3.a Concept pluralism

One way to establish aim pluralism is to argue that more than one concept can be involved in the use of aim-talk:

My proposal is to defend concept pluralism by distinguishing between three concepts of blame’s aims: its intentional aim, functional aim, and normative aim. This distinction draws on a close analogy from the literature about the aim of belief. It is widely agreed that beliefs aim at truth, but philosophers have carefully distinguished between different concepts of aim that can be involved in such a claim.Footnote ⁵ I suggest that a similar taxonomy can be applied to the case of blame as well in a way that will turn out to illuminate our discussion of the communication argument.

Theorists have offered three competing kinds of interpretations of the platitude that belief aims at truth: the intentional interpretation, the functional interpretation, and the normative interpretation. The intentional interpretation understands the claim as stating that an agent only believes that p when she accepts p with the intention of obtaining truth. It follows that believers always have an intentional goal: to accept a proposition only when it is true. The functional interpretation understands the claim as stating that obtaining true beliefs is the function of the activity of regulating beliefs. This function may, but need not, rise to the level of the agent’s intention. Both the intentional and the functional interpretations are sometimes dubbed “teleological” interpretations.Footnote ⁶ They are attempts to offer naturalistic, non-normative explications of the claim that belief aims at truth. By contrast, there is the normative interpretation that understands the claim as stating an epistemic norm that is “essential and constitutive for belief” (Fassio Reference Fassio2011, 471): a belief is correct or appropriate if and only if its content is true. According to this interpretation, the expression of “aiming at truth” is best “interpreted as a metaphor” (Wedgwood Reference Wedgwood2002, 267) meant to capture this epistemic norm.Footnote ⁷ Disagreements between teleological and normative interpretations persist in the literature.

My proposal is that we can distinguish between different senses of “aim of blame” in a similar way, and they would provide different ways to explicate and reconstruct the communication argument. First, X is an intentional aim of blame if and only if an agent always, or at least typically, has an intention to X when she blames. The claim that blame aims at communicating a moral demand, then, implies that we blame with an intention to communicate such a demand to the wrongdoer. Second, X is a functional aim of blame if and only if a function of blame as a type of practice is to X. The function of blame may, but need not, rise to the level of intention of the blamer. Instead, the claim that blame aims at a distinctive kind of moral communication implies that it is a function of blame to facilitate this kind of communicative processes. Lastly, X is a normative aim of blame if and only if blame as a type of practice is governed by a norm like the following: blame is appropriate only if it is directed at agents who can X. This last interpretation takes the aim of blame as a metaphor. The claim that blame aims at communicating demands, then, implies that blame is appropriate only if it is directed at agents who can fulfill the required role in a distinctive kind of moral communication.

Given this taxonomy, we have good reasons to accept concept pluralism. More accurately, there can be an intentional version, a functional version, and a normative version of the claim that blame “aims” at a distinctive kind of moral communication. Concept pluralism does not directly pose any objection to the communication argument, but it does invite theorists to take special care to convey the meaning that they intend when contending that X is an aim of blame. Note that some instances of aim-talk by defenders of the communication argument—for instance, the claim that blame would be “pointless” (Shoemaker Reference Shoemaker2007, 85)—seem open to all three kinds of interpretations. One important exception is Macnamara (Reference Macnamara, Clarke, McKenna and Smith2015a, Reference Macnamara2015b), who clearly identifies the functional aim of blame as the concept relevant to the communication argument; but the potential ambiguity remains an issue for at least some defenders of the argument. Take Darwall (Reference Darwall2006) as an example. He frequently puts it that communicating the second-personal demand is what blame “seeks.” He writes, for example, that reactive attitudes “seek respect” and they “seek to engage the other second-personally, and they succeed when the other takes up the address, acknowledges its terms, and thereby respects the dignity of the addresser, both the demand she addresses and her standing to address it” (86). The most natural reading here seems to be intentional or functional. However, Darwall also frequently says that communication figures in the “normative felicity condition” of appropriate blame. As he puts it, a “normative felicity condition” of second-personal address is “a necessary condition of the second-personal reasons actually existing and being given through address” (56). This is more like saying that communicating the demand is the normative aim of blame. Note that the point here is not that Darwall’s account is inconsistent; I am inclined to think that his argument can be reconstructed in a way that avoids any potential linguistic ambiguity. Instead, the point here is that attending to concept pluralism can help clarify and reinterpret some existing theories and arguments that involve aim-talk—otherwise, we risk trading between claims about different concepts of blame’s aim.

One reason why we should be especially cautious of this risk, I think, is the following: the normative aim of blame should not be the relevant concept that figures in the communicative argument, so defenders of the argument should pay special care to dispel this interpretation. Otherwise, the argument will be the following:

There are a couple of reasons why this should not be how communicative theorists reconstruct their argument. The obvious problem is that the argument is not very informative and lacks dialectical value. Its crucial premise is that communication is a normative aim of blame, but that just means that there is such a norm that blame is only appropriate if it targets individuals who possess communicative capacities. Those who do not yet find the conclusion of the communication argument convincing are unlikely to simply accept this premise; the explanatory circle between the premise and the conclusion is too small to make the argument very informative.

In addition, the normative version of the communication argument would potentially undermine a, broadly speaking, Strawson-style naturalist project. This project is at least important for Watson (Reference Watson and Watson2004), who attempts to find a justification for the exemption condition of moral responsibility that is internal to our blaming practices involving reactive attitudes. Watson’s goal there is to develop and critically evaluate the best possible version of what he calls Strawson’s “expressive theory of responsibility” (226). The basic tenet of the expressive theory is that “it is not that we hold people responsible because they are responsible; rather, the idea (our idea) that we are responsible is to be understood by the practice … of expressing our concerns and demands about our treatment of another” (222). But if the starting point of the communication argument is a normative claim about when blame is appropriate, then we may worry that the account does not really offer an internal justification—the normative claim requires further justification, which may well turn out to be external to our practices of holding people responsible.Footnote ⁸

Again, concept pluralism does not directly pose any objection to the communication argument, but it invites conceptual clarity. More specifically, the upshot is that theorists defending the communication argument should (i) adopt the intentional version of the argument, (ii) adopt the functional version of the argument, or (iii) appeal to a sense of aim other than that of the intentional, functional, or normative aim. Option (iii) is a possibility worth more investigation in the future. For now, I will focus on the intentional and the functional versions of the argument. I will argue that even if we settle on which concept of aim is involved in the communication argument, there is still a further difficulty: blame plausibly has more than one intentional aim and more than one functional aim, which comes into tension with the communication argument.

3.b Content pluralism

Start from the intentional version of the argument:

This intentional formulation, I think, remains a reasonable reinterpretation of the communication argument to the extent that defenders of the argument take the communicative aim of blame as stemming from the nature of blame as a demand. After all, a literal demand qua speech act typically accompanies an intention to achieve some goal. More generally, participants of speech act activities typically intend to participate in such activities. To the extent that blame is thought of as a demand analogous to a form of speech act, it is reasonable to think that the aim of this activity is determined by the intention of the blamer. So one interpretation for the communication argument as defended by Watson, Darwall, and Shoemaker can be as follows: blamers intend to communicate a moral demand, and, therefore, only those who possess the (cognitive, behavioral, and/or affective) capacity to engage in this communicative process are morally responsible agents.

One difficulty with the intentional version of the argument concerns cases of private blame. It seems that many of those who have hidden resentment or indignation do not intend to communicate their demand; their intention is typically to not communicate it.Footnote ⁹ But there may be available moves to deal with this difficulty. One possible move can be found again in the literature about the aim of belief. In his defense of the intentional interpretation, Steglich-Petersen (Reference Steglich-Petersen2006) suggests that there can be “subintentional surrogates” (510) of the intention to obtain truth, if a mental state shares certain characteristics such that it can be conducive to the “hypothetical aim” (515) of someone intending to obtain truth. Similarly, a defender of the communication argument may turn to rely on the weaker assumption that blamers either (i) have an actual intention to communicate demands or (ii) have an attitude or behavior that shares certain characteristics conducive to the hypothetical aim of someone intending to communicate demands.

No matter whether such a move is ultimately defensible, my contention is that the intentional version of the argument faces another serious difficulty: the intention to communicate a moral demand does not appear to be the only intentional goal that would count as the intentional aim of the practice. That is, the argument is in tension with the following claim:

Remember that X is an intentional aim of blame if and only if an agent always, or at least typically, has an intention to X when she blames. Let’s assume that blamers do typically have an intention to communicate demands. Still, we blame with other kinds of intentions as well, sometimes having little to do with addressing a moral demand of reasonable regard. Depending on the context, some might blame with an intention to stand up for themselves and to promote one’s self-respect; some might blame with an intention to signal that they share some values widely accepted in their social community, and to maintain social cohesion; some might blame with an intention to vent their negative feelings.Footnote ¹⁰ At least some of these intentions seem to appear in a wide enough range of contexts such that we can say that blamers typically possess these intentions, and, therefore, they should count as blame’s intentional aims. Accepting content pluralism does not lead to any logical contradiction with the communication argument, but it does lead to a serious tension. This is because defenders of the communication argument tend to privilege the communicative aim over the other aims, but it is unclear how they justify doing so if communication is just one of the many intentional aims of blame.Footnote ¹¹ The challenge for defenders of the communication argument, then, is to show why the intention to engage in a distinctive sort of moral communication is special in a way that it leads to a constraint about morally responsible agency.

One may reply on the grounds that the communicative intention has a special conceptual connection to blame. It is conceptually contingent that blamers have intentions to promote their self-respect. However, to the extent that they engage in the practice of blame, there is a conceptual requirement that they also have a communicative intention. Here one may also invoke the distinction between distal and proximal intentions (e.g., Mele Reference Mele1992, Reference Mele2009; Pacherie Reference Pacherie2008). Roughly speaking, distal intentions are about what to do in the future, while proximal intentions are about what to do now (Mele Reference Mele2009, 10). It might be suggested that even when blamers have the distal intention of promoting their self-respect, their proximal intention in blame must be to engage in a distinctive kind of moral communication, such as communicating moral demands.

In reply, consider a case that I will call Idiosyncratic Intention: Matt was told by his therapist that practicing anger could relieve stress and be good for his mental health. The day after, Matt blames his friend for failing to fulfill a promise, with the intention to act on the therapist’s suggestion. Let’s grant, for the sake of argument, that Matt also has a proximal intention to communicate a moral demand to his friend. But let’s also suppose that the former intention, rather than the latter, is Matt’s primary intention: he renders it much more significant in his practical reasoning, and he cares about his mental health much more than whether he successfully conveys a moral demand to his friend. My intuition is that to the extent that blamers’ intentions provide a constraint for who counts as the proper targets to blame, what matters in Idiosyncratic Intention is Matt’s primary, not secondary, intention. However, the intentional version of the communication argument suggests that Matt’s secondary intention, rather than his primary intention, determines the standard for whether his friend is a morally responsible agent. But what explains this? The fact that Matt’s secondary intention happens to be one that has some conceptual connection to blame does not seem enough. Generally speaking, even if we take for granted that there is such a conceptual connection and it explains why the communicative intention is a special intentional aim, we still lack an adequate explanation for why it is special in terms of its normative implications about morally responsible agency. In any event, it seems to me that we can imagine that Matt does not even have the proximal intention to communicate; instead, it is possible that Matt’s intention about what he does at the moment of blaming is precisely to improve his own mental health, instead of communicating a demand. If so, then the distal/proximal distinction does not help defenders of the intentional version of the argument. There does not seem to be a strong conceptual connection between blame and a communicative intention after all.

Now turn to the functional version of the argument:

Macnamara (Reference Macnamara, Clarke, McKenna and Smith2015a, Reference Macnamara2015b) has defended an argument like this by focusing on blame’s function of “eliciting sincere acknowledgment of fault from the wrongdoer,” typically in the form of eliciting guilt (Reference Macnamara, Clarke, McKenna and Smith2015a, 222). I will focus on Macnamara in the remainder of this section. However, note that this functional formulation is another reasonable interpretation of the communication argument for theorists like Watson, Darwall, and Shoemaker as well. Given the problems with understanding blame as always involving a literal demand qua speech act, it is reasonable to reinterpret their demand-talk as the claim that blame has the function of communicating demands. I expect that my arguments below would also apply to this reinterpretation.

Macnamara explicitly endorses an etiology theory of function. According to the etiology account (e.g., Wright Reference Wright1973, Reference Wright1976), a function of M is F just in case that (i) F is a consequence of M’s being there, and (ii) M is there in part because it does F. As an example, the function of the heart is to pump since, first, pumping is a consequence of the heart’s being there, and, second, current tokens of the heart are there in part because past tokens of the heart did the work of pumping. Applying the etiology theory to blame, Macnamara suggests that (a) past tokens of blame did evoke uptake of their representational content in recipients, and (b) this partly explains why we have current tokens of blame now. To defend (a), Macnamara appeals to empirical evidence about the facial signatures of reactive emotions (Reference Macnamara2015b, 557–60); to defend (b), she contends that reactive attitudes and their emotional uptake have made a significant contribution to the “building and healing of the moral community” (Reference Macnamara2015b, 560–2).

But there is a complication. We need to distinguish between the claim that evoking emotional uptake is the only function of blame, and the claim that it is only one of the many functions of blame. Macnamara means the latter. As she puts it,

And Macnamara is right to focus on the weaker claim, because it is better supported by the evidence she cites. The fact that reactive attitudes have, through evoking emotional uptake, played a significant part in the building and healing of moral community is good evidence that evoking emotional uptake is one important function of blame, but it is fairly weak evidence for its being the only function. Macnamara herself mentions various intrapersonal functions like “informing the emotion-bearer about, and preparing her to respond to, her environment” (Reference Macnamara, Clarke, McKenna and Smith2015a, 219). We thus have good reasons to accept content pluralism about blame’s functions:

Again, there is no logical contradiction between accepting content pluralism about blame’s functions and the communication argument. But there is a serious tension here given that defenders of the argument tend to privilege the communicative aim of blame. The challenge is for defenders of the communication argument to explain what is special about the communicative function that it leads to a constraint about morally responsible agency, but other functions do not.Footnote ¹² Is it special in that the communicative function is the only interpersonal function of blame, whereas other functions are intrapersonal? I do not think this is tenable. Blame has interpersonal functions in addition to evoking emotional uptakes as well. For example, various theorists (McGeer Reference McGeer, Coates and Tognazzini2013; Kogelmann and Wallace Reference Kogelmann and Wallace2018; Shoemaker and Vargas Reference Shoemaker and Vargas2021) have argued that blame has an important signaling function. Shoemaker and Vargas defend an account according to which the definitive feature of blame consists in its “hard-to-fake costly signaling function” (591). A crucial feature of their account is that they take blame to signal “many types of information” and to “multiple targets” (590). For example, they argue convincingly that blame functions to signal, first, one’s commitment to a relevant norm, and, second, one’s own competence in detecting norms in a relevant domain. Along a similar vein, McGeer suggests that blaming emotions are “expensive signals” that are “hard to generate under conditions that do not naturally prime a person to experience them” (Reference McGeer, Coates and Tognazzini2013, 182).

Importantly, blame as a signal can be properly addressed to a recipient even if its recipient lacks the capacity to feel guilt. One can signal that one is committed to a norm by issuing blame, and a recipient can properly receive this signal as long as she is sensitive to the relevant contextual cues. The recipient does not need to be able to understand that the content of the relevant norm is correct in order to be able to get the signal that one is committed to that norm. Feeling guilt, by contrast, requires understanding the force of the content of the relevant norm. If we run an argument structurally parallel to the communication argument, it will go as follows: the capacity of receiving signals is necessary for morally responsible agency because blame has the signaling function. It seems that defenders of the communication argument would also need to accept this signaling argument. However, this then seems to undermine the alleged centrality of communicative capacities to morally responsible agency.

At this point, some may want to deny that blame has a distinctive signaling function. But the more general point stands anyway: blame plausibly has other interpersonal functions in addition to its communicative function,Footnote ¹³ and defenders of the communication argument need to either (i) explain what is special about the communicative function, or (ii) expand the scope of their argument by adding other constraints about responsible agency. Either way, it is not that easy to retain the force of the functional version of the communication argument while at the same time accepting content pluralism about blame’s functions.

Another potential response is that signaling is just a form of moral communication or a part of what moral communication consists of. It might be suggested that signaling one’s moral status is communicating one’s moral status in some sense, and the point of issuing such signals is for the sake of engaging in moral communication. Thus, it might be suggested that what I’ve characterized as a pluralist functionalist picture is, in fact, just a redescription of the communicative view when different forms and elements of the communicative process are spelled out. However, I do not think that signaling can be categorized as a communicative function. The point of communication, at least according to Watson and Macnamara, is to make the blamee understand that she has done something morally bad or morally wrong—for Watson, for the blamee to understand the moral demand; and for Macnamara, for the blamee to understand that she ought not to perform the relevant conduct and thereby feel guilt. Thus, the blamee needs to have the cognitive capacity to understand the relevant norms. But, again, signaling does not presume this cognitive capacity. There is no denying that the signaling function often works together with blame’s communicative function. However, they can also come apart, and my contention is that it is a non-trivial task to reformulate the communication argument to accommodate this possibility.

Maybe there is a way to appeal to a broad definition of “communication” such that all interpersonal functions of blame fall into that category. But I doubt this will be a useful definition. Its content would be too thin to have much explanatory value. Further, defenders of the communicative view (e.g., Watson Reference Watson and Watson2004, Reference Watson, Wallace, Kumar and Freeman2011; Shoemaker Reference Shoemaker2007) usually want to exempt psychopaths from blame, but even psychopaths would count as possessing communicative capacities under this broad sense of communication. This would at least be a striking result for anyone sympathetic to the communicative theory of responsible agency. The takeaway from the pluralist challenge is not that the communication argument is definitely refuted, but that a careful reconstruction of the argument is needed, which may require significant modifications to the communicative view of morally responsible agency as standardly conceived.

4.b The conjunctive strategy

One natural response to the pluralist challenge is simply to accept the implication that the communication argument needs to be expanded in a way that requires us to add further constraints on morally responsible agency:

I have said that this threatens to undermine the alleged centrality of communicative capacities to morally responsible agency, but one may not regard this as a huge problem. After all, the communication argument, even as originally presented, is supposed to offer only a necessary condition for moral responsibility. What’s the problem, one may ask, if we just allow more necessary conditions?

One worry is that the resulting requirement about responsible agency would seem ill motivated. The conjunctive view requires a responsible agent to be a fitting target when it comes to all functions of blame. However, if blaming an individual already fits one function of blame, why isn’t this enough for that individual to be eligible for the role of blamee? Why should it matter whether it fits another function of blame? As an analogy, a ring has both a function of indicating marital status and a decorative function. Imagine a context, like a fashion show, where wearing a ring is only fitting for the decorative function but not for the function of indicating marital status. It still seems appropriate for the model to wear a ring; the complaint that “but this does not fit the other function of rings!” is ill grounded. There are similar cases when it comes to blame. Imagine an employee, Erin, in a non-ideal workplace, where she is routinely asked to take on extra tasks, doing more work than she is paid for. Also imagine that she has an uncommunicative boss, Pam, who is too ingrained in her motto of “there is no such a thing as too much work” to be able to acknowledge that she has done anything wrong to Erin. Given this setting, it seems possible that Pam fails to be a fitting target when it comes to blame’s communicative function, but she remains a fitting target for blame’s signaling function. Even if communication is futile, Erin can still signal to Pam that she is committed to the relevant norms by blaming Pam. Erin’s blame would not fit all functions of the practice. However, her blame still seems appropriate in contrast to what the conjunctive strategy entails. After all, Erin’s blame fits the signaling function, so the complaint that “but blaming Pam does not fit that other function of blame!” seems beside the point. Note that this is all consistent with thinking that we may have other, non-function-related grounds to think Pam ought not to be blamed, for example, by arguing that she is non-culpable, does not deserve blame, etc. But to the extent that we rely on the communication argument, or a variant thereof, to draw conclusions about responsible agency, the conjunctive strategy seems to lead to an overly strong and ill-motivated requirement.

4.b The disjunctive strategy

If the conjunctive strategy seems too strong, then how about a disjunctive approach? The idea will be as follows:

The natural worry here is that the criterion is too weak. In the absence of other constraints for responsible agency, it implies that one is eligible for the role of blamee as long as one has the capacity to, for example, receive signals about norm commitments. But this is a really low bar. Again, even psychopaths—who many communicative theorists (Watson Reference Watson and Watson2004, Reference Watson, Wallace, Kumar and Freeman2011; Shoemaker Reference Shoemaker2007) tend to exempt—would count as morally responsible agents. We would also include young children who are developmentally mature enough to receive moral signals, but not yet mature enough to engage in the communication of moral demands or have the capacity to feel guilt. Further, the disjunctive strategy seems to render communication—either in the sense of communicating demands or evoking guilt—only one of the many elements that we need to attend to in deciding who is eligible for the role of blamee. There would seem nothing special at all about communication in this process; but the uniqueness of the communicative function of blame is an important part of why the argument becomes appealing in the first place.

4.c The context-specific strategy

Both the conjunctive strategy and the disjunctive strategy lead to some unwanted consequences. There is more to be said about these two strategies, and perhaps they can be developed in more careful ways to avoid the potential problems. But I think there is another strategy available:

The general idea is this. When a type of practice has more than one function, and to the extent that its functions constitute restrictions of the propriety of the practice, these restrictions ought to be context specific. Consider the ring-wearing example again. It is intuitive that whether ring wearing is appropriate in a fashion show has little to do with whether it would fit the function of signaling marital status. The natural explanation is that the latter function is not significant in this context. Instead, the contextually significant function is the decorative function. Thus, the fittingness of ring wearing depends only on this contextually significant function; whether it fits some other function of the practice is simply beside the point.

The context-specific strategy is based on a similar thought. The communicative function is only one of the many functions of blame, but there is a distinctive kind of context—let’s call this the communicative context—where this function is contextually significant. We then may be able to explain why the communicative function is special by explaining the uniqueness of this kind of context.

But there is a further problem. What if there are two or more equally or similarly significant functions in C? I am inclined to think that the formulation of the strategy then needs to take into account degrees of responsible agency. If blaming S would fit all contextually significant functions, S is a fully morally responsible agent in C; but if blaming S would fit some but not all contextually significant functions, then S has diminished morally responsible agency in C. The account then needs to be modified as follows:

I find the context-specific strategy promising. But what exactly decides the contextually significant function(s) in C?Footnote ¹⁵ One way to do so is to examine the function(s) that we, either based on some convention or based on some other shared understanding of an expectation, associate with C. Footnote ¹⁶ For example, the decorative function of wearing a ring is conventionally associated with the context of a fashion show. On this approach, the communicative function is a contextually significant function of blame in C when there is a convention or a shared understanding of an expectation that communication is an important goal in C.

Another way to decide the contextually significant function(s) is to examine the function(s) that should matter in C given the morally salient features of this context. The thought is that different kinds of contexts can highlight and make salient different elements of one’s situation that require our moral attention, which can then make a difference about the contextually significant function(s).Footnote ¹⁷ I think we should combine those two approaches to understanding contextual significance. Sometimes the communicative function becomes contextually significant due to our associating the function with C on the basis of convention or shared understanding of an expectation. At other times it becomes contextually significant because communication is an important goal given the morally salient elements of C that require our attention. Both kinds of circumstances constitute the “communicative context”—those circumstances where the communicative function of blame is contextually significant.

We can then have a version of the communication argument according to which the communicative function affects, to a significant degree, the extent to which one is a morally responsible agent in the communicative context. We may also preserve the uniqueness of the communicative function by defending the uniqueness of the communicative context. More specifically, the communicative context seems to be fairly pervasive: in many, if not most, blaming contexts, communication is at least one of the contextually significant functions. On one hand, we often associate the communicative function with blaming contexts, especially when we blame people whom we love or care about; we typically expect blame to serve a communicative goal and to initiate a conversation between us. On the other hand, there are also many situations that make salient the need for moral communication, for example, when blame targets a norm violation that occurs because of miscommunication or lack of shared moral understandings between the blamer and the blamee. In these circumstances, communication should be an important functional goal given the features of the situation that deserve our moral attention. The pervasiveness of the communicative context, I think, can potentially justify why one can privilege the communicative aim of blame over the other aims.

However, the context-specific strategy also allows contexts where the communicative function is not significant, or significant only to a fairly small degree. This can be what happens between Erin and Pam. I take it that the morally salient features of the context are the non-ideal features of the workplace. The need for communication is much less salient, especially given the radical asymmetry between the communicative power of the potential interlocutors. This can explain why Erin’s blame still seems appropriate: blame is fitting for its most contextually significant functions, including the function of signaling Erin’s commitment to norms. The context-specific strategy thus allows the possibility that one lacks the communicative capacity but is still a partially or even fully morally responsible agent.

The conclusion of the context-specific strategy relies on the assumption that morally responsible agency is context-specific. As I have mentioned in section 2, there are good reasons to accept a context-specific or “patchy” picture of responsible agency (see e.g., Vargas Reference Vargas2013), but this remains a controversial matter. My formulation of the strategy also assumes that responsible agency comes into degrees. More work needs to be done to defend these assumptions, but the context-specific strategy seems a more promising way out of the pluralist challenge compared to its alternatives. Despite its relatively moderate stance on the centrality of communication to responsible agency, it may provide the best reconstruction of the communication argument in a way that preserves much of its theoretical insight.