Taking up the second-person standpoint – relating to another as an I to a you – makes a difference to how one ought to act. Or, at least, so goes the central thought of a sub-tradition, nested within the last two centuries of Western philosophy.Footnote 1 What exactly that difference is, which the second-person relation brings about, is a point of divergence between various of the figures in this mini-canon. In this paper, I will identify a problem that has run through much of this tradition.

I argue that a series of the most prominent views in the tradition provide structural accounts of the practical significance of the second-person relation. That is, they claim that relating to others in the form of an I-you relation alters the structure of one’s own agential perspective. In this spirit, Fichte (2000) sees the summons of the other as a change that reconfigures the subject’s thought from a form of solipsistic, absolute freedom, to a form of practical thought that is structurally constrained by moral norms. What it means to introduce a structural constraint to practical thought is to introduce a new kind of practical reason. The kind of reason that proponents of the structural approach have in mind might be called ‘accountability-thinking’, or ‘second-personal thought’. In secularising the theological notion of a dialogical relation between a subject and God, Fichte inaugurates the philosophical discussion of the importance of dialogical interpersonal relations for the structure of human practical thought. I suggest that Buber (1958) and Darwall (2006) are both successors to Fichte in advancing structural accounts.

Against the structural accounts, I will argue that the practical significance of the second-person relation can be understood in terms of a substantive change. As such, my project is a deflationary one. In order to see the difference made by taking up the second-person stance, we need not think that the very structure of our agential thought is altered – that is, we need not think that we now have reasons of a new kind. The substantive change it brings about is comprised of a set of different practical reasons that can emerge through second-personal interaction with others. These reasons are no different in kind than those which can be entertained from the pre-second-personal, solipsistic form of practical reasoning. They are just reasons: facts which favour actions and practical attitudes.

What is more, this set of reasons that can be seen to depend on second-person relations are notably heterogenous. All sorts of considerations have a bearing on us when we relate second-personally to others, that otherwise would not. These include: reasons created through intentional address, such as requests, consent and commands; reasons to value others that can only be seen in light of second-personal interaction with them; and, possibly, further reasons that are metaphysically dependent on second-person relations.

I will provide a schematic way of understanding this variety of reasons. There are two main purposes of advancing this schema. First, it will significantly clarify the senses in which a reason could be said to depend on a second-person relation. Second, it will reveal the areas where there remains much uncertainty about how deeply shaped by I-you relations certain reasons could be, and so suggests areas for fruitful further enquiry as this tradition of thought progresses. To be clear, there are a number of interesting arguments for structural views – arguments by the likes of Fichte, Stephen Darwall, and Seabstian Rödl. I will not evaluate all of those arguments in this paper. Rather, I hope to show, simply, that such ambitious approaches to understanding the ethical significance of I-you relations are somewhat optional: the key intuitions that have underscored this philosophical tradition can, it seems, be accommodated by a relatively modest view, the one I call the substantive approach.

1 The Second-Person Relation

The second-person relation might be important for ethics, in ways that I will come on to discuss. The first thing to address is what the second-person relation is. At first blush, it is the relation between two people when each can address the other as you. If I relate second-personally to you, the way that I think of you assumes that you also relate second-personally to me. This is manifested in communicative address, where all speech to an addressee presupposes that the addressee will interpret the speech as from the addressor.

This way of understanding the relation can be encapsulated in the following definition. Let us say that:

  • Second-person relation: A’s relation to B is second-personal if and only if A regards B as a partner in the intentional interaction of addressing.

That is to say, the relation is second-personal if and only if A regards B in such a way that A could address B, and could be addressed by B in like fashion.Footnote 2 This construction enables us to think of second-person relations in verb form: A relates to B by having the relevant mental state, one of regarding B as a partner in the intentional interaction of addressing. In order for any intentional interaction such as address to be successful, B must regard A in like fashion. That is to say, for example, that I would not really be relating second-personally to you unless I thought that you were also relating second-personally to me.

Before moving on, let it be noted that this definition has it that in order for the relation to be second-personal, B need not actually see A in the way that A thinks that B does (which is also the same way that A sees B). A’s relation to B is second-personal even if A is wrong to think that B regards A as a partner in intentional interaction. An example of this that comes to mind for me is the recording that my brother, Pete, used to have as the greeting message on his answerphone, which he’d made to sound as though it was actually him answering the call: ‘Hiya!… yeah…not bad thanks, you?’ The message only tricked me occasionally, but when it did there would be a few moments when I thought I was talking to my brother on the phone. I thought I could hear him acknowledging my words, and I imagined Pete listening to me as the addressee of my words, just as I thought of myself as addressing myself to him. But he wasn’t really there.

A significant school of philosophers who have written about the second-person relation may take issue with the way that I have defined the subject matter. They hold that the relation that ought to be of interest is not something that can be reduced to the mental states of individuals. Rather the second-person relation, for such theorists, is something irreducibly social, something that exists only when two minds really do meet (see, e.g.: Rödl, 2014; Laing, 2021). However, perhaps despite initial appearances, proponents of such a view need not take issue with my definition of the second-person relation in terms of individual mental states. This definition does not rule out their view that there is a further phenomenon of philosophical importance that consists in something above and beyond the sets of individual mental states that can arise in social interaction. Everyone in these discussions agrees on the existence of the kinds of attitudes of regarding other people as partners in interaction, of which I have said that the second-person relation is comprised. This definition therefore helps to specify the phenomenon under discussion. I will remain neutral, here and in my arguments against the structural approach, on the question of whether there is a further phenomenon that is irreducibly social.

While in what follows I will leave open the possibility that the practical significance of the second-person relation can be understood only when that relation is construed as an irreducibly social entity, this paper’s core argument will offer some modest evidence against that view. Later, I will sketch my substantive approach to the importance of second-personal relations for practical reasoning. This approach will appeal only to the conception of second-personal relations as reducible to sets of individual mental states. If the sketch is successful in accommodating our intuitions about how second-personal relations should figure in practical thought, then it will thereby have swept up some of the motivation for a less deflationary, irreducibly social theory of the second person.

The definition offered explains the sense in which a family of paradigm examples of I-you interaction are indeed second-personal. Stephen Darwall’s central example of the addressing of a moral demand, ‘Get off my foot!’ is second-personal in that the addresser regards the one treading on their foot as a partner in this addressing kind of interpersonal interaction, and thus as one who reciprocates this recognition (Darwall, 2006, p. 5). Similarly, Margaret Gilbert’s (2014, p. 329) favoured example is of a moment of eye contact establishing mutual acknowledgement between herself and the other person sitting at the table in the library. This too is second-personal in that the manner of looking into the eyes of the man opposite presupposes that he is looking back in that self-same way: a way in which each acknowledges the other as a partner in intentional interaction.Footnote 3

Equipped with a clarified understanding of what a second-person relation is, I would like to turn to consider why this relation might be of interest to practical philosophy.

2 Practical Significance

What changes when one relates to others second-personally? Does taking up the I-you stance make any difference to an agent, and how they ought to act? This question has, I suggest, driven much of the interest from philosophers in thinking about the second-person relation. The question can be expressed more precisely using a concept that Darwall (2006, p. 4) introduces – the concept of a second-personal reason. I will clarify what it is for a reason to be second-personal in a moment, but it is useful here because it enables the following realisation. The fundamental question to this domain of enquiry is whether there are any second-personal reasons.

The reasons we have to cooperate with each other seem to be second-personal (Heal, 2014), as do those we have to adhere to the requests that others make (AUTHOR). Some have recently suggested that our reasons to adhere to the norms of the traditions and practices that we share with others are also, in a sense, second-personal (Brandom, 1994; Satne, 2014, 2017). Others, such as Darwall, have argued that moral obligations are second-personal reasons. I will take up this claim in section four. The underlying idea is that there is a conceptual relation between moral obligation, and the second-personal, addressing attitude of moral blame.

Later in this essay I will argue that the standard approaches to explaining the practical significance of the second-person relation are misguided. What is apparent is that considering various recent discussions of the second-person can be seen to have an overlapping theme. That is, they are all concerned, in one way or another, to explain the practical significance of the second-person relation by advancing accounts of second-personal reasons. It remains for this section to formalise my understanding of what a second-personal reason is.

  • Second-person reason: A reason is second-personal if and only if it is a fact which speaks in favour of an agent performing an action only if that agent stands in a second-personal relation to some other person.

When Darwall talks about second-personal reasons, he often means to restrict the term to a certain sort of deontic reason: a moral demand voiced from the second-person perspective (Darwall, 2006, p. 4). However, at points he does also use the term in a broader sense such that non-moral reasons can be second-personal if they are grounded in a relation between an addresser and an addressee, an I and a you (Darwall, 2006, p. 55). The definition of a second-personal reason that I am offering here captures that broader sense of Darwall’s, which is also the more natural way to grasp the term. Moreover, besides being natural to understand, this sense of ‘second-personal reason’ is highly useful for clarifying tracts of philosophical discourse. That alone is more than sufficient justification for sticking with the above definition.

3 A Tradition of Enquiry

In §VI below, I will return to the discussion of the senses in which reasons might be second-personal, as my definition allows for a variety of ways in which a reason might depend on a second-person relation. First, though, in the present section I want to illustrate the extent to which this question of whether there are reasons that are second-personal has been the locus of a tradition of thought in Western philosophy, one beginning with J.G. Fichte and culminating in the apparently disparate discussions of second-personality that I described in the previous section. My motivation here is this. Attending to some of the historical approaches to the practical significance of the second-person relation can be an effective way to animate the attractiveness of the idea that the second-person relation is indeed practically significant.

Although they rarely call it by this name, the second-person relation is an important element in the ethical thought of both Fichte and Hegel.Footnote 4 Each in their own way, these post-Kantian German idealist system builders propose that recognition from another person is constitutive of an individual’s standpoint as a mature moral agent. Understanding this idea and its motivation is centrally important for understanding the rest of the tradition of philosophy about second-personal reasons, not least because this is the germ of what I will identify in the next section below as the structural approach.

As a legacy of Kant’s first Critique, the German idealists inherited a certain perspective on metaphysics and epistemology: the perspective of an experiencing subject. That is to say, they were concerned to articulate how the subject could have knowledge of the world, but they supposed ‘the world’ to be just that which could be understood by a cognizant subject. Clearly that is an extremely simplified statement, but it is not important for my purposes to discuss idealist metaphysics in any detail. What is important is that one can see how, given their starting point in considering the world as it appears to a subject, the idealists had to provide some account of the apparent importance of other people’s interests. Since other people are appearances in the manifold of things in the world, why should the subject accord them the great significance that we do feel inclined to accord to others? Why do other people matter?

These questions are especially pressing for idealists since the mattering of other people requires a special place in the system. It cannot be part of the structure of subjective thought – in the way that Kant thought that the conditions of experience can – since the fact that the other’s life matters is a fact about something outside of the subject, and independent of the subject. And yet, at the same time, this fact cannot be part of the material realm that the subject experiences, since it is not a material fact.Footnote 5 That the other has interests – in survival, health, flourishing – might be thought of as an empirical matter that the subject could apprehend in experience. But the fact that these interests matter – a fact which underpins all of the subject’s moral reasoning – is resolutely immaterial.

Even if one is not an idealist and does not endorse an idealist metaphysics of subject and subjectively-determined world, the problem facing Fichte and Hegel is compelling – at least from a certain removed, sceptical standpoint. In order to understand the reasons that we have to respect others’ interests, it seems some explanation is needed of why those interests should matter. And on the face of it, such an explanation is difficult to offer. The idea of a second-personal reason is the solution that Fichte comes up with.

He proposes that it is the summons of the other that has the effect of calling the subject to responsibility (Fichte, 2000, p. 31). There is some debate over exactly how that should best be understood (Darwall, 2005; Ware, 2010). But it seems uncontroversial that Fichte is proposing that moral responsibilities are second-personal reasons in the sense that they would not have the normative purchase that they do have on agents, were it not for those agents having been summoned by others (or at least one other). And what is more, it seems clear that whatever else Fichte might mean by ‘the summons’, it requires the subject relating in the position of an addressee, towards another, the summoner, in the position of an addresser. And thus, the subject must relate second-personally to at least one other in order to be constrained by moral obligations. So moral obligations on Fichte’s view are second-personal reasons.

This constitutes a solution to the problem of how to explain the fact that other people matter. This solution does not work by providing some otherwise-hidden epistemological access-route that the subject can take to establish foundational knowledge of the other’s moral significance. Rather, the idea that the summons of the other is a condition of the subject’s responsibilities is a way of saying that the domain of moral responsibilities begins with certain presuppositions in place. It does not begin with a solipsistic search for foundations, but with an intentional interaction with others. Action takes place in the context of practices of address, demands, and cooperation, and such practices simply presuppose that the other’s life and ends are valuable things. Developing a theory of morality as dependent on second-person relations between the subject and others is a way of circumnavigating the problem of why others matter. This, I think, is what initiated the tradition of exploring the idea of second-personal reasons.

It is often overlooked that this very starting point led, more or less directly, to the later discussions of the second-person in Western philosophy. Consider this remark from Feuerbach (1972, §59):

The single man in isolation possesses in himself the essence of man neither as a moral nor as a thinking being. The essence of man is contained only in the community, in the unity of man with man – a unity, however, that rests on the reality of the distinction between “I” and “You”.

I will not discuss how other German idealists such as Hegel, or indeed their materialist successors such as Feuerbach, spell out their versions of the intersubjective basis of ethics. Suffice it to say that the role of recognition, which is implicit in Fichte, is drawn out explicitly in Hegel. In this quoted passage, Feuerbach goes a step further and makes explicit that the relation of recognition that is relevant to understanding the intersubjective deliberations of practical life is a second-person relation, between an I and a you.

Interestingly, Buber cited this particular passage from Feuerbach as one of the sources of inspiration that led him to think about the philosophical importance of I-you, or I-thou relations (Buber, 2002, p. 32). Indeed, the German idealist approach to appreciating the practical significance of the second-person had a more pronounced influence on Buber’s evocative considerations of I-thou relations than merely prompting them. Buber shares with Fichte and Hegel the view that when the subject enters into the standpoint of dialogue with another, this brings about a thorough reconfiguration of the subject’s own outlook as an agent. This is clear from a famous passage early in I and Thou where Buber describes taking up an I-thou relation to another person:

This human being is not He or She, bounded from every other He or She… But with no neighbour and whole in himself, he is Thou and fills the heavens. This does not mean that nothing exists except himself. But all else lives in his light. (Buber, 1958, p. 8).

This description of what it is like to enter whole-heartedly into a second-person relation – to give one’s interlocuter one’s full attention – concurs with the insight from the idealists: mutual recognition fundamentally alters the subject’s outlook. But Buber also adds an element that was at best inchoate in Fichte’s and Hegel’s accounts. He connects the importance that the recognition of others has for our practical thought, on the one hand, with the character of the experience of mutually acknowledging interpersonal interaction, on the other. This connection between ethics and phenomenology is the theme that links Buber with other figures in a significant early Twentieth-Century generation in the tradition under consideration. One such figure is Emmanuel Levinas.

In Buber’s aphorisms, the impact made on the subject by the other-as-thou is described in suggestive, religious terms, as above: ‘he is thou and fills the heavens’. While Buber does also describe the way that love of the other is occasioned by I-thou relations (Buber, 1958, p. 15), he does not draw attention to any distinctly moral character of the I-thou relation. That is, he does not say that the experience of relating to another as thou is accompanied by the feeling of being responsible for that other, constrained by moral obligations towards them, or moved to help them.

By contrast, Levinas does bring out this moral dimension of the phenomenology of second-person relations. For the purposes of my argument in the next sections of this essay, it will be helpful to present, briefly, this ethical phenomenology. The key point that I wish to emphasise here is that there is some intuitive evidence – which Levinas articulates – to support the view that the experience of second-personal interaction can and often does engender a sense of responsibility for, and to, the other. This is summarised in Levinas’ (1969, p. 195) claim that ‘the formal structure of language… announces the ethical inviolability of the Other’. But consider a passage in which the idea is developed more fully:

The face [of the other person] resists possession, resists my powers. In its epiphany, in expression, the sensible, still graspable, turns into total resistance to the grasp. This mutation can occur only by the opening of a new dimension…. The expression the face introduces into the world does not defy the feebleness of my powers, but my ability for power. The face, still a thing among things, breaks through the form that nevertheless delimits it. This means concretely: the face speaks to me and thereby invites me into a relation incommensurate with power exercised, be it enjoyment or knowledge. (Levinas, 1969, p. 198)

The epiphany that Levinas is referring to, and the opening of a new dimension, are the emergence of a second-person relation between subject and other, a relation which Levinas calls the face. As he rightly notes, one need not actually communicate verbally in order to stand in this communicative, dialogical relation. It is enough to establish eye-contact, which is already communicative. This is the sense in which the face ‘speaks to me’ and ‘invites me into relation’ (a point that is echoed in Margeret Gilbert’s example of mutual recognition that I mentioned above). Of that emergent second-person relation, Levinas is claiming that it comes with a certain feeling of ethical responsibility. More specifically, the claim here is that exercising power over another person – through violence or any other means – is incommensurable with relating with them second-personally. That is, thinking of another as a partner in dialogue renders any thoughts inapt which would consider the other with anything other than the respect and attention that they deserve as one’s interlocutor.

My purpose here is not to defend Levinas’ claim in any detail, beyond having simply presented it. What is worth emphasising, though, is that the whatever strength his account has lies in its affinity with our general experiences of interpersonal encounters. He is not claiming that every time we meet another person, we are overwhelmed by their ethical importance – just that when we relate second-personally to others and devote our attention to them as our interlocutor, then we can be struck by their moral importance and our duty not to harm them. Moreover, Levinas suggests that this feeling makes sense, since harming the other or doing violence to them are actions performed on the other rather than to them. In this way, such unethical conduct toward the other cannot be so much as entertained for as long as one remains singly in the second-person standpoint of an addresser-cum-addressee. To summarise this section, I have shown that there is a shared concern between a number of philosophers spanning from Fichte, through Hegel and Feuerbach, to Buber and Levinas. Thinking of each of these figures through the prism of their focus on the question of whether there are second-personal reasons illuminates a significant degree of continuity between their intellectual projects. In this sense, they deserve to be thought of as the backbone of a tradition of thought, to which others too have contributed. I have drawn out two influential ideas that have run through the tradition. The first is the Fichtean idea that the nature of first-personal agency is reconfigured through the second-person relation, so that moral constraints are brought into force. The second is the Levinasian idea that reflecting on the character of the experience of second-personal interaction reveals an ethical quality. At least in paradigm cases, there is something that it is like to relate second-personally to others, and that something involves a feeling of being responsible for, and to, the other.

4 The Structural Approach

The key figures in this tradition do not only share an interest in second-personal reasons. They also share, for the most part, an approach to explaining what second-personal reasons are. I suggest that it is helpful to give this approach a name, so let’s call it the structural approach. In short, this is the view that relating to others second-personally effects one big change, a fundamental alteration in the subject’s agential perspective. In this section I want to clarify what that means.

The first step is to clarify exactly what the structural approach is an approach to. The answer to this can be drawn from my earlier discussion. The structural approach is one way to approach the question of whether there are any second-personal reasons. This label usefully applies to a family of views that claim that there are second-personal reasons, and such reasons are a product of the way in which a subject deliberates from the second-person standpoint, which is somehow different from how they deliberate otherwise. Talk of ‘ways of deliberating’ is one way to make sense of the idea of the ‘structure of one’s agential perspective’. This will become clearer through an example, shortly.

In addressing the question of whether there are any second-personal reasons, proponents of the structural approach seek to explain the two things. These two things emerged in the previous section as the themes of the tradition of enquiry into the second-person. Namely, those are: (i) the fact that other people’s interests matter, in a way that seems to be presupposed by our shared deliberations about what to do; and, (ii) the phenomenological character of second-person relations as experiences that seem to change how one ought to act.

The structural approach, then, accommodates these two explananda in its account of second-personal reasons. It says that relating as an I to a you brings about an alteration in the way one thinks about what one ought to do. Such a relational standpoint ushers in a paradigm of relational thought, in which what I should do is thought of in connection with what you could hold me accountable for doing. This meets the first explanandum. The structural approach makes sense of the way that practical thought presupposes the value of other people’s lives and interests. The form of second-personal thought naturally does not call into question the value of the other’s life and ends, since it is the form of contemplating the demands made by the other. The second explananda is straightforward for the structural approach. On this view, second-person relations seem to make an impact on what the subject out to do because they emphatically do: such relations alter the very structure of practical thought. This can be seen more clearly through the most well-known theoretical account that takes this approach.

Darwall (2006, pp. 11–15) argues that there is a conceptual relation between the second-person relation, accountability, and moral obligation. Moral obligations just are those reasons that one is bound to comply with on pain of being held accountable through the moral blame of others. Moral blame is, inter alia, a second-personal emotion: a way of addressing others and thereby holding them to account for their moral conduct. Moral responsibilities are second-personal reasons. And, therefore, any practical thought in which one considers obligation is structured by – at least the possibility of – second-person relations. Consideration of one’s moral obligations is necessary to and characteristic of mature practical thought. In this sense, Darwall provides an account of the structural change that second-person relations effect on the agential perspective of a mature moral agent.

This is the most developed expression of a long history of the structural approach. But it adds flesh to the bones of the approach as it was articulated by earlier figures. Darwall’s view is one way to make sense of Fichte’s (2000, p. 37) claim that ‘the human being […] becomes a human being only among human beings’; likewise Buber’s (1958, p. 11) claim that ‘I become through my relation to the Thou; as I become I, I say Thou’; and likewise also Levinas’ (1969, p. 201) claim that ‘The face opens the primordial discourse whose first word is obligation’. Darwall’s explanation of a conceptual connection between moral obligation and the second-person relation is one way to redeem the structural approach.

The conceptual route is only one variant of the structural approach, however.Footnote 6 This is why it is better to call it an approach – encompassing a family of views – rather than thinking of it as the structural view. An alternative to the conceptual route would be an ontogenetic one. That is, one might claim that as a matter of developmental psychology, in order to become a morally mature agent, one requires second-personal interaction with others. This seems to have been the idea that R.D. Laing (1960, p. 26) had in mind in his thought that ‘[h]ere we have the paradox, the potentially tragic paradox, that our relatedness to others is an essential aspect of our being, as is our separateness, but any particular person is not a necessary part of our being.’ Perhaps there could be a way of elaborating the ontogenetic story of a mature moral agent where the capacity to engage in developed practical reasoning can only come about through second-person relations with others.Footnote 7 This story might thus find some of the reasons that ought to be entertained by a mature agent to be second-personal reasons, in that they depend for their normative force on some formative second-person relations between the agent for whom they are reasons, and some others.

If the ontogenetic route could be worked out, then that would provide an account of the practical significance of second-person relations as a kind of significance to the structure of the subject’s rational agency. Darwall’s conceptual route does just this. Perhaps there could be still other routes for cashing out the common thought to many in this tradition that the second-person is structurally significant to the first-person’s practical perspective. In the remainder of this essay, I will suggest that any version of the structural approach faces some significant obstacles, and that moreover, there is an alternative to the structural approach: a much more straightforward way to explain the way that second-person relations make a difference to our practical thought.

5 Problems for the Structural Approach

The structural approach has attractions. Specifically, it accommodates both of the explananda that have, in one way or another, motivated much of the tradition of thought about the second-person. It explains how the value of other people’s lives becomes embedded in our practical thought, as well as explaining the intuition that encountering another person as a you can make an impact on what one ought to do. However, this approach also faces some significant obstacles. In this section I will present a twin pair of objections to the structural approach. It sees all second-personal reasons as being second-personal in the same, structural way. As such, it ends up providing an account of second-personality which is at once either too broad, or too narrow, or both.

Consider first the charge that any account of second-personal reasons on the structural approach will be too broad. On any version of the structural approach, relating second-personally with others is the transition from self-concerned solipsism, to mature, morally aware agency. This way of understanding how I-you relations make a difference says that the I-you relation makes one, dramatic difference. As an example, recall that one route that the structural approach could take was the ontogenetic route, which was gestured at by Fichte, Buber and Levinas. This view envisages a transformation that is brought about by the subject relating second-personally to at least one other person. Once the other has encountered the face, to use Levinas’ (1969, chpt. 2.I) terminology, then the subject can enter the domains of Language, Discourse and Justice.

The problem with the idea that the second-person relation makes just one difference, albeit a dramatic one, is that this seems to be at odds with the diversity of ways in which a reason can be second-personal. This approach implies that all mature practical reasoning – or at least that which is about morality – is structurally second-personal in the same way. But it is not. To illustrate this, consider a contrast between two practical reasons that are both equally second-personal according to the structural approach. On the one hand is the reason one might have to do donate some money to the Red Cross. Let’s suppose that in a certain set of circumstances, one is morally obliged to donate to the Red Cross. The obligation is grounded in, among other things, the fact that donating money would contribute to the alleviation of suffering for people in desperate situations. On a structural approach this reason might count as a second-personal reason. The structural approach says that the morally obligatory reason one has to alleviate the suffering of others is one that relies on one’s sense of responsibility for and to others, and that sense of responsibility is rooted in second-personal thought.

On the other hand, consider the reason I have to trust you to keep a secret in virtue of the fact that you look into my eyes and ask me to trust you. This reason seems to depend on a particular second-person relation a great deal. Whatever reason I have to trust you that is generated by you imploring me could not possibly exist but for the relation of mutual acknowledgement between us. The reason is a product of the fact that we are you for one another.

The structural approach says that both reasons mentioned here are second-personal in the same sense and to the same extent. It says so, because all reasons that have a second-personal character, are such in virtue of resulting from the one big structural effect that second-person relations bring to practical agency. But it is worth attending to the contrast between the two reasons just outlined. The reason to give to charity is putatively second-personal because, as a moral obligation, to understand this reason requires an understanding of the blame that would be appropriate if one were to ignore it. That legitimate blame is something that the agent could face from any representative member of the moral community. So, in the case of this reason, its second-personal character is grounded in a relation that can be borne in the agent’s imagination between themselves and a generic representative second person: a faceless ‘you’.

By contrast, in the second case an embodied interaction yields a reason for one person to trust the immediate other before them. It seems plausible, on the face of it, to think that the normative weight of the resultant reason has something to do with the needs of the particular other person who asks for trust, with the value of their particular life and interests, with the fact that they are present in mutually-acknowledged interaction, embodied and therefore immanently vulnerable to one another, and with the fact that the reason is presented self-consciously for one person by the other in the context of this mutually recognitive relation. In virtue of these contrasting features, it seems as though the second reason is second-personal in senses that the first is not. It also seems, on the same basis, that the second reason is second-personal to a greater extent than the first. The structural approach, though, lacks the resources to make sense of these appearances.

The argument just presented treats the Darwall of The Second Person Standpoint as an exemplar of the structural approach. A defender of Darwall may point out at this point that his work on the second person has evolved in the years since that landmark book, and he has now proposed ways to accommodate the kinds of varieties of second-personal reason that my argument invokes. This is true. In subsequent work, Darwall (2013, chpt. 2) has offered analyses of reasons that he calls ‘bipolar obligations’, which are keyed to the dyadic relations between two individuals, such as a promisor to a promisee. Bipolar obligations, like the more general category of ‘moral obligations period’ are (as the term obligation suggests) reasons to be considered in a deontic key. That is, failure to comply with such reasons licenses deontic reactions such as resentment (in the case of bipolar obligations) and moral blame (in the case of all moral obligations). But in still further work, Darwall has offered analyses of the second-personal dimension of other normative phenomena, which fall outside the realm of deontic practical constraints. For instance, Darwall has argued that love (2016), trust (2017), gratitude (2019), and being in the presence of another person (2021), are all parts of interpersonal life that rely on people relating second-personally to one another. Clearly, then, it would be unfair to accuse Darwall – when his work in the last two decades is taken as a whole – of failing to attend to the diversity of kinds of second-personal reasons.

However, my argument is not against Darwall himself, but against the structural approach. The Darwall of The Second Person Standpoint is a paradigm representative of the structural approach precisely because he offers a view on which relating second-personally to others brings about one decisive change to the structure of a subject’s agency: it introduces the possibility of deontic practical thought. His later articles, by contrast, implicitly challenge the structural approach. They suggest that the practical effects of second-person relations are multifarious: some effects pertain to the possibility of deontic thought in general, some to bipolar normativity in particular, and some are unrelated to deontic thought altogether. The kinds of practical significance that Darwall attributes to the second-person across his work are not united by a common form of deliberation (sometimes second-personality is associated with ‘accountability-thinking’, other times, contrastingly, with ‘openness’ (e.g. Darwall, 2021, p. 11)). Neither are these kinds of practical significance united by a common kind of reason (sometimes second-personal reasons are deontic, other times, contrastingly, not (e.g. Darwall, 2019)). Taken as a whole, Darwall’s work opposes the structural approach because it denies that the second-person relation is practically significant simply for making one transformative, structural change to the subject’s practical agency. Darwall’s work does not construe the second-person relation in a way that is too broad to capture the diversity of forms of second-personal reasons; the structural approach does.

As well as being too broad in this way that obscures the diversity of second-personal reasons, the structural approach risks being too narrow. One prominent way to spell out the structural approach, as noted above, is as Darwall does: to articulate a common second-personal structure to all and only moral obligations. Darwall argues that moral wrongness just is moral blameworthiness, and blame is a reactive attitude that is addressed second-personally, and that therefore all moral obligations presuppose the possibility of second-personal address. In this sense moral obligations are second-personal reasons, and the category of second-personal reasons is restricted to moral obligations.

This is a problem for such variants of the structural approach, because restricting the category of second-personal reasons to moral obligations is too narrow a restriction, on two counts. First, it excludes moral reasons that are not obligatory, but which do seem to be second-personal. Second, it excludes reasons that also seem clearly to be second-personal but which are not moral at all. Examples of each will be sufficient to support this point.

Suppose we are friends, your rent is due and through no fault of your own, you can’t pay it. I might well be thought to have some kind of reason to support you here by giving or lending you some money. Exactly what kind of reason I have to help you will depend on some further factors. If I have plenty of money to spare, I’m the only person who could help you, and we both know that you’ll be evicted if you miss the payment, then it seems reasonable to say that I am obliged to help you meet the rent. This is reflected in the fact that you would be entitled to blame me if I refused to help. On the structural approach which identifies all and only moral obligations as second-personal reasons, my reason to help you in this case counts as a second-personal one.

If, however, the situation is slightly different, my reason to help you would not be counted as second-personal according to this variant of the structural approach. Suppose, for instance, that I myself am short of money at the moment, or that there are others who might be able to support you instead of me, or that it is only possible but not likely that you will actually be evicted if you miss the payment. In these cases, I would still have reason to help you, but it may not be obligatory. You might well acknowledge that if I did not help you, that would be fair enough and you would not blame me for it.

Despite not being entitled to blame me, though, if I failed to come to your aid in this situation, you may well be entitled to some other reactive attitudes that are addressed towards me second-personally. You might be hurt, feel let down by me, disappointed in me. Whilst I wouldn’t exactly be bound to apologise (since we acknowledge that in the circumstances I did not do anything wrong in not helping you) I may nonetheless seek to make it up to you. You would not exactly be entitled to forgive me, but something very much like forgiveness, would be at your discretion: the power to allow the whole episode to be forgotten.

These appropriate reactions to the event are replete with second-personal interaction. If what makes the moral obligation second-personal is its conceptual connection to the second-personal address of moral blame, then it seems that some supererogatory altruistic reason – like the one I have in this example to help you with your rent – should be thought of as second-personal too, and for much the same reason.Footnote 8 That is, supererogatory altruistic reasons, just like obligations, can imply the appropriateness of various forms of second-personal address: the supererogatory analogues of blame, apology, forgiveness, and so on. The fact that such supererogatory reasons are not classified as second-personal reasons on the obligation-centred variant of the structural approach suggests that at least this variant of that approach is too narrow in scope.

As mentioned, there is also a further worry about the structural approach being too narrow. This worry holds even if the structural approach pursues a line of thought on which the structural change brought about to the subject’s agency by the second-person relation is broader than simply instituting moral obligations, and includes also non-obligatory moral reasons. Suppose we’re still friends, and I invite you to spend the weekend hiking with me. My invitation provides you with a reason to go hiking, and it looks as though this should be considered a second-personal reason – emphatically so. Not only does the act of inviting require a second-personal relation of addresser and addressee as inviter and invitee, but also the relationship of friendship that gives rise to my invitation is such that it could not itself arise without a history of interaction where that interaction is, inter alia, second-personal. However, the invitation does not look like a moral reason. It would not be supererogatory of you to accept my invitation since I do not want you to come for my sake. I would like you to accept for your own sake, in acknowledgement of the fact that it would be a bit of fun to come for a hike with me for the weekend.

Now, a defender of the structural approach may make a similar rejoinder here to the one made a few pages ago. They may object that while my argument might suggest that a view like Darwall’s in The Second-Person Standpoint is too narrow to account for the diversity of second-personal reasons, that argument does not target the structural view as such. So, at this point, it is worth considering whether there is any alternative elaboration of the structural approach that may fare better than the Darwall of the mid-2000s at accommodating the range of second-personal reasons. What we are searching for now, then, is a view that explains how the possibility of relating second-personally to others makes one transformative change to the structure of the subject’s practical reasoning, but where that view does not follow the Darwall of ‘06 in identifying that change with the emergence of deontic practical thought. Perhaps, for instance, the structural change is to be identified with the emergence of some other characteristic of dialogue: not one’s accountability to the other, but, maybe one’s preparedness to join them in forming a cooperative unit and occupying the stance of the ‘we’, or, simply, one’s openness to being affected by the other.

However, neither of these proposals are redeeming. Not all second-personal reasons are characterised by a connection with the perspective of a ‘we’, or at least not in the same way. One might think that moral obligations in general are second-personal because they invoke the standpoint of ‘we members of the moral community’. What is more, one might think something similar of the particular domain of bipolar obligations, such as your directed duty to me not to break your promise. In the bipolar context, perhaps the we-subject that is invoked is something like ‘we two, who mutually recognise one another as persons’. Similarly again, perhaps the reason to take seriously the camping invitation mentioned earlier invokes the standpoint of ‘we friends’.

There is a straightforward problem for the idea of elaborating the structural by conceiving of second-personal reasons in this way as characterised everywhere by their conceptual connection to a we-subject. In order to make sense of the variety of second-personal reasons, this approach has to postulate a variety of we-subjects, as I have just made explicit. But if second-personal thought gives rise to second-personal reasons through the subject’s identifying with not one but a plethora of we-subjects, then the second-person standpoint is not one thing making one change to the structure of the subject’s agency: it is many. Thus, if second-personality is explained in terms of a subject’s identification with a variety of we-subjects, this explanation cannot rescue the structural approach – indeed, this explanation reinforces the thought that the structural approach is too narrow to accommodate the range of normative phenomena that appear to be second-personal reasons.

On the other hand, the other elaboration of the structural approach suggested a moment ago identifies the structural change to the subject’s agency with an openness to being affected by the other person. There is something promising about this as a characterisation as it does seem to encompass all putative second-personal reasons. However, like the Darwall-inspired elaboration, this view too lacks the conceptual resources to explain the different senses in which, and the different extents to which reasons can appear to be second-personal. This approach will have it that the reason to give to the Red Cross, and the reason to keep a promise to a friend are second-personal in the same sense and to the same extent – namely, that both involve an openness to being affected by the other person (the imagined representative of the moral community in the former case, the particular promisee in the latter).

Considering the deficiencies of the structural approach has cast light on second-personal reasons. The array of counterexamples to the structural approach have suggested that second-personal reasons vary on at least two dimensions: some reasons seem to be second-personal in different senses than others, and some reasons even seem to be second-personal to different degrees than others. In the remainder of this paper, I will propose an alternative approach to understanding the practical significance of the second-person relation that does justice to the heterogenous variety of second-personal reasons.

6 The Substantive Approach: A Schema

The structural approach regards the I-you stance as a fundamentally different paradigm for action than first-personal practical thought. As a result, the structural approach regards second-personal reasons as homogenous: they are all second-personal in the same decisive way that any of them are, as reasons brought about by the structural impact of the second-person relation (whatever that might be thought to be). But second-personal reasons are not homogenous, they are decidedly heterogenous. If we abandon the structural approach, we can acknowledge this heterogeneity. I suggest we call the alternative the substantive approach. On the substantive approach, relating second-personally to others affects our practical reasoning. It does this not by altering the form of our practical deliberations, but their content. That is, the substantive approach says that relating second-personally to others can ground, and reveal, reasons for actions that would not otherwise have existed or been available to the subject.

In support of the substantive approach, I will offer a summary schema of different senses in which a reason can be second-personal. These are not all mutually exclusive, so some reasons can be second-personal in multiple senses. One might say that in this respect, those reasons are more second-personal than others. In each case I will try to animate the intuition that there really can be reasons that are second-personal in that sense. If that much is granted, then the substantive approach will have some strong evidence in its favour, against the structural approach (Table 1).

Table 1 Variety of second-personal reasons, with examples
Full size table

A category of reasons that are unequivocally second-personal are those that emerge through what David Enoch (2011) calls ‘robust reason-giving’. These are reasons such as those presented in commands and requests. These are reasons that are intentionally created by one person (the addresser), for another person (their addressee), through the act of address. One can only receive a command or a request, or make a promise, by regarding another person as an addresser or an addressee, which just is relating to them second-personally. Thus, the reasons that emerge through these transactions are always second-personal reasons. These reasons are grouped in the column titled ‘Specific acts of address’ in the table.

The table has another column, to the right, titled ‘No specific act of address’, because there can be reasons that depend on second-person relations, but which are not created through address in the way that commands, promises and requests are. These are reasons that emerge through the course of interpersonal relationships where those relationships could not give rise to those reasons were it not for the second-personal interactions between parties. In fact, a great portion of our social lives are spent relating second-personally to others, and all sorts of relationships could not be the relationships that they are without this second-personal interaction. This includes familial relationships, friendship, many common collegial and neighbourly relationships, and so on. These relationships are often the sources of reasons for us to do things for others, even without being asked, or commanded, and without having promised. Because those reasons derive from relationships that are essentially second-personal, they too are essentially second-personal, even though they do not require a specific act of address in order to come about.

The two columns thus distinguish two senses in which a reason can be second-personal. Within each column, there are two rows which distinguish a further pair of ways in which reasons can be second-personal. On the one hand, there are reasons that would exist independently of the second-person relation, except they come to be known by the subject through such a relation. Second-person relations are playing a merely epistemic role in making these reasons possible. On the other hand, there are reasons that depend on second-person relations but not just in an epistemic way. Thus, in the case of these metaphysically second-personal reasons, what makes them second-personal is something about their grounds as reasons. These reasons have their second-personal character irrespective of how anyone might come to know about them.

As the table illustrates, there can be four categories of second-personal reason when these two distinctions are applied. Some second-personal reasons are dependent on second-person relations in a merely epistemic way, beyond the fact that they are brought about through an instance of second-personal address. These are located in the top-left square of the table. An example here is the reason that emerges through a verbal agreement. Imagine that you are cooperating with a stranger in an emergency in carrying a further, wounded stranger to safety. They say to you, ‘you lift the legs, I’ll take the arms’. This gives you a reason that you did not have before to lift the legs of the wounded person. Because the reason that you now have was brought about through an act of address, it is second-personal. But it does not derive its force from any further necessarily second-personal relationship that you had with the person with whom you are cooperating. Indeed, we may suppose that until the utterance of this suggestion about how to manage the carrying, neither of you had ever related second-personally to the other. The reason you have to lift the legs was specified by the stranger’s suggestion, but it derives its force from the fact that you should help the wounded person in the most effective way possible. In this sense, the reason is only second-personal insofar as it comes about through the act of address.

Staying in the upper row, there is another category, in the right-hand column, of reasons that are neither brought about through second-personal address, nor metaphysically dependent on second-personal relations. An example here are reasons that become apparent through second-personal interaction. Though this is not mapped in the table, we could also further distinguish here between, on the one hand, reasons that in fact are revealed through second-personal interaction but could have been known some other way – such as when someone tells you that the bus departs in five minutes – and on the other hand, reasons that could only have been revealed through second-person interaction, despite not being metaphysically dependent on them. This latter group might be harder to see, but the example in the chart helps to illustrate them. Suppose that you have heard nice things through a mutual friend about someone else – call her Wynn. Specifically, you have heard that Wynn is good to be around: she is good humoured, a good listener, and generally easy company. Despite the fact that you have heard this on good authority, we might well think that you can come to know it in a much fuller sense by getting to know Wynn herself and becoming acquainted with these qualities of hers. Of course, when you get to know Wynn and a relationship develops between you, it is possible that the relationship will ground reasons to value her. But somewhat separable from that, the fact that she is good company is a reason to like or admire her: to value her in some way. This reason does not depend in any metaphysical way on your having a relationship with Wynn.Footnote 9 It is second-personal in that only relating second-personally to Wynn could provide you with the appropriate kind of epistemic access to recognise the full extent of her worthiness for your admiration.

So much for reasons that are not metaphysically second-personal. In the lower row of the table are reasons that are. In the lower cell in the left-hand column, are reasons that are not only brought about through an act of address, but which draw their normative force from some further relationship, where that relationship must be, inter alia, a second-personal one. An example here is a request from a loved one. As (Lewis, 2018) has argued, a paradigmatic act of requesting appeals to a discretionary valuing attitude that the addressee of the request places in the addresser. One might think that in the course of a loving relationship such as a friendship, one’s attitude of valuing the other is grounded, in part, in the fact that you have a particular history of relating to one another.Footnote 10 As such, it is possible for a loved-one to make a request of you that is second-personal both in the sense that it is brought about through an act of second-personal address, and in the further sense that it appeals to a relationship that is, as we might say, deeply second-personal.

Sometimes, though, your friend does not need to ask you to act in their interest in order for you to have reason to do so. There are reasons that are borne of love or care, where such valuing attitudes are grounded – at least in part – by the history one has in relating with another as friends and so second-personally. These reasons are not merely epistemically but metaphysically second-personal, and they do not require any specific act of address in order to arise. They therefore occupy the bottom-right cell of the table.

This schema of second-personal reasons makes it clear that they are a heterogenous class. That alone speaks against the structural approach. On this, substantive approach to understanding why second-person relations make a difference to what we ought to do, we are able to recognise a variety of kinds of reasons that emerge through our relating with others as an I to a you.

7 Conclusion

Encountering other people and taking up the distinctive I-you stance towards them makes a profound difference to our lives as practical agents. Recognising this and trying to understand what that difference might consist in has been the project of a rich tradition within the last two centuries of Western philosophy. The predominant approach within that tradition has pursued the idea that the very structure of practical reasoning is reconfigured when we think of ourselves as agents in relation to some other in the position of a you. I have not found fault in the principal arguments that are associated with the structural approach, such as Darwall’s Fichtean argument to the conclusion that moral obligations are second-personal reasons. For all I have said, that may be true. My point has been, rather, that even that ambitious argument looks unlikely to account for the full breadth of the practical significance of the second-person relation – that significance does not seem to be explicable in terms of a single transformation of the structure of practical reasoning. The structural approach has obscured the diversity of types of reasons that emerge in second-person relations. If we abandon that approach, then we can recognise that diversity.Footnote 11