Second-Personal Approaches to Moral Obligation

According to second-personal approaches to moral obligation, the distinctive normative features of moral obligation can only be explained in terms of second-personal relations, i.e. the distinctive way persons relate to each other as persons. But there are important disagreements between different groups of second-personal approaches. Most notably, they disagree about the nature of second-personal relations, which has consequences for the nature of the obligations that they purport to explain. This article aims to distinguish these groups from each other, highlight their respective advantages and disadvantages, and thereby indicate avenues for future research.

Suppose Inga is standing on Ole's foot. 3 Inga might recognise that, since the current placement of her foot is causing Ole pain, moving her foot would realise a more valuable outcome.But, arguably, by recognising this "outcome-based reason" to move her foot, Inga would not yet recognise that she is morally obligated to do so (Darwall, 2006, p. 8).Her outcome-based reason seems to lack at least three features of moral obligation.
First, outcome-based reasons are best interpreted as normative reasons in what we might call the standard sense, i.e. as considerations that favour or recommend a course of action (Parfit, 2011, p. 31, Scanlon, 1998, p. 17).By contrast, moral obligations present a course of action as mandatory.This does not mean that they cannot be outweighed.
The feature is illustrated by the conceptual possibility of supererogatory acts, i.e. acts that are favoured by the balance of reasons but not obligatory (Darwall, 2013a, p. 5, Wallace, 2019, pp. 27, 41).For example, I may have all-things-considered reason to smile at strangers in the park, because doing so would slightly increase their level of well-being, without being obligated to do so.Whether we perform actions like this is in a sense 'up to us', whereas "[w]hen we are morally obligated, we are not morally free to act otherwise" (Darwall, 2006, p. 27).
Second, it seems that Inga's outcome-based reason would equally serve as a reason for another agent, Carlotta, who is also able to affect Inga's foot-placement.Moreover, if the overall number of painful foot-stompings in the world were minimised by leaving Inga's foot in place, then Inga's outcome-based reason would seem to favour that action instead.In sum, outcome-based reasons are agent-neutral.By contrast, if Inga is morally obligated to move her foot, then she is required to do so.Carlotta's intervention won't discharge Inga's obligation; nor will Inga's performance of an alternative action that minimises the number of painful foot-stompings in the world.In sum, moral obligations are agent-relative (Darwall, 2006, p. 6, Wallace, 2019, p. 38).
Third, by recognising the outcome-based reason, Inga does not yet recognise that she is accountable to Ole-or, for that matter, to anyone else-for withdrawing her foot.By contrast, by recognising that she is morally obligated to move her foot, she would recognise that her unexcused failure to do so would warrant Ole's adoption of reactive attitudes, such as blame and resentment, towards her (Darwall, 2006: Chapter 5, Schofield, 2021, pp. 33-34, Wallace, 2019: Chapter 3).Through such attitudes, Ole would implicitly address Inga, demanding that she comply with her obligation (Darwall, 2006, p. 17, Strawson, 1962;Wallace, 2019, pp. 70-71, Watson, 2004).
Arguably, moral obligations' link to accountability distinguishes them, not only from outcome-based reasons (or, more generally, reasons in the standard sense), but also from strict and agent-relative normative requirements, insofar as these are conceived third-personally or merely first-personally.To conceive of normative requirements third-personally is to regard them as part of "the fundamental nature of the universe", to use W.D. Ross's phrase, and thus as independent of the agent's practical perspective (Ross, 2002, p. 29).It seems that the only sense in which normative requirements, so conceived, might be addressed to an agent is by invoking the epistemic authority to point out facts about the world (Darwall, 2006, p. 6, Zylberman, 2017, p. 924).
To conceive of normative requirements merely first-personally 4 is to regard them as arising from within the agent's practical perspective, independently of her recognition of persons in a distinctively second-personal, or I-You, manner.Certain Humean or Kantian views might be interpreted along these lines (Street, 2012;Velleman, 1996).
It seems that the only sense in which normative requirements, so conceived, might be addressed to an agent is by invoking epistemic authority to point out that violating these requirements is inconsistent with the agent's own commitments (Darwall, 2006, pp. 26-27, Zylberman, 2017, p. 924).Therefore, neither third-personal nor merely first-personal normative requirements are addressable to agents in the accountability-invoking way that is distinctive of moral obligations.
Call the challenge of explaining the three distinctive features of moral obligation the problem of obligation. 5  Second-personal approaches to moral obligation are unified by the claim that this problem cannot be solved by without appeal to second-personal relations.The recent popularity of such approaches in philosophy is paralleled in psychology, where moral obligations are increasingly taken to possess a distinctive motivating force that cannot be reduced to other moral motives, such as sympathy, and depends on the capacity for second-personal relations (Dill & Darwall, 2014;Isern-Mas & Gomila, 2020, 2022;Tomasello, 2020). 6  But even among philosophical second-personal approaches, on which I focus here, there's no consensus about the exact nature of second-personal relations.To understand the disagreement, it's helpful to note that, arguably, Inga owes it to Ole to move her foot.Her obligation is directed: it is an obligation to Ole.Consequently, if she fails to comply (without excuse or justification), Inga wrongs Ole.Michael Thompson calls directed obligations "bipolar" because they connect two distinct "poles", e.g.Inga and Ole, through a "distinctive type of practical nexus" (Thompson, 2004, p. 335).
But one might think that it is also possible to be obligated to do something period, without owing it to anyone in particular.By violating such an obligation, one would simply be acting wrongly, without wronging anyone in particular.
Second-personal approaches generally agree that bipolar obligations lend themselves particularly well to an explanation in terms of second-personal relations (Darwall, 2013a;Eilan, 2014;Haase, 2014a;Thompson, 2004;Vandieken, 2019;Wallace, 2007;Zylberman, 2014).But only some second-personal approaches characterise second-personal relations as essentially giving rise to bipolar obligations.Call these bipolarist.Other second-personal approaches characterise second-personal relations as giving rise to bipolar and non-bipolar obligations alike.Call these inclusive.
In what follows, I discuss bipolarist and inclusive approaches in turn, and highlight their respective advantages and disadvantages.

| BIPOLARIST APPROACHES
Bipolarism is promising because it appears to provide a clear distinction between second-personal and non-second-personal phenomena, while solving the problem of obligation: a normative consideration is second-personal just insofar as it is owed to a particular person; and obligations' being second-personal in this sense explains their distinctive features.
First, when we owe an action to a person, this action appropriately becomes "a presumptively fixed point in [our] ongoing planning about [our] activities" (Wallace, 2019, p. 50).Thus, bipolarity promises to explain moral obligations' strictness.Second, when we owe an action to a person, we are bound up with this person in a "peculiar nexus" and "the rest of the world is, at least to a certain extent, closed out" (Thompson, 2004, p. 334).Thus, it might be the fact that Inga owes it to Ole to move her foot that explains why an alternative action, or Carlotta's intervention, cannot discharge Inga's obligation.In short, bipolarity promises to explain moral obligations' agent-relativity.Finally, when we owe an action to a person, this arguably makes it especially appropriate for this person to adopt reactive attitudes like blame or resentment should we fail to perform the action without excuse (Wallace, 2007(Wallace, , p. 29, 2019, p. 97), p. 97).Thus, bipolarity promises to explain moral obligations' link to accountability.
However, there are at least two general challenges for bipolarist approaches.First, while bipolar obligations clearly exhibit the features that give rise to the problem of obligation, this may be simply because they are obligations.Hence, to solve the problem of obligation, bipolarist approaches need to characterise bipolarity in a way that demonstrates that bipolarity explains obligatoriness (Schaab, forthcoming: §4).I discuss this challenge in more detail in Sections 3.1 and 3.2, by exploring two different versions of bipolarism.Both turn out to face distinctive challenges.
Second, bipolarist approaches solve the problem of obligation only if all moral obligations are necessarily bipolar.
I discuss this challenge in Section 3.3.

| Interactional Bipolarism
The bipolarist approach that I call interactional holds that A's obligation to B is bipolar in that it constitutively depends on A's and B's disposition to interact with each other using shared, distinctively bipolar concepts.This mode of interaction amounts to the adoption of a specific "posture of the mind" to one another (Thompson, 2004, p. 336; also see Haase, 2014a, 2014b, Lavin, 2014). 7When interacting in the relevant manner, A and B think of each other as a You, another subject, who in turn thinks of them as a You, and whom they are capable of wronging or being wronged by, respectively.B is thus recognised as "the victim", rather than merely "the occasion", of A's moral "fall" (Thompson, 2004, p. 340).
In other words, the relation of bipolar obligation can hold between two persons only if they are at least disposed to "reciprocally represent each other in terms of the relation" (Lavin, 2014, p. 284).We might say that the obtaining of a bipolar obligation between A and B constitutively depends on A's and B's ability to achieve common knowledge of its obtaining.That is, A and B must be able to know that the relation obtains, that each of them knows this, that each of them knows that each of them knows this, and so on.Thus, the thought of a bipolar obligation is a "thought for two" (Haase, 2014a, p. 122, citing Rödl, 2007, p. 187).
Recall that one of the general challenges for bipolarism is to provide an account of the bipolarity of moral obligations that can explain their obligatoriness.Interactional bipolarism might meet this challenge.First, if Inga's obligation to Ole is inseparably linked to Inga's disposition to represent Ole as a You who stands to be wronged if Inga does not comply, this might explain why, in Inga's deliberative field, removing her foot stands out from among other actions as a "a presumptively fixed point" (Wallace, 2019, p. 50).Second, the distinctive mode of representation linked to Inga's obligation might explain why "the rest of the world is, at least to a certain extent, closed out" (Thompson, 2004, p. 334).Third, if bipolar obligations depend on persons' representing each other as potential victims or perpetrators, this might explain these obligations' close link to accountability.Indeed, some argue that this representation is itself an instance of address (Haase, 2014a;Haddock, 2014).
However, interactional bipolarism faces challenges of its own.If A's obligation to B constitutively depends on A's and B's disposition to think of each other in a certain way, and the relevant thought is a 'thought for two', two factors severely limit to whom we can owe things.
First, some beings lack the capacity to form bipolar thoughts.Interactional bipolarism appears to rule out that these beings can be owed anything.Thompson replies that some of these beings, e.g.infants and some mentally disabled people, only accidentally lack this capacity, and that what counts is their membership in a "genus" whose members possess this capacity "as a rule" (2004, p. 372).However, even if this reply is successful, interactional bipolarism still seems to rule out that beings who non-accidentally lack the capacity for bipolar thought, e.g.most non-human animals, can be owed anything (Wallace, 2019, pp. 120-21).
Second, if A and B are to be able to attain common knowledge of their bipolar relation, they must be able to be aware of each other as sharing a bipolar thought, employing the same bipolar concepts.They must be capable of "a conscious encounter […], a unity of wills which rests on an opposition of I and you" (Lavin, 2014, p. 283).
We might commonly acquire bipolar concepts by being initiated into certain social practices.However, Thompson argues that this will not enable A and B to share bipolar thoughts unless they participate in the same social practice (Thompson, 2004, pp. 361-62).His reasoning is that, just as the speakers of distinct languages that happen to share the same vocabulary and grammar will only seem to converse with each other, users of bipolar concepts stemming from distinct social practices will only seem to share bipolar thoughts.Thus, social practices cannot underwrite distinctively moral bipolar obligations, which link persons as such (Thompson, 2004, p. 376).
Thompson discusses two potential attempts to render the disposition toward bipolar thought independent from social practices.The first, Aristotelian proposal holds that this disposition is part of our "shared specifically human life form" (Thompson, 2004, p. 376).According to Thompson, this proposal faces the formidable challenge of showing that human beings know non-empirically, "from the inside", that members of their life form are disposed to employ bipolar concepts (Thompson, 2004, pp. 376-78).Furthermore, it seems to rule out bipolar relations between human and non-human users of bipolar concepts (Thompson, 2004, pp. 378-79).The second, Kantian proposal holds that "I have synthetic knowledge 'from within' of a tendency to pairwise judging across [the manifold of persons]" as a result of the activity of pure practical reason (Thompson, 2004, p. 383).Thompson rejects this proposal by asserting, without argument, that it contradicts "a mild naturalism" (ibid.).Though I cannot do so here, it seems that much more could be said for and against both proposals than Thompson does in his brief remarks.

| Non-Interactional Bipolarism
The bipolarist approach I call non-interactional holds that moral obligations are bipolar insofar as they are "constitutively connected to claims that others have against us, just insofar as they are persons" (Wallace, 2019, p. 1; also see Wallace, 2007, 2013, Zylberman, 2014, 2017, 2018, 2021). 8  It might be tempting to understand claims in terms of what persons have a standing to claim (Feinberg, 1970).
According to Wallace, our moral obligations "derive from the fact that we inhabit an extensive notional community together with other beings who are 'equally real' (in Thomas Nagel's striking formulation), and whose interests are no less significant than ours" (Wallace, 2019, p. 37, citing Nagel, 1978, p. 14).By complying with these obligations "we are acknowledging our equal standing within such a moral community" (Wallace, 2019, p. 38).
Similarly, according to Ariel Zylberman (2021, p. 410-411): Moral necessity attaches to certain actions as such […] because original claims demarcate a domain of actions as required, prohibited, or permitted.This necessity is […] generated by the fact that moral agents act and interact in ways that are both consistent and inconsistent with reciprocal relations of respect and recognition.
Since non-interactional bipolarism does not regard bipolar obligations as dependent on the disposition to share bipolar thoughts, it avoids interactional bipolarism's challenge of showing that the set of beings that can share such thoughts coincides with the set of beings whom we commonly take to owe obligations to each other.
However, recall that one of the general challenges for bipolarism is that it needs characterise bipolarity in a way that demonstrates that bipolarity explains, rather than presupposes, moral obligations' obligatoriness.It is not obvious that non-interactional bipolarism does so.What is it about persons' claims that makes the corresponding obligations bipolar and simultaneously explains their strictness, agent-relativity, and link to accountability?It seems that, for all that's been said so far, non-interactional bipolarism simply posits a part of normative reality that takes the form of obligations, and states that these just are bipolar.What it does not offer is a clear explanation of their obligatoriness in terms of their bipolarity.
Zylberman seems to attempt such an explanation when he suggests that moral obligations are bipolar in that they are not "intelligible independently of our practical relations to others" (Zylberman, 2017, p. 926).He elaborates (Zylberman, 2017, p. 939): My right to not be tortured is not a reminder, a moral sign-post, as it were, of your non-relational duties.I am not the occasion of your moral fall.I am your victim.My standing is irreducibly practical because it is my practical standing before you as an equal bearer of rights.

SCHAAB
Here, Zylberman appeals to the idea, familiar from interactional bipolarism, that when persons owe things to each other, they are disposed to adopt a distinctively bipolar mode of representation toward one another.Indeed, his contrast between the "victim" and the "occasion" of one's "fall" is borrowed from Thompson (Zylberman, 2017, p. 939, citing Thompson, 2004, p. 340).
However, it is hard to see how non-interactional bipolarists can appeal to this idea (Schaab, forthcoming: §5; also see Haase, 2014a, p. 137).Once persons' claims are considered primitive parts of an independent normative reality, it's unclear why their intelligibility should depend on persons' disposition to represent each other in terms of these claims.Zylberman might insist that this is due to claims' bipolarity.But once the (otherwise plausible) link between bipolarity and forms of interaction, such as claiming, is severed, bipolarity can no longer be simply relied upon to explain claims' alleged dependence on a special mode of representation.
Can non-interactional bipolarism provide a satisfactory account of bipolarity without appealing to this special mode of representation?Although he agrees with Zylberman that claims are primitive (Wallace, 2019, p. 159), Wallace suggests that they might nevertheless be illuminated by their link to interests.He writes (Wallace, 2019, p. 163): [A] duty is directed to another party only if the considerations that go into establishing the duty center around that party, and it is personal interests of the putative claimholder that will be prominent within such a person-involving justification.
However, Wallace warns against reducing claims to interests (Wallace, 2019, p. 147).And he seems right to do so.For one thing, such a reduction would deprive claims of any fundamental explanatory role (Zylberman, 2021, p. 414).For another, it would make it doubtful that moral obligations can be explained in terms of claims at all (Wallace, 2019, p. 178, Zylberman, 2021, p. 414).After all, interests do not generally make agent-relative demands that we're accountable for meeting.Now, if we are to derive claims from interests without reducing the former to the latter, interests need to enter moral reasoning "as potential bases of moral claims" (Wallace, 2019, p. 178; also see Thompson, 2004, pp. 349-50).
But then claims' connection to interests cannot provide an account of the bipolarity of claims and their corresponding obligations, much less one that explains their obligatoriness (Schaab, forthcoming: §6).
There are two ways to challenge the conceptual impossibility of non-bipolar obligations.First, as we saw in §2, one might raise the possibility that some obligatory actions are not owed to anyone.Second, one might argue that, even if all obligatory actions are owed to someone, the notion of non-bipolar moral obligation plays an indispensable role in moral reasoning.Stephen Darwall advances two arguments along these lines.
First, Darwall argues that, even if all our pro tanto moral obligations happen to be bipolar, the notion of all-things-considered moral obligation is non-bipolar (Darwall,manuscript: §3).Suppose Inga has two conflicting bipolar obligations, one owed to Ole, the other to Carlotta.Darwall argues that we cannot intelligibly answer the question of what Inga is morally obligated to do, all things considered, in purely bipolar terms.Instead, we need to attend to the non-bipolar weight of Inga's obligations, i.e. their weight as (pro tanto) moral obligations period.
Second, Darwall argues that, while Ole has a special, "individual authority" to hold Inga accountable when Inga wrongs Ole, by adopting the reactive attitude of resentment, all members of the moral community, including Inga herself, share a "representative authority" to hold her accountable, by adopting the reactive attitude of blame (Darwall, 2013a, pp. 27-28, manuscript: §4).According to Darwall, this indicates that, even if Inga owes it to Ole to act in a certain way, Inga is also morally obligated, period, to do so.

| INCLUSIVE APPROACHES
Darwall is the main defender of an inclusive second-personal approach (Darwall, 2006(Darwall, , 2013c(Darwall, , 2013b)). 9In his view, all moral obligations, bipolar or not, are "second-personal reasons" in that their "validity depends on presupposed authority and accountability relations between persons and, therefore, on the possibility of the reason's being addressed person-to-person" (Darwall, 2006, p. 8, italics deleted).Thus, Inga's moral obligation to remove her foot from on top of Ole's not only implies her accountability to him, but it "would not exist but for [his] authority to address it" (Darwall, 2006, p. 13).
Thus, Darwall explains moral obligation's link to accountability by contending that moral obligations are grounded in relations of accountability.He then purports to explain the other two distinctive features of moral obligation in terms of that link.He argues that obligations are strict because it makes "no sense to blame someone for doing something and then add that he had, nonetheless, sufficient reason to do it, all things considered" (Darwall, 2006, p. 28).
And he argues that obligations are agent-relative because they are based on the agent's "[accountability] relations to others (and himself) viewed from his perspective within those relations" (Darwall, 2006, p. 7, italics added).
Darwall contends that this requirement can be reformulated as a contractualist moral principle: "an act is wrong if the act would be disallowed by a principle no one could reasonably reject our holding one another accountable for complying with" (Darwall, 2006, p. 301).Putative demands that don't express such non-rejectable principles "misfire", and thus fail to address a genuine obligation (Darwall, 2006, p. 44).By contrast, demands that express such principles are "in force" independently of whether anyone actually (implicitly or explicitly) addresses them (Darwall, 2007, p. 65).
An immediate challenge for Darwall's approach is that, paralleling interactional bipolarism, it appears to rule out obligations to beings who lack the capacity for second-personal address. 10In what follows, I discuss two additional challenges.
Darwall explains that the second-person standpoint "involves practically directed and directive thought, thought that is addressed to, and that makes a claim on, a free and rational agent" (Darwall, 2006, p. 9).This mode of thought can be employed internally, when we are addressed by ourselves or by "the moral community", which is not "any actual community composed of actual human beings" but, "like Kant's idea of a 'realm of ends,' a regulative ideal that we employ to make sense of our ethical thought and practice" (Darwall, 2007, p. 64). 11Therefore, "'second person' does not entail 'second party'" (Darwall, 2010a, p. 217).
But can one be addressed by oneself or a regulative ideal in an interesting sense? 12On Darwall's behalf, one might respond that second-personal address is like a practice in that its rules do not merely regulate but constitute certain offices and moves (Schaab, forthcoming: §9; also see Rawls, 1955;Schapiro, 2001).The contractualist principle then fixes demands that we, qua persons, cannot or must be taken to address to each other and ourselves.Just as one does not count as performing a move in chess unless one follows the rules of chess, demands that are (not) supported by the contractualist principle (don't) count as being addressed by persons.Whether this response is satisfactory partly depends on what's more important to us: solving the problem of obligation or maintaining pre-theoretical meanings for terms like 'demand' and 'person'.

| Bipolarity
Darwall purports to explain the distinction between bipolar and non-bipolar obligations in terms of the appropriateness of different reactive attitudes, resentment and blame, respectively (Darwall, 2013a, pp. 23-24; also see §3.3 above).However, Simon May notes (2015, p. 528): To feel resentment is, inter alia, to believe that one has been inexcusably wronged.But this means that the distinction between the two attitudes lies downstream from the concept of [bipolarity].
Given that bipolar obligations seem especially congenial to second-personal approaches, Darwall's apparent failure to explain the phenomenon of bipolarity might be disappointing.However, even if Darwall's approach does not explain what makes bipolar obligations bipolar, it might help explain their special moral significance.Some authors contend that the bipolarity of some obligations is at least partly due to social conventions that serve the common good (Cruft, 2019;Wenar, 2013).For example, the fact that journalists are owed certain obligations (e.g.not to force them to reveal their sources) is explained by their work's value to society.But this value arguably does not explain why violating these obligations disrespects individual journalists (Cruft, 2013).By linking bipolarity to the second-personal authority to resent, an account like Darwall's might fill this gap (Schaab, 2018; also see Cruft, 2019, pp. 76-77).That is, violations of our obligations to journalists might disrespect them because these obligations-despite being justified by the common good-give journalists a special standing to complain about their violation.

| CONCLUSION
Second-personal approaches to moral obligation may appear to face a dilemma: either only bipolar obligations are second-personal, in which case it seems questionable that a second-personal approach can solve the problem of obligation; or second-personal relations underpin bipolar and non-bipolar obligations alike, in which case (1) the significance of the label 'second-personal' threatens to elude us and (2) second-personality does not explain what makes obligations bipolar.Yet, the debate is at an early stage.Some of the challenges-and potential ways to overcome them-that I outlined here are only beginning to be recognised and explored.I hope that, by distinguishing different kinds of second-personal approaches and identifying their unique challenges, this article helps the debate advance to the next stage.