” type of
responsibility is.20
Notice, finally, that the distinction between retributivists and Strawsonians isn’t the
same as the distinction between accountability and other types of responsibility. As many
authors have noted, the essence of accountability consists in being accountable to someone,
that is, it consists in being subject to other people’s legitimate demands and expectations
and to the reactions that are appropriate when these demands and expectations are
disappointed or exceeded (e.g., Watson 2004, 2019; Darwall 2006; Nelkin 2016). This
presupposes “the capacity for moral reciprocity or mutual recognition that is necessary for
intelligibly holding someone accountable to basic moral demands and expectations”
(Nelkin 2015: 363). Crucially, however, being accountable to someone in this sense isn’t
the same as being subject to another’s authority to administer sanctions and punishment
(Shoemaker 2013), nor does it entail anything about the goodness of the blameworthy
suffering some harm or about the obligatoriness, permissibility, or just preferableness that
they do (Scanlon 2013b). Therefore, Strawsonians have every right to a respectable notion
of accountability and thus the divide between retributivists and Strawsonians can’t be
equated to the dispute about accountability per se.21
5. The future of responsibility studies
To sum up, I have argued that the deep theoretical and methodological divisions that
characterize the field of responsibility studies can all be traced to a foundational
20 Though see Shoemaker’s important observation that recognizing the plurality of responsibility is essential
for respecting “marginal agents” (232).
21 It’s worth noting that some Strawsonians mischaracterize their own position as being about a form of
responsibility basically different from accountability (e.g., Arpaly and Schroeder 2014: Ch. 7).
19
disagreement about the role that desert (understood as deserved harm or suffering) should
play in theorizing about moral responsibility. Strawsonians, I have argued, have a very
strong case that once we focus on the ordinary reactions that characterize engaged human
relationships—the reactive attitudes—we thereby have changed the topic vis-à-vis the
traditional free will problem and its associated conception of moral responsibility. Failure
to clearly recognize this point by both retributivists and Strawsonians has led to seemingly
intractable disputes down the road. The solution I propose is thus simple: theorists working
on responsibility studies should acknowledge that the discipline has effectively bifurcated
into two discrete subareas, which I suggest calling “retribution studies” and “interpersonal
studies”.22
These are subareas of a single philosophical field—what I have called
“responsibility studies”—because they agree on the basic definition of moral responsibility
I gave in section 2. But they are discrete subareas nonetheless because they are concerned
with categorically different reactions to morally loaded conduct with categorically different
appropriateness (including agential) conditions. Retribution studies will retain the tight link
between moral responsibility and the questions that have informed the traditional free will
debate, and thus will continue focusing on understanding the (im)possibility of persons
coming to deserve good or bad things simply on account of what they have done. Theorists
working in this subarea will benefit from clearly acknowledging—as some philosophers
have already done23—that the conception of moral responsibility they are concerned with is
closely connected to the one at stake in discussions of criminal responsibility and in theistic
22 These are fields of study that fall squarely within ethics (or within the intersection of ethics and
metaphysics, in the case of retribution studies), so I am puzzled by Shoemaker’s suggestion that responsibility
theory may “become its own substantial and distinctive area of philosophical inquiry along the lines of other
well-established areas of philosophy like ethics, metaphysics, and epistemology” (233).
23 See footnote 10 above.
20
debates about free will, rather than to the one pioneered in “Freedom and resentment”. On
the other hand, interpersonal studies will be in charge of mapping out, understanding, and
laying out the appropriateness conditions of pretty much all responses that characterize
engaged human relationships (Tognazzini 2015; Watson 2019). A beneficial consequence of
this extremely wide focus is that it will naturally admit into our theorizing people who are
usually left aside in responsibility studies—those Shoemaker calls “marginal agents”—
since “normal adults” aren’t the only people who interact with and react to each other (231-
2).24
My diagnosis and proposed solution to the chaos that has taken over responsibility
studies isn’t intended as a magic wand that will end all disputes in one fell swoop. My hope
is, rather, that once there is widespread acknowledgement that the fundamental division
between retributivists and Strawsonians amounts to a cleavage between two separate
subareas of study, the remaining disagreements within these subareas will become more
focused and less intractable (though not less interesting), since there will be internal
agreement on the basic foundational issue, i.e., what the object of study is. Thus, rather than
something to be lamented, this partition of the field could produce a profitable division of
labor between the two subareas sketched above. So I think Shoemaker’s metaphor of “fault
lines” is ultimately misleading, because it suggests a bad sort of division (no one would
characterize a well-functioning division of labor as being “riven by fault lines” [205]). It’s
bad when something that should be unified is fractured; but it’s not bad (it’s actually good!)
when something that is basically diverse is recognized as such.
24 This will also open the door to a more sustained study of the sorts of reactions that are appropriate
regarding children’s conduct. See for instance Shapiro (1999) and Burroughs (2020).
21
It remains to be seen whether this proposed solution proves too optimistic. For my
own part, I refuse to accept that those of us working on responsibility studies are
condemned to confronting one dialectical stalemate after another.
References
Arpaly, N. and T. Schroeder (2014) In Praise of Desire. Oxford: OUP.
Brink, D. and D. Nelkin (2013) “Fairness and the Architecture of Responsibility.” Oxford
Studies in Agency and Responsibility, Volume 1. Oxford: OUP, pp. 284–314.
Burroughs, M. (2020) “Navigating the Penumbra: Children and Moral Responsibility.”
Southern Journal of Philosophy 58 (1): 77-101.
Carlsson, A. (2017) “Blameworthiness as Deserved Guilt.” The Journal of Ethics 21: 89-
115.
Clarke, R. (2016) “Moral Responsibility, Guilt, and Retributivism.” The Journal of Ethics
20: 121–137.
Darwall, S. (2006) The Second-Person Standpoint. Cambridge, MA: Harvard University
Press.
Doris, J. (2015) “Doing Without (Arguing about) Desert,” Philosophical Studies 172:
2625–2634.
Fischer, J. (2007) “Compatibilism.” In J. M. Fischer, R. Kane, D. Pereboom, and M.
Vargas (eds.) Four Views on Free Will, pp. 44–84. Oxford: Blackwell Publishers.
Fischer J. and M. Ravizza (1993) “Introduction.” In J. Fischer and M. Ravizza (eds.)
Perspectives on Moral Responsibility, pp. 1–41. Ithaca, NY: Cornell University
Press.
Fischer, J. and M. Ravizza (1998) Responsibility and Control: A Theory of Moral
22
Responsibility. Cambridge: CUP.
Hieronymi, P. (2004) “The Force and Fairness of Blame,” Philosophical Perspectives 18:
115-148.
Hieronymi, P. (2019) “I’ll Bet You Think This Blame is About You.” Oxford Studies in
Agency and Responsibility 5: 60–87.
Hieronymi, P. (2020) Freedom, Resentment, and the Metaphysics of Morals. Princeton:
Princeton University Press.
Levy, N. (2005) “The Good, the Bad, and the Blameworthy.” Journal of Ethics and Social
Philosophy 1 (2): 1-16.
Levy, N. (2011) Hard Luck: How Luck Undermines Free Will and Moral Responsibility.
Oxford: OUP.
McGeer, V. (2013) “Civilizing Blame.” In D. Coates and N. Tognazzini (eds.) Blame: Its
Nature and Norms, pp. 162-188, Oxford: OUP.
McGeer, V. (2019) “Scaffolding Agency: A Proleptic Account of the Reactive Attitudes.”
European Journal of Philosophy 27 (2): 301-323.
McKenna, M. (2012) Conversation and Responsibility. Oxford: OUP.
McKenna, M. (2019) “Basically Deserved Blame and its Value.” Journal of Ethics and
Social Philosophy 15 (3): 1-28.
Nelkin, D. (2015) “Psychopaths, Incorrigible Racists, and the Faces of Responsibility.”
Ethics 125 (2): 357-390.
Nelkin, D. (2016) “Accountability and Desert.” The Journal of Ethics 20: 173-189.
Pereboom, D. (2014) Free Will, Agency, and Meaning in Life. Oxford: OUP.
Rosen, G. (2015) “The Alethic Conception of Responsibility.” In R. Clarke, M.
23
McKenna, and A. Smith (eds.) The Nature of Moral Responsibility, pp. 65-88.
Oxford: OUP.
Scanlon, T. M. (2008) Moral Dimensions. Cambridge, MA: Harvard University Press.
Scanlon, T. M. (2013a) “Interpreting Blame.” In D. Coates and N. Tognazzini (eds.)
Blame: Its Nature and Norms, pp. 84-99. Oxford: OUP.
Scanlon, T. M. (2013b) “Giving Desert its Due.” Philosophical Explorations 16 (2): 101-
116.
Scanlon, T. M. (2015) “Forms and Conditions of Responsibility.” In R. Clarke, M.
McKenna, and A. Smith (eds.) The Nature of Moral Responsibility. Oxford: OUP,
pp. 89-111.
Shapiro, T. (1999) “What is a Child?” Ethics 109: 715–738
Shoemaker, D. (2013) “Blame and Punishment.” In D. Coates and N. Tognazzini (eds.)
Blame: Its Nature and Norms, pp. 100-118. Oxford: OUP
Shoemaker, D. (2015) Responsibility from the Margins. Oxford: OUP.
Shoemaker, D. (2020) “Responsibility: the State of the Question. Fault Lines in the
Foundations.” The Southern Journal of Philosophy 58 (2): 205-237.
Shoemaker, D. and M. Vargas (2019) “Moral Torch Fishing: A Signalling Theory of
Blame.” Noûs, early online https://doi.org/10.1111/nous.12316
Smith, A. (2012) “Attributability, Answerability, and Accountability: In Defense of a
Unified Account.” Ethics 122: 575–589.
Smith, A. (2013) “Moral Blame and Moral Protest.” In D. Coates and N. Tognazzini (eds.)
Blame: Its Nature and Norms, pp. 27-48. Oxford: OUP
Smith, A. (2015) “Responsibility as Answerability.” Inquiry 58: 99–126.
Smith, A. (2019) “Who’s Afraid of a Little Resentment?” Oxford Studies in Agency and
24
Responsibility 6: 85-111.
Strawson, P. (1962/2003). “Freedom and Resentment.” In G. Watson (ed.) Free Will.
Oxford: OUP, pp. 72-93.
Strawson, G. (1994) “The Impossibility of Moral Responsibility.Philosophical Studies
75: 5-24.
Talbert M. (2012) “Moral Competence, Moral Blame, and Protest.” Journal of Ethics 16
(1): 89-109.
Talbert, M. (2016) Moral Responsibility. Cambridge: Polity Press.
Tognazzini, N. (2015) “The Strains of Involvement.” In R. Clarke, M. McKenna, and A.
Smith (eds.) The Nature of Moral Responsibility, pp. 19-44. Oxford: OUP.
Vargas, M. (2013a) Building Better Beings. New York: OUP.
Vargas, M. (2020) “Negligence and Social Self-Governance.” In A. Mele (ed.) Surrounding
Self-Control, pp. 400-420. Oxford: OUP.
Wallace, J. (1994) Responsibility and the Moral Sentiments. Cambridge, MA: Harvard
University Press.
Wallace, J. (2011). “Dispassionate Opprobrium: On Blame and the Reactive Sentiments.”
In J. Wallace, R. Kumar, and S. Freeman (eds.) Reasons and Recognition, pp. 348-
372. Oxford: OUP.
Watson, G. (2004) “Two Faces of Responsibility.” In Agency and Answerability. Oxford:
OUP.
Watson, G. (2019) “Second Thoughts.” In Oxford Studies in Agency and Responsibility 5:
214-239.
Wolf, S. (1990) Freedom within Reason. Oxford: OUP.
Wolf, S. (2011) “Blame, Italian Style.” In R. J. Wallace, R. Kumar, and S. Freeman (eds.)
25
Reasons and Recognition: Essays on the Philosophy of T.M. Scanlon. Oxford: OUP,
pp. 332-347.
Zimmerman, M. (2015) “Varieties of Moral Responsibility.” In R. Clarke, M. McKenna,
and A. Smith (eds.) The Nature of Moral Responsibility. Oxford: OUP, pp. 45-64.
26
Article
Does one's causal responsibility increase the degree of one's moral responsibility? The proponents of resultant moral luck hold that it does. Until quite recently, the causation literature has almost exclusively been interested in the binary question of whether one factor is a cause of an outcome. Naturally, the debate over resultant moral luck also revolved around this binary question. However, we have seen an increased interest in the question of degrees of causation in recent years. And some philosophers have already explored various implications of a graded notion of causation on resultant moral luck. In this paper, I will do the same. But the implications that I will draw attention to are bad news for resultant moral luck. I will show that resultant moral luck entails some implausible results that leave resultant moral luck more indefensible than it was previously thought be. I will also show that what is typically taken to be the positive argument in favor of resultant moral luck fails. I will conclude that we should reject resultant moral luck.
Article
This paper examines the methodological claim made famous by P. F. Strawson: that we understand what features are required for responsible agency by exploring our attitudes and practices of holding responsible. What is the presumed metaphysical connection between holding responsible and being fit to be held responsible that makes this claim credible? I propose a non‐standard answer to this question, arguing for a view of responsible agency that is neither antirealist nor straightforwardly realist. It is instead “constructivist.” On the “Scaffolding View” I defend, reactive attitudes play an essential role in developing, supporting, and thereby maintaining the capacities that make for responsible agency. Although this view has relatively novel implications for a metaphysical understanding of capacities, its chief virtue, in contrast with more standard views, is in providing a plausible defense of why so‐called “responsible agents” genuinely deserve to be treated as such.
Chapter
Self-control can involve more than just impulse inhibition. For some notions of self-control, especially those concerned with moral responsibility, sensitivity to reasons is the idea central to self-control. For these accounts, it is not obvious how to capture the idea that people are responsible for negligence and other instances of apparently non-volitional culpability. One blames people for failing to take into account some important moral consideration in deciding what to do, for failing to remember some commitment, and for failing to recognize situationally relevant things. This chapter proposes an account of this broader notion of self-control, one that solves the problem of control in non-volitional culpability cases, and that retains the idea that people in such cases could have complied with the demands of morality.
Article
In this article, I discuss five major fault lines in the foundations of responsibility theory—a relatively new field—fault lines traceable to P. F. Strawson's groundbreaking “Freedom and Resentment.” They are about the proper methods and content of responsibility theory, and disputes over these foundational issues have led to a messy and wildly divergent set of theories and approaches in the field. My aim is simply to identify and sort out these fault lines so that we can at least agree on the source of our disagreements, and so that we may also perhaps realize the pressing need to address these fault lines first, in order to help resolve some of our downstream disputes and nurture responsibility theory into a more unified and theoretically sophisticated philosophical enterprise.
Article
Child moral agency is dismissed in many historical and contemporary accounts based on children's supposed lack or marginal possession of agency‐bearing capacities, including reason, deliberation, and judgment, amongst others. Given its prominence in the philosophical canon, I call this the traditional view of child agency. Recent advancements in moral developmental psychology challenge the traditional view, pointing toward the possession of relevant capacities and competencies for moral and responsible agency in early and middle childhood. I argue that both views—traditional and developmental—underdetermine our practices of holding children responsible in our common interactions. For one, we face significant epistemic barriers in accurately assessing children’s agential status qua possession of responsibility‐bearing capacities and competencies. Second, overreliance on assessments of individualistic capacities emphasizes an atomistic view of agency at the expense of relational views that are of particular relevance for children as uniquely developing persons. Our practices of holding children responsible and the values that guide these practices in the context of supportive relationships are central to both supporting current and drawing out future responsible agency in childhood and, importantly, provide us with a path to regard children as participants in our moral communities, as opposed to mere agents‐in‐waiting.