Brandom on Hegel and the Retrospective Determination of Intention

Abstract

In this paper I examine Brandom's account of Hegel's claim that the content of an intention can only be determined retrospectively. While Brandom's account, given in Chapter 11 of A Spirit of Trust, sets a new standard for thinking about this topic, I argue that it is flawed in three important respects. First, Brandom is not able to make sense of a distinction that is central for Hegel, namely, between the consequences of an action that ought to have been foreseen by an acting agent, given the right of objectivity of the action, and unforeseeable consequences that are completely contingent. Second, Brandom incorrectly conceptualizes the disparity and unity that all actions display as temporally successive features of an action, rather than as speculatively identical features. Third, Brandom's account cannot make sense of cases of retrospective determination that involve self-deception, and this demonstrates that he misses something critical about Hegel's account of action, namely, that action is expressive of the logic of essence.

I. Introduction

In Chapter 11 of A Spirit of Trust Brandom examines Hegel's theory of action and agency, focusing on the claim that the content of an intention can only be determined retrospectively. Roughly, Hegel's claim is that the content of an agent's intention depends on the action that realizes it, its consequences, and on how that action and its consequences are interpreted by others through time. Given this dependence, an agent can only come to grasp the content of their intention retrospectively.

In Brandom's hands, the retrospective thesis plays a key role in his overall argument to establish that the movement of experience as traced by Hegel in the Phenomenology of Spirit is recollective.¹ In this paper I ignore, as far as I can, this larger context and focus on the cogency of Brandom's account of the retrospective thesis. This is worth doing because his account in my view sets a new standard for thinking about this topic. Instead of gesturing to an argument for why the content of an intention can only be determined retrospectively, as many commentators do, Brandom works the argument out in systematic detail.² He does so both through giving an interpretation of Hegel's thought and by calling upon independently developed theoretical tools. Two questions, therefore, can be asked about his account: is it true to what Hegel says in the Phenomenology of Spirit and the Philosophy of Right about action and intention; and is it theoretically cogent on its own terms? In this paper I try to address both of these questions.

In the first half of this paper I discuss materials necessary to grasp Brandom's account of the retrospective thesis, namely, the difference between the modern and ancient-heroic conceptions of action and agency (section two), the distinction between purpose and intention (section three), and lastly, the difference between the vulgar success and failure of an action and the unity and disparity that all actions necessarily display (section four). I then, in section five, examine the retrospective thesis itself, focusing on Brandom's use of anaphora to illuminate the process by which the content of an intention develops, and his use of the distinction between speaker reference and semantic reference to illuminate the fact that acting agents can be mistaken about the content of their intention. Lastly, in section six, I examine four cases of the retrospective thesis and argue that: 1) Brandom is not able to make sense of a distinction that is central for Hegel, namely, between consequences that ought to have been foreseen by an acting agent given the nature of the action and completely contingent unforeseeable consequences; 2) Brandom's account of the disparity and unity that all action displays is flawed because he thinks of them as temporally successive features of an action, rather than as speculatively identical features; and 3) Brandom's account cannot make sense of cases of retrospective determination that involve self-deception, and that this shows that he misses something critical about Hegel's account of action, namely, that action is expressive of the logic of essence.

II. Two conceptions of action and agency

On Brandom's view, Hegel's theory of action attempts to integrate two distinct conceptions of action and agency: a modern conception and an ancient-heroic conception. To provide context for the retrospective thesis, let me briefly lay out these two conceptions.

The modern conception of action and agency is based on the fact that modern subjects make a distinction between what in a deed is theirs—the aspects of it that they know and intend—and what in a deed is not theirs—its unforeseen and unintended consequences. In Hegel's language modern subjects assert a ‘right of knowledge’, which is the right to be held responsible ‘only for those aspects of its deed which it knew to be presupposed within its end, and which were present in its purpose’ (PR: §117, 144), and a ‘right of intention’, which says that ‘the universal quality of the action shall have being not only in itself, but shall be known by the agent and thus have been present all along in his subjective will’ (PR: §120, 148).³ For modern subjects, if certain consequences of one's deed are not known or intended then one is under no obligation to accept responsibility for them.

But there are two distinct senses of responsibility: a causal and a moral sense.⁴ One is responsible in the casual sense for all of the alterations in external existence brought about by one's doing. One is responsible for them in the sense that one is their condition, ground, or cause (see PR: §115A, 143). But, Hegel thinks, the ‘fact that I am responsible for something does not mean that the thing can be imputed to me’ (PR: §115A, 143). In other words, the fact that I am causally responsible for something does not mean that I am morally responsible for it—i.e., normatively accountable for it. ‘I am only what has reference to my freedom, and my will is responsible for a deed only in so far as I have knowledge of it. Oedipus, who unwittingly killed his father, cannot be accused of parricide’ (PR: §117A, 144). While killing his father, parricide, is something that Oedipus did, rather than something that merely happened, he is not, according to the modern conception, morally responsible for it because he did not know or intend to commit parricide. Oedipus is normatively accountable for killing, which he did know and intend, but not for parricide.

The ancient-heroic conception, in contrast, does not distinguish in a deed between what is one's own and what is not. As a result, it imputes all of a deed's consequences to the acting agent—one is morally and not just causally responsible for whatever one brings about through one's deed, regardless of what one knows and intends. Oedipus's deed is not only the condition, ground, and cause of parricide, it is morally imputable to him—he is normatively accountable for parricide. As Hegel puts it, the ‘heroic self-consciousness […] has not yet progressed from its unalloyed simplicity to reflect on the distinction between deed [Tat] and action [Handlung], between the external event and the purpose and knowledge of the circumstances, or to analyse the consequences minutely, but accepts responsibility for the deed in its entirety’ (PR: §118R, 146).

In this passage Hegel makes a crucial distinction between deed (Tat) and action (Handlung). This is not a distinction between two different things but between two different interpretive perspectives on the same thing, the same act-event. As Brandom puts it:

Hegel's ‘Tat’ refers to the deed done, with all of its accordioned descriptions, and […] ‘Handlung’ is that same deed as the agent's doing—that is as specifiable by the restricted set of descriptions under which it is intentional, and hence something done at all. (2019: 389)

So, on the one side, a deed is ‘what is done as an actual event’ (Brandom 2019: 386), i.e., the bodily doing and all of the alterations that it brings about in the world—its consequences. The content of a deed is specified by all of the true descriptions of the consequences that this doing brings about through time—regardless of what the acting agent meant to bring about. In Brandom's language, the content of a deed is specified through consequential descriptions. These descriptions articulate what the performance is in-itself:

The truth of the performance, what it is in itself, is expressed in all of the descriptions of what is actually achieved, all the specifications of the content in terms of its consequences. These descriptions are available in principle to anyone in the community to recognize the performance under or to characterize its content […] The consequential descriptions specify what the action is for others, and for the agent qua other. (Brandom 2019: 394)

On the other side, the side of action (Handlung), the act-event is specified by a much more limited set of descriptions, intentional descriptions. Brandom gives a Davidsonian gloss here: an event is an action if it is intentional under some description, and a description is intentional if it specifies the action as the conclusion of an agent's practical reasoning. ‘What qualifies an occurrence as an action—something an agent is responsible for—is the existence of a privileged subset of specifications. And they are privileged precisely by their normative relation to the agent. Specifically, they are justified by practical reasons whose normative force or validity the agent acknowledged’ (Brandom 2019: 389). This subset of descriptions specifies not what the act-event is in-itself but what it is for the acting agent—it is that for which they acknowledge responsibility. This is the action (Handlung) in Hegel's technical sense.

The modern conception of action and agency takes it that an agent's intentional descriptions ‘have special authority not shared by those who merely observe the results of that endorsement’ (Brandom 2019: 395). The agent has authority about what they mean to do given their process of practical reasoning, and they are responsible in the sense of normatively accountable only for what in their action results from that process. While what the agent does goes beyond the subset of intentional descriptions that they apply to their action, they are accountable only for what knowingly results from that subset. The ancient-heroic conception, in contrast, does not grant the individual agent this authority. Because it does not recognize the distinction between deed and action, it takes it that one is responsible in the sense of normatively accountable for whatever one does, whether or not it is the result of one's process of practical reasoning. Oedipus is answerable for parricide, even though there is no intentional description of his deed that includes this consequence.

According to Brandom, what Hegel aims to do in his theory of action is to integrate the modern and ancient-heroic conceptions of action and agency into a single picture, one in which the ‘authority over what happens that is constitutive of agency can be genuine without being total’ (Brandom 2019: 375). We must see subjects as having genuine authority over their doings, for their endorsement of ‘a purpose is an essential element in a resulting performance being something done rather than just something that happens’ (Brandom 2019: 455). The agent's take on their doing is what distinguishes an action from a mere event. But this must not lead us to accept the illusion characteristic of the modern conception that authority is total, the illusion that one can determine what it is that one does independently of consequences and how other agents take such consequences. To leave this illusion behind we need not, however, return directly to the ancient-heroic conception. Rather, we have to rehabilitate, on a new and higher plane, its commitment to ‘“accepting responsibility for the deed in its entirety”, including those features of the doing that stem from its contingent, unforeseen, indeed unforeseeable consequences’ (Brandom 2019: 465). For Brandom, this rehabilitation will happen in a period after modernity, in post-modernity. Here, each subject will come to see their independence, their authority over their thought and action, as bound up with their dependence on, their responsibility to, others. In this post-modern form of ethical life we will

understand ourselves in such a way that we all take responsibility for what each of us does, and we each take responsibility for what all of us do. Although the individual is still understood to play an essential role—without which nothing would be done—the recognitive community is understood to play an equally essential role in the individual's capacity to do anything. In a real sense, to be the doing of an individual agent, each action must also be the doing of all. (Brandom 2019: 465)

But for this understanding to be possible we need to cash out the exact way in which the recognitive community plays an essential role in the individual's capacity to do anything. And as we shall see, this requires understanding the retrospective determination of intention.

III. Purpose and intention

Before getting to the retrospective thesis, however, we need two more pieces of Brandom's apparatus on the table: his account of Hegel's distinction, made in the Philosophy of Right, between purpose (Vorsatz) and intention (Absicht), and his account of the difference between the vulgar success and failure of an action, and the unity and disparity that all action displays. I take these up in the next two sections.

To put it simply, the purpose of an action is the ‘subjective content of the action (what one decides to do)’ (Brandom 2019: 402). It is what one means to do, the end or goal to be brought about as envisaged by the acting agent. When one specifies the content of an action through intentional descriptions, one specifies its purpose. Indeed, ‘[w]hat makes what is done (the deed) mine—that is an action, rather than just something that happens—is its relation to a purpose’ (Brandom 2019: 287). The intention with which the action is done, in contrast, concerns ‘the universal, manifold (articulated) content of the action as planned […] (which includes how one decided to do it)’ (Brandom 2019: 402). Such content endures through the different sets of connections that make up the action as it unfolds through time. So, for example, my purpose in acting may be to burn down the blue house, but my intention in so acting is to commit arson. In intending to commit arson, I do not just intend to make true the state of the world as specified by its envisaged end, the blue house having burned down, I also intend the plan to burn it down, which includes the steps needed to complete this plan, for example, obtaining an accelerant, getting to the house, spreading the accelerant, lighting a match, throwing the match on the accelerant, etc. But while I intend these steps, they can and often do change in carrying out one's intention. For instance, when Home Depot is out of accelerant, I change my plan and go to Lowes. But my intention has not changed. ‘The plan changes but the intention endures’ (Brandom 2019: 412). It endures because ‘intention functions as a norm that governs the process of realizing a purpose’ (Brandom 2019: 405). The content of the intention with which an action is done is therefore equivalent neither to its purpose nor any given plan to bring about that purpose, though one's intention includes one's purpose and the plan that one turns out to utilize in bringing about that purpose. ‘What is intended is the whole structure (the universal), not just the end or purpose aimed at, nor (at the other end of the planned process) the immediate initial means adopted’ (Brandom 2019: 402).

But how far does the content of an intention extend with respect to the consequences of an action? If, for instance, the red house next door to the blue house that I intended to burn down itself burns down—is that also part of my intention? For Hegel, the answer is yes because, given what he calls the ‘the right of the objectivity of the action to assert itself as known and willed by the subject as a thinking agent’ (PR: §120, 148), I, as a thinking agent, ought to have foreseen these consequences, given the natural propensity of fire to spread. Because one is a thinking agent who has insight into the likely connections between events, one cannot disown responsibility for this consequence. As Hegel puts it: ‘An old proverb rightly says, “The stone belongs to the devil when it leaves the hand that threw it”. By acting, I expose myself to misfortune, which accordingly has a right over me and is an existence of my own volition’ (PR: §119A, 148). These foreseeable consequences were implicitly part of the content of one's intention.

Brandom agrees with this conclusion, but he interprets it in the most maximal way possible:

[C]orresponding (at least roughly) to the Tat/Handlung distinction in Hegel's account is an Absicht/Vorsatz distinction. The content of the feature of an action that Hegel calls its ‘purpose’ need not extend to everything that the developed deed contains, while the content of the feature of an action that Hegel calls its ‘intention’ does extend to everything that the developed deed expressing it contains. (Brandom 2019: 387)

His interpretation is maximal because if the content of an intention extends to everything the fully developed deed contains, then its content will include its purpose, the plan to bring about that purpose (including its steps), the envisaged consequences of the action, the consequences that ought to have been envisaged (given the right of objectivity of the action), and the contingent consequences of the action that go beyond what is possible to envisage. We shall examine in section six whether this extension of the content of an intention to include contingent consequences that are not foreseeable in any way accords with Hegel's view.

On Brandom's view, an action's complete content is the result of both the agent's take on the action and on how it and its consequences are taken up by the recognitive community—in other words, on both its intentional and consequential descriptions:

The content is what is both acknowledged by the agent and attributed by the community: the product of a process of reciprocal specific recognition. The content of my action accordingly does not depend on me alone. It is not just what I take it or make it to be, but depends as well on its determinate acknowledgement by others who attribute to me responsibility for the performance specified in ways that go beyond those in terms of which I made it mine.⁵ (Brandom 2019: 396)

Because a deed can be specified by a subset of descriptions that the acting agent has the authority to issue, and for which they take moral responsibility, it is an intentional action rather than a mere happening. But what one does in doing an intentional action goes beyond what one means to do, and one is responsible for this as well, because, from the retrospective point of view of the developed deed, others (including potentially oneself as ‘other’) can discern an intention that goes beyond the purpose:

The distinction among features of the deed that is induced by the purpose is what determines the deed as the agent's doing, in the normative sense of being something that agent is responsible for. What the agent thereby becomes responsible for (doing) is the whole deed (what is done). And that fully developed deed reveals an intention that extends beyond what is merely ‘meant’ or purposed. (Brandom 2019: 387)

IV. Success and failure vs. unity and disparity

We need to put one last piece of theoretical apparatus on the table. For Brandom, one of the main payoffs of the distinction between Vorsatz and Absicht is that it gives us a way to distinguish the ‘vulgar’ success and failure of an action from the unity and disparity that he argues all actions necessarily display.

In the vulgar sense, an action succeeds when what one means to do and what one in fact does correspond to one another. But there is of course an equivocation in the notion of ‘what one in fact does’. Take a successful action: opening a window. My aim is to open the window. I get up from my chair, take a few steps, and open it. But in doing so I also change the chemical composition of the room, I bring about a slight disturbance in the air outside the window, and I seem slightly rude to my host for opening their window without asking, etc. All of these happenings brought about by my doing can be picked out by consequential descriptions, but there is a difference between them in so far as opening the window is my end or purpose, while the other consequences are not. If my end or purpose is among the action's true consequential descriptions then the action is successful in the vulgar sense, and if not then it fails:

An action succeeds in this sense if the consequential descriptions that are true of it include the purpose whose achievement is the endorsed end in the service of which all the other elements of the intention-plan function as means. An action fails in this sense if, although some things are done intentionally, i.e., as part of the plan, the purpose is not achieved, because the means adopted do not have the consequences envisaged. (Brandom 2019: 402)

This vulgar success or failure is distinct from the unity and disparity that Brandom argues every action necessarily displays, whether successful or not. While the vulgar success or failure of an action depends on whether an actor's take on what they are doing, articulated by its purpose, is included among the consequential descriptions of the action, unity and disparity are ontological features of the action that hold regardless of an acting agent's take on what they are doing. An action's unity concerns the fact that ‘every action (“globally”), as an action […] simply translates something inner or implicit into something outer or explicit, hence exhibiting the unity of action and the identity of content in two different forms’ (Brandom 2019: 382). In every action, whether successful or not, something inner is revealed in the outer, and there is ultimately a kind of unity or identity in its content. But every action also displays disparity because action takes place ‘in the realm of existence’ (PR: §114A, 141–42), i.e., in a physical and social world that the agent does not control. As such, the action is ‘exposed to external forces which attach to it things quite different than what it is for itself, and impel it on into remote and alien consequences’ (PR: §118, 145). Because this is so every action necessarily ‘involves an actual disparity between purpose and achievement’ (Brandom 2019: 382). Regardless of whether or not one's purpose is included among an action's consequential descriptions, such descriptions run beyond what one meant to do.

The central challenge in thinking about action, Brandom argues, is bringing unity and disparity ‘together into an intelligible whole’ (2019: 376), developing a picture of action in which they can be seen ‘as two sides of one coin: as reciprocally sense-dependent concepts playing essential roles in the concept of intentional action’ (Brandom 2019: 382). This is a difficult challenge because, as Hegel points out, action's disparity leads the acting agent to ‘become a riddle to itself’ in so far as ‘the consequences of its deed are for it not the deeds themselves’ (PS: ¶365, 220; quoted in Brandom 2019: 375, 735). In other words, the agent cannot see the deed's ‘remote and alien’ consequences as theirs, cannot see them as expressive of something they have done. But if this is so then there is no way for an agent to see the content of their inner intention as identical with the content of their outer deed.

For Brandom, this challenge ultimately concerns the content of an intentional doing: how can such a doing be the expression of a determinate content that nevertheless becomes more determinate through the doing?

We must understand how the process and practice that are the exercise of intentional agency are intelligible both as mere expression, revelation, and translation from subjective to objective form of already fully determinate contents and simultaneously as the means by which initially less determinate contents become more determinate: the process of determining conceptual contents. The former perspective is that of the unity of action and the identity of content […] and the latter is that of the disparity of action and the difference between the content subjectively intended and the content objectively achieved (in every action, whether it succeeds or fails in the ordinary sense). (Brandom 2019: 383)

As we are about to see, the only way to understand how action can be the unity of a fully determinate content in two different forms, subjective and objective, while also being a process of determining content, is to view action retrospectively. From this point of view we will be able to see that the process of determining a content through the actualization of a prior-intention in action and through the interpretation of the action by others, was always going to turn out to express the unity of a determinate content.

V. The retrospective thesis

We can now turn to the retrospective thesis itself. The retrospective thesis can be understood in two ways, ontologically and epistemologically. On my reading, Brandom accepts both understandings, but he thinks that the ontological version of the retrospective thesis is primary, leading to the epistemological.

From the ontological point of view, the retrospective thesis concerns the nature of the content of an intention, not our awareness of this content. The content of an intention depends not just on one's intentional descriptions of the action but on what one has turned out to do, which depends on the circumstances in which it was done and the consequences to which it leads. These ‘contingencies to which the process of trying to realize a purpose is subject are somehow to be understood as features of the content that are retrospectively discernible as always already having been implicit in the intention’ (Brandom 2019: 404). Many of these contingencies are not ‘objectively present’, rather their being part of the content of a doing depends on the ongoing consequential descriptions of the community. So the content of an intention not only depends on what is in fact done but also on the interpretation of that doing by others.

It is important to see, however, that Brandom is not endorsing the radical thesis discussed by Laitinen, which claims that an intention ‘comes to existence only retrospectively, and is dependent on the reaction of others’ (Laitinen 2004: 59). In other words, he is not claiming that an intention and its content come into being through the consequential descriptions of others. Acting subjects have prior-intentions, but they lack determinacy. Prior-intentions gain determinacy through being realized in a doing, and through consequential descriptions of it.⁶ But from a retrospective point of view we can discern that that determinacy was always already implicit in the prior intention. This more determinate content, which is the actualization of a prior indeterminate content, is discernible as always already having been implicit in the intention. We shall come back to the exact nature of this claim below.

The ontological thesis leads to an epistemological thesis concerning the access that a subject has to their intentions. On the standard view, an intention is a discrete mental state (or a combination of states, a belief-desire pair) that causes bodily action. This state is accessible through introspection, a kind of reflective witnessing of the mental states that one is having. But for Brandom, an indeterminate intention is a norm or commitment that is to be actualized in action, not a mental state to be witnessed, and this norm guides action rather than causes it. But given that the content of this norm is ultimately ‘settled by what is true of the actual external performance that expresses it, then it is epistemically available, even to the agent, only retrospectively’ (Brandom 2019: 384). In other words, given that the content of an intention gains determinacy through being expressed, externalized, and interpreted by others we cannot know its content until after the deed is done. As Hegel famously puts it:

Consciousness must act merely in order that what it is in itself may become explicit for it […] Accordingly, an individual cannot know what he is until he has made himself a reality through action. (PS: ¶401, 240)

Ethical self-consciousness now learns from its deed the developed nature of what it actually did. (PS: ¶469, 283)

But could one not say that one's prior intention is immune to misidentification, regardless of the deed that realizes it and its consequences? For it does not seem possible to think that I might be mistaken in saying ‘I intend to lift my arm’ because, although I do know of someone who intends to lift their arm, that person is not me.⁷ I settle whether I am committed to lifting my arm, and it is in settling this that I have a kind of practical knowledge of my intention.⁸ Hegel's answer to this, if he were to develop one, would be to say that while the purpose of the action may be immune to misidentification, one's intention is not. While an agent has authority over the content of their purpose in so far as they settle what their purpose is, they do not have similar authority over the content of their intention. One might think that they settle the content of their intention, but they are simply wrong about that.

To make this more than a bare assertion, Brandom offers us two arguments. First, he calls upon his account of anaphora to explain the development of the content of an intention through time, and second, he calls upon the difference between speaker reference and semantic reference to make sense of the fact that agents can be mistaken about the content of their intentions.

Anaphora is the process by which later pronouns pick up their meaning from unrepeatable demonstratives. In A Spirit of Trust, Brandom uses anaphora to explain how, in sense-certainty, unrepeatable sensory acts can nonetheless have a content that endures through such acts. When I say ‘this (pointing to a chair) is elegant, but it is also inexpensive’ the pronoun ‘it’ picks up its referent from the prior demonstrative. So the unrepeatable act of sensing the chair, captured by my utterance of the demonstrative ‘this’, determines the cognitive significance of this pronoun and an indefinite number of future pronouns that refer back to this pronoun. When we move from perception to action, from so-called entry moves to exit moves, the movement of anaphora reverses, meaning that an earlier phrase picks up some of its content from later demonstratives that give it specificity. Take the prior-intention ‘I will raise my hand in 1 minute’. This intention develops into to another prior-intention, ‘I will raise my hand in 30 seconds’, and another, ‘I will raise my hand in 15 seconds’, until it is realized by ‘I raise my hand now’. All intentions have this movement, from a prior-intention whose content is relatively indeterminate to a content that becomes determinate through doing this now. Here, ‘an earlier description of what is to be done can be thought of as inheriting some of its content from the later demonstrative specification of what is done, on which it is understood to be anaphorically dependent’ (Brandom 2019: 406). From the prospective point of view, the agent can only form intentions in general terms that may or may not pick out the action specified by the final demonstrative. But from the retrospective point of view this ‘final demonstrative picks out what we were all along referring to’ (Brandom 2019: 408). It is only in retrospect, therefore, that ‘we can tell what the actual content of the intention was, given the possibly unknown circumstances in which it was to be actualized’ (Brandom 2019: 408).

Let us now move to the issue of how an agent can be mistaken about the content of their intention. To make sense of this Brandom calls upon the distinction between speaker reference and semantic reference, which he then applies to practical contexts. Take Donnellan's famous example: I say ‘the man in the corner drinking champagne is an economist’, where the person I mean to be talking about is drinking ginger ale and, unbeknownst to me, there is a second man in the corner drinking champagne. My statement semantically refers to the second man, while I speaker refer to the man drinking ginger ale. There is here a difference between what I say and what I mean. From the semantic point of view, I ‘said that the man in the corner drinking champagne is an economist’, referring to the second man, while from the speaker point of view ‘I said of the one drinking ginger ale that he was an economist’ (Brandom 2019: 407). But given that these are two specifications of the same content, how can one be in a position to make this distinction?

The first thing to say is that the distinction can only be drawn from the ‘third-person point of view’ (Brandom 2019: 407). I myself cannot distinguish between what I say and what I mean because I do not have access to the information that would allow me to grasp what I am in fact saying (i.e., that I am talking about the second man). Drawing the distinction between semantic and speaker reference

requires adopting the perspective of someone else, who has different information than I do, someone who can attribute a different responsibility to me than I acknowledge, by linking my utterance anaphorically to other possible utterances of mine. But, from that third-person point of view, there are two ways to assess the commitment I have made, the responsibility I have undertaken by my claim. (Brandom 2019: 408)

The belief I acknowledge is that the man in the corner is drinking champagne. But an interpreter can know that that person is in fact drinking ginger ale, and so can distinguish, with respect to my commitments, between the commitment I take myself to have (what I mean) and the commitment I in fact have (what I in fact say). But of course, for an interpreter to make this distinction requires that they know the person I mean to be referring to, the one drinking ginger ale. And this requires that they are able to anaphorically link my statement ‘the man in the corner drinking champagne is an economist’ to statements like ‘that man’ or ‘the man I am looking at right now’. They must ‘understand my definite description as anaphorically dependent on one of those possible demonstrative specifications […] and so as inhering its content from that demonstrative’ (Brandom 2019: 407).

How does all of this bear on action? Brandom's point is that an agent is not, by themselves, in a position to distinguish between what they take themselves to intend—the ‘content subjectively intended’ (Brandom 2019: 383)—and what they in fact intend. They are not in a position to make this distinction because what they in fact intend to do, unlike what they take themselves to intend to do, depends on anaphoric links to later demonstrative specifications of what they have done, specifications that can only be described consequentially after the fact. Just as my claim that the man in the corner drinking champagne is an economist is in fact, though unbeknownst to me, a claiming of a man drinking ginger ale that he is an economist, my de dicto specification of the purpose or subjective intention to be made true by my action may be, unbeknownst to me, a de re intending of something I am not in a position to be aware of prior to, or even in, acting. This is because

whatever I actually did determines the content of my intention, under the actual circumstances in which I acted. Under the actual circumstances, having the purpose I did amounted to intending to do that—whatever I actually achieved. Intending to turn on the light by flipping the switch was under the actual circumstances in which I intended it, though unbeknownst to me, intending of a particular burglar alerting that I do that. (Brandom 2019: 408)

Regardless of what I take myself to be doing, the realization of the actual deed—the res—determines what I, in fact, intend in so far as the later demonstratives at play in the action realize the content of my prior-intention. Unbeknownst to me, I, in fact, intended to do that, whether my purpose can be found in the action's consequential descriptions or not.

Because I, in fact, intended to do that, I can, through the consequential descriptions of others, come to be aware that I did that. I can retrospectively come to see what I have done as mine—even though prospectively this was not the case. This retrospective learning process is what makes it possible to solve the central challenge in thinking about action, namely, to see the unity and disparity of action as two parts of ‘an intelligible whole’. For now the contingent consequences of one's doing are not a riddle for the acting agent, but are seen by them as part of the developing content of their deed. Retrospectively, I can see that doing that was part of a process of determining the content of my prior-intention, and that this process has resulted in a determinate content shared by intention and developed deed. I grasp retrospectively that what I intended and what I did are the same and that this sameness or identity was arrived at through a process of development.

VI. Evaluation and critique of Brandom's retrospective thesis

To evaluate Brandom's account I would like to examine four cases of retrospective determination, the first three of which are Brandom's.

1. 1. I intend to get to the other side of the river by the end of the day, but to do so I must swim across the river. Under the actual circumstances my intention to get to the other side of the river was, unbeknownst to me, intending of a river swimming that I do that.⁹

2. 2. I intend to turn on the lights by flipping the switch, but in doing so I alert the burglars (see Brandom 2019: 408). Under the actual circumstances my intention to turn on the lights was, unbeknownst to me, intending of a particular burglar alerting that I do that.

3. 3. I intend to write a good poem, but in doing so I produce a bad poem not fit for publication (see Brandom 2019: 385). Under the actual circumstances intending to write a good poem was, unbeknownst to me, intending of a not-fit-for-publication poem writing that I do that.

4. 4. I intend to be kind to someone but in doing so I am in fact subtly mocking. Under the actual circumstances, intending to be kind was, unbeknownst to me, intending of a subtle-mocking that I do that.

These cases are quite different in nature. The first concerns the means to the end of an action, the second the unforeseen consequences of an action, the third a work produced by a complex action, and the fourth concerns the social meaning of an action. I do not think that Brandom's account can handle all of these cases equally well. Let us examine them in order.

Cases One and Two

Cases one and two, which Brandom treats together, are both cases in which the content of a general prior intention gets specified through a later demonstrative. I intend to ‘get to the other side of the river’ and ‘to turn on the light’, but in doing the actual deed I do this now, i.e., swim across the river and alert the burglars respectively. What I want to argue is that there is critical disanalogy between these cases that casts doubt on whether the second case is a genuine case of retrospective determination.

When one acts on one's general intention to get to the other side of the river one finds that one must do this now, where this now refers to swimming across the river. This means to one's end was not envisaged in advance, but in looking back on the developed deed one grasps that it was always in the cards that they swim across the river given the content of their prior-intention and the circumstances in which it was going to be actualized. If one is to fulfil one's subjective intention (which includes one's purpose), one must swim across the river. In grasping this, one retrospectively makes explicit to oneself that this means was implicit in one's prior-intention. Because one sees it as implicit in one's prior intention, one sees oneself as responsible for this means, not just in the causal sense but in the moral sense. In intending a means-end complex one undertakes responsibility to will means that necessarily conduce to that end. One is normatively accountable for this means in so far as it follows rationally from one's intention.

But how is this supposed to work in the case of alerting the burglars? I have a general intention to turn on the lights and the content of this intention gets specified by doing this now. But what does this now refer to here? If it refers to flipping the switch then there is no problem, as that is a means to one's end. But Brandom also thinks this now refers to my alerting the burglars. From the point of view of the consequential descriptions of others, I am also doing this now. But was it always in the cards that I would alert the burglars, given the general nature of my prior-intention? Unlike swimming across the river when there is no other way of achieving one's end, this consequence is related to what I have done in a completely contingent way. This consequence was not implicit in the intention to turn on the lights as it in no way realizes that intention. It is rather simply a contingent consequence of it. Not only did I not mean to alert the burglars, I did not intend it. While I should recognize myself as responsible in the causal sense for this consequence, there is no reason for me to accept normative accountability for it.

Now I keep using the first person here, asking about what I recognize as my responsibility. But if the content of an intention is specified both by intentional and consequential descriptions then I cannot privilege this point of view. I may be normatively accountable for elements of my deed that outrun what I take myself to be accountable for—something that can only be identified by others. But would others say in light of their consequential descriptions of the deed that I am normatively accountable for alerting the burglars and that I ought to accept moral responsibility for it? I do not think an interpreter would say that. It would be far more natural for them to say that this is a case where what I did outruns what I intended to do. In this case, it is natural to say that no one is morally responsible for the alerting.

Brandom would respond by saying that in a post-modern form of ethical life, ascriptions of responsibility would not just apply at the individual level. When we enter the age of trust by becoming self-conscious about the identity of independence and dependence, authority and responsibility, we come to grasp that a deed is ‘not done by the agent alone’, but that it is ‘also done in a different, though equally constitutive sense by the agent's community. All are responsible for the doing of each, and each for the doing of all’ (Brandom 2019: 734). While I may not be morally responsible for alerting the burglars, it is not merely a contingent event for which no ascription of responsibility can be made. For in a post-modern form of ethical life we would, for example, all take responsibility for the fact that there are burglars to alert, would take responsibility for the fact that the social order contains crime.¹⁰ But while there is a very important insight here, this ascription of responsibility takes place at the wrong level to undermine my argument: it concerns the conditions for an event to happen, not the event itself. In other words, this collective ascription of responsibility does not concern the alerting of the burglars itself but the social conditions in which there are burglars to alert. With respect to the alerting itself, it remains contingent, entailing that even within a post-modern form of ethical life no one ought to be held responsible for it. Not all contingency can be incorporated into an intention, becoming—from the retrospective point of view—necessary.

The fact that Brandom cannot recognize the difference between cases one and two shows that his account of intention is too strong. It is too strong because it does not do justice to a distinction that in my view Hegel never gives up, namely: between 1) the foreseeable consequences of action and the consequences that are not foreseen but ought to have been foreseen, given the right of the objectivity of the action over the agent as a thinking agent, and 2) the completely contingent and unforeseeable consequences of an action. Brandom's account of consequential descriptions is not fine grained enough to allow for this distinction. While Hegel significantly widens the narrow version of the modern conception, which keys consequential descriptions and therefore intentions solely to what agents can foresee in advance, he accepts, for actions that are normatively permissible, the idea that subjects are not normatively accountable for completely contingent and unforeseeable consequences.¹¹ Grasping that action is in general enmeshed in external existence is critical for modern subjects, as it tempers the illusion of mastery to which they are prone. But recognizing this general truth does not require denying that at the individual level the contingency of an outcome has an effect on our ascription of moral responsibility for that outcome.

Case Three

The third case, where a poet means to write a good poem but in fact writes a bad one not fit for publication, is perhaps most clearly illustrative of the retrospective thesis. In his discussion of it Brandom quotes an exceptionally rich passage from Hegel:

From what has now been said, we may learn what to think of a man who, when blamed for his shortcomings, or it may be, his discernable acts, appeals to the (professedly) excellent intentions and sentiments of the inner self he distinguishes therefrom. There certainly may be individual cases where the malice of outward circumstances frustrates well-meant designs, and disturbs the execution of the best-laid plans. But in general even here the essential unity between inward and outward is maintained. We are thus justified in saying that a man is what he does; and the lying vanity which consoles itself with the feeling of inward excellence may be confronted with the words of the Gospel: ‘By their fruits ye shall know them’. That grand saying applies primarily in a moral and religious aspect, but it also holds good in reference to performances in art and science […] if a daub of a painter, or a poetaster, soothe themselves by the conceit that their head is full of high ideals, their consolation is a poor one; and if they insist on being judged not by their actual works but by their projects, we may safely reject their pretentions as unfounded and unmeaning. (Quoted at Brandom 2019: 385)¹²

This passage outlines what is perhaps Hegel's prime reason for thinking that there is a unity or identity between inner and outer, implicit and explicit, namely: to overcome a view of action and agency in which agents can exculpate themselves from the consequences of their deeds by claiming that there is an inner self comprised of fine ‘intentions and sentiments’ that is distinct from the self displayed in their outer actions. In my view this passage is not primarily about the identity of token intentions and deeds, but about the relation between all such intentions and deeds to the character of the self who bears them. It is interested in what a subject is—their propensities, capacities, talents and skills—and how we can come to know them, i.e., through action. As Hegel puts it, ‘What the subject is, is the series of its actions. If these as a series are worthless productions then the subjectivity of volition is likewise worthless, and conversely, if the series of the individual's deeds are of a substantial nature, then so also is his inner will’ (PR: §124, 151). This allows for the fact that in individual cases there can be misalignments between the inner and the outer due to the ‘malice of outward circumstances’, or because someone does something ‘out of character’. Such cases are consistent with there being an overall unity or identity of inner and outer, not a strict identity, but a speculative identity, one that is arrived at through a reciprocal process in which inner and outer transform into one another. Outer doings are expressions of character, and character is reciprocally determined by these doings, in so far as they lead to the development of propensities to do those very things.

But of course the self's character is expressed in specific actions and deeds, so the question of the unity of token intentions and actions recurs. In the case of the painter and the poet, the question is how to make out the unity of their inner ‘projects’ and outer ‘works’. When the poet writes a bad poem they may try to exculpate themselves by saying that their inner project or intention was brilliant, but that their brilliant intention went awry in its execution. The background conception of action at play in this exculpation is that action is caused by an inner intention whose content is determinate and complete, but that sometimes action does not, for whatever reason, accurately translate that determinate content into an outer work. Something gets lost in translation. But this picture, according to Brandom, gets ‘the model of expression—making the implicit explicit’ (Brandom 2019: 385) completely wrong. It is wrong because there is no completely determinate poem bottled up in the poet's intention that can be identified prior to their coming to specify its content by writing ‘this now’. As such, the poem ‘cannot be understood as the botched execution of a fine aim or plan, but must be understood rather as showing exactly what its creator actually intended—however it might seem to the author’ (Brandom 2019: 385). Given that one's intention extends to all of the consequences of the developed deed, one wrote just the poem one intended to write, even if it does not seem that way to the author.

Because Brandom understands expression as a form of making determinate he is able to avoid a criticism that McDowell directs toward Pippin's similar use of the poet example. McDowell thinks it is an abuse of the concept of intention to say that the poet intended to write a bad poem. We can admit that the supposedly good inner poem is mythical, and that the only poem that is relevant is the bad one that is actually written, ‘without needing to abuse the concept of intention by saying his intention turns out to have been the intention to write that poem. It is true that he has intentionally written that poem—he has intentionally strung just those words together in just that order. But in realizing his intention he gave it that specificity; he did not find out what it was all along’ (McDowell 2009: 180). But Brandom does not think that we find out what the poet's intention turns out to have been all along, namely an intention to write a bad poem. It is rather in realizing one's intention, in giving it specificity through doing ‘this now’, that we find out what one's intention has turned out to be, which was to write the exact poem that was written, which has turned out to be bad.

But I think Brandom's account is problematic in another way. To see how, let me get more specific. Let us say that the poet's intention is to write a melancholic poem about loss. In giving this intention specificity through writing the poem they mean to bring about this emotion through writing about loss. They have a plan to bring this emotion about through writing the poem, a plan that alters through the writing of the poem, and they take into account, as far as possible, the consequences of the poem—take into account how it will affect readers. But despite all of this the poem turns out to express bitterness about loss rather than melancholy. Bitterness, not melancholy, is included amongst its consequential descriptions. The poem therefore fails in the vulgar sense. On McDowell's view, this kind of failure is the only kind of disparity that an action can display: the content of the poet's intention is one thing, the meaning of the work produced by giving that intention specificity another. It can happen that one genuinely intends to do something but, in the doing, something else is done than what one intended. There is a possibility for retrospective learning here, but what one learns about is the meaning of what one has done—one's actual work—not what one intended in making the work.

Brandom, in contrast, says that this failed action displays both unity and disparity. It is a failed action because melancholy is not among its consequential descriptions. It is a disparate action because the consequences of the poem, as specified by these consequential descriptions, necessarily go beyond what the poet could have foreseen. The way that poem is read by others, its effects on others, is unpredictable. But as we saw above, the poet can come to view these consequences as theirs, as the product of their doing, because they can retrospectively come to see that they intended to do that, i.e., come to see that they intended to write the exact poem they wrote. From the retrospective point of view, the poet's action therefore displays unity: the content of the poet's intention is identical with the meaning that the completed work turns out to have.

I agree with Brandom that this failed action displays both unity and disparity. But my account of unity and disparity is different than his, reflecting my view that unity and diversity must not be understood temporally, as successive features of a single deed's content, but speculatively. On my view, this action displays unity because the poem expresses something about the poet, an underlying bitterness of which they perhaps were not aware. The writing of the poem expresses an aspect of who the poet is.¹³ The action also displays disparity, but not only because the consequences of the work in the outer world go beyond what the poet can foresee. In addition to outer disparity action contains an inner disparity. Writing a poem is a kind of experiment, a process of stringing words together in ways that while informed by past practice has no exact precedent. It cannot be predicted beforehand how purpose, plan, poetic skill, emotion and material will interact in the production of the poem. There is a kind of contingency here that is not completely controllable by the poet. This contingency is not only the contingency of how the poem will be taken by others, it is the contingency internal to the process of producing the poem in the first place. It is because of this contingency that even good poets can write unsuccessful poems.

The poet can only know retrospectively whether their experiment ‘worked’, whether the poem in fact expresses melancholy about loss. In coming to know this, they come to know about the intention that is actually expressed in the work. So the poet acquires self-knowledge through the work. But because of the contingency at play within the action that produced the work the poem has a kind of life of its own, a kind of quasi-autonomy. As a result, the content of the poet's intention is not completely identical with the meaning that the work turns out to have. The work, we could say, both expresses and does not express the poet's intention, and in coming to retrospectively learn about the intention expressed in the work they also come to learn—due to the inner-disparity that action involves—that their work has meanings that go beyond their intention.

On Brandom's account, we can, once the process of determining conceptual content is done, retrospectively posit unity. From this retrospective point of view, action is the ‘mere expression, revelation, and translation from subjective to objective form of already fully determinate contents’ (Brandom 2019: 383).¹⁴ From this point of view, unity has a kind of priority in so far as we cannot posit unity until diversity, the developing of content, has been completed and overcome. But if we view unity and diversity speculatively then unity has no such priority. Unity must be continually forged out of the diversity, both inner and outer, that action displays, and diversity is always the diversity of a single process of translating inner into outer. These two features do not apply to action sequentially, first diversity (as the content develops) and then unity (the completed and determinate content), rather they are elements of a single process. I come back to the nature of this process below.

Case Four

In the fourth case I intend to be kind to someone but in doing so I am, unbeknownst to me, subtly mocking. This case seems to have a different logic than the other two cases of genuine retrospective determination. In cases one and three when I intend of something that I do that, the ‘that’ refers to a later phase of the deed that realizes and specifies the content of one's more general prior-intention. But in the fourth case what does the ‘that’ refer to? It refers to the subtle mocking: I did ‘that’. But does the ‘that’, the mocking, realize the prior-intention to be kind? One could say this, but I think it would be misleading, for it is not as if mocking specifies the intention to be kind. It would be far more natural to say that it is just the expression of a different intention, and that the prior-intention never was what the agent who formed it took it to be. In this case there is and always was a difference between what one really intended and what one took oneself, appeared to oneself, to have intended. What one really intended was to be subtly mocking, while one appeared to oneself to intend to be kind. There is here an essence-appearance distinction that applies to one's prior-intention.

It is clear that an agent cannot, by themselves, distinguish between what they really intend and what they appear to intend, as they, from the prospective point of view, only have access to appearances.¹⁵ For the agent to learn that they were subtly mocking when they intended to be kind requires input from others. But for others to attribute this meaning to an agent's behaviour requires that they are able to put it into context, i.e., the context of the agent's character. To grasp the meaning of an agent's outer doings others must not merely enumerate them but must see them as infused with inner purposes and intentions, the tenor of which stem from their character. That is what gives others an interpretive baseline to develop consequential descriptions of this action as involving subtle mockery rather than kindness. So here interpretation goes from inner to outer, from character to action. But this interpretation simultaneously goes the other way. Through an interpretation of the meaning of an agent's doings, an other is able to posit their character, which allows one, on specific occasions, to posit token manifestations of it, in this case, the intention to mock. Through coming to be aware of the results of this interpretive process, an acting agent can retrospectively realize that their prior-intention was not at the time what it appeared to them to be. The agent becomes aware that all along they were wrong about the nature of their intention, that it had a different content than they took it to have. The agent realizes, by looking at themselves through the viewpoint of the other, that they were self-deceived about what they intended.¹⁶

It is important to see that self-deception is a more-or less-concept. At one extreme, essence and appearance completely come apart. What one really intends and what it appears to one that one intends do not interact in any way. One is, as it were, completely caught in the appearances. This is a pathological case because it enacts within the subject a metaphysical mistake: turning the difference between appearance and essence into a permanent distinction. But usually appearance and essence do not completely come apart: what one appears to oneself to intend is informed by what one really intends (as specified by what actually happens), and what one really intends is informed by what one appeared to oneself to intend. To utilize Brandom's underdeveloped distinction between subjective and objective intentions, we could say that one's subjective intention to be kind is inflected by the intention to subtly mock, and the subtle mockery that one objectively intends retains to some degree the subjective intention to be kind.¹⁷ It is important to note this because it allows us to account for two cases that need to be distinguished: first, where someone subjectively intends to be kind and turns out to be subtly mocking; and second, where someone subjectively intends to be subtly mocking and is subtly mocking.¹⁸ Taking oneself to intend to be kind alters the character of one's subtle mocking, making it different than the subtle mocking that would have flown from a subjective prior intention whose content was simply to subtly mock. The subjective intention to be kind tinges the subtle mockery. In this case, the content of one's prior intention, while not what one took it to be, retains, to however small a degree, the intention to be kind.

Pointing out that there are cases of retrospective determination of this fourth type does not by itself undermine Brandom's account of cases one and three. But it does point to a significant limitation in Brandom's view. Brandom misses the fact that for Hegel action is—to put it in the language of the Science of Logic—expressive of the logic of essence and not the logic of being.¹⁹ For him, all cases of retrospective determination are based in an essence-appearance distinction. It is just that the essence-appearance distinction operates differently in cases one and three, as opposed to case four.

In cases one and three what an agent takes the content of their intention to be in doing the action inevitably turns out to be different from what, from the retrospective point of view, the content turns out to really be. Here, we retrospectively look at a single content at different temporal points in its development. The ‘essence’ of the action, we could say, is comprised by all the characterizations that pertain to the fully developed deed, and ‘appearance’ is any characterization prior to that full development.²⁰ In all cases an agent learns retrospectively through the consequential descriptions of others the full content of their intention. But in the fourth case these descriptions of the developed deed are merely the reflective basis upon which essence is posited—they do not articulate the essence itself. In this case, essence pertains to the prior intention, not to the totality of descriptions of the developed deed. Here, making the implicit explicit depends not simply on articulating what the deed turned out to be, but on an interpretation of that developed deed that leads interpreters, and potentially the acting agent themselves, to get beyond the intention that the acting agent takes to be embodied in the deed to the genuine intention that informs it.

It is critical to understand, however, that for Hegel the essence of essence is to appear, and that appearances are the appearance of essence.²¹ Essence is not something that stands behind reality—it is not found in a ‘world beyond’. Rather, for essence to be essence it must get into outer form, be expressed to one degree or another in existence. Appearance, in contrast, is not semblance, but the way that essence manifests itself in immediacy. Because this is so, the essence that ‘stands behind’ the intention that appears to the acting agent does not stay hidden, but is expressed or actualized in the very deed that is the basis of the interpretation that posits it in the first place. So the ascription of a distinction between appearance and essence to the acting agent's prior-intention can only come about through an interpretation of the developed deed. But, as we saw above, that interpretation requires having a handle on the agent's character, which is that through which the developed deed is interpreted. So here we have come back to the speculative identity discussed in case three: we determine essence (character and its manifestation in specific intentions) through appearances (outer doings), and we determine the meaning of those appearances (outer doings) by interpreting them through essence (the agent's character), and it is this interpretive process that gives up the basis to make ascriptions about an agent's prior-intentions.

It is this process that ultimately underlies the unity and diversity that all action necessarily displays. An action displays unity because this process reveals that it is expressive of an underlying essence that necessarily shines through it. It displays disparity because, while the essence of essence is to appear, the expression of the one in the other is never fully transparent and complete. The outer never completely exhausts the inner, and conversely, the inner is never fully expressed in the outer, because the reciprocal mediation of the inner and outer is interminable, not temporally but speculatively. Given that the essence of essence is to appear, and that essence is never fully exhausted in appearance, it follows that an action both does and does not express the essence that shines through it. The unity and diversity that all actions display is therefore not best thought of as sequential aspects of an action's content, developing and fully developed, but rather as characteristics of an action whose content is conferred through the speculative movement between the inner and outer, the implicit and explicit, essence and appearance.²²