1. Action Explanationism

For as long as there has been anything called ‘the philosophy of action’ its practitioners have broached their topic through an investigation of the reasons that explain action. Anscombe accounted for the nature of action by way of a reflection on ‘a certain sense of the question “Why?”’, the relevant sense being ‘that in which the answer, if positive, gives a reason for acting’.Footnote ¹ Anscombe's question ‘Why?’ – like any question ‘Why?’ – requests an explanation. It asks, of something not understood, that it be made intelligible. For that is what reasons do: they explain. What is characteristic of reasons for action is that they explain action.

Anscombe was not the first to suppose that the question, ‘What is action?’ ought to be addressed by asking, ‘What is an explanation of action?’ The opening question of Wittgenstein's Blue Book, ‘What is the meaning of a word?’ was followed by a proposal about how to answer it: ‘Let us attack this question by asking, first, what is an explanation of the meaning of a word?’ Wittgenstein was not proposing a general philosophical method: he did not claim that every question of the form, ‘What is x?’ ought to be attacked by asking, first, ‘What is an explanation of x?’ Nevertheless, a dozen pages later, when he took up the topic of action, Wittgenstein began in exactly that way, by inquiring how an action is explained. That was the context of his influential claim that the question ‘Why?’ is ambiguous:

This remark by Wittgenstein set the terms for one of the most controversial questions of twentieth-century action theory: ‘How does a reason for action explain action?’

Like Wittgenstein, Gilbert Ryle argued that not all accounts of why something happened are causal explanations. ‘There are’, Ryle claimed, ‘at least two quite different senses in which an occurrence is said to be “explained”; and there are correspondingly at least two quite different senses in which we ask “why” it occurred. When we ask “Why did someone act in a certain way?” this question might, so far as language goes, either be an inquiry into the cause of his acting in that way, or be an inquiry into the character of the agent which accounts for his having acted in that way on that occasion.’Footnote ³ Ryle held that when we explain why a person did something by giving her reason for doing it, we are giving an explanation of the second type, not of the first. According to him, reasons for action explain action otherwise than by citing a cause.

This view was opposed by Carl Hempel and Donald Davidson.Footnote ⁴ According to them, an explanation of action adverting to an agent's reasons for acting is a species of causal explanation. In the introduction to his Essays on Actions and Events, Davidson identified this as the unifying idea of his work in action theory: ‘All the essays in this book… are unified in theme and general thesis. The theme is the role of causal concepts in the description and explanation of human action. The thesis is that the ordinary notion of cause which enters into scientific or common-sense accounts of non-psychological affairs is essential to the understanding of [action]… Cause is the cement of the universe; the concept of cause is what holds together our picture of the universe.’Footnote ⁵

If in the classical period of action theory – the middle third of the twentieth century – action explanation was the locus of the main doctrinal disputes, and thus also the primary object of inquiry, successive generations have plowed the same field. The view of Hempel and Davidson – that action explanation is causal explanation – soon displaced the one defended by Wittgenstein and Ryle, and for decades it has stood at the center of ‘the standard story of action’, according to which, in the words of Michael Smith, ‘an action is a bodily movement caused in the right way by a belief and a desire’.Footnote ⁶ Nowadays, even the standard story's most radical opponents aspire to replace it with an alternative story of action explanation in terms of which to account for action. Thus, in spite of many relatively superficial differences between followers of Anscombe and followers of Davidson, between anti-causalists and causalists, between event-causalists and agent-causalists, between champions of mechanical explanation and champions of teleology, between defenders of this or that account of an agent's mental states, and between partisans of ‘naive’ and of ‘sophisticated’ rationalization – in spite of all such differences, the entire tradition, from Wittgenstein to the present, has taken it for granted that the proper way to account for action is by accounting for an associated kind of explanation.

I will call this dogma – for that is what it is – action explanationism. My aim here is not to refute it, but to expose it to critical scrutiny. In fact, action explanationism is only one expression of a much more general tendency in contemporary philosophy. Philosophers now tend to theorize any expression of practical reason in terms of practical reasons, whose office it is to show something to be rational, justified, motivated, caused or otherwise intelligible. Though contemporary practical philosophy is, in general, explanationist, my immediate target is more specific. Here my sites are narrowly set on the species of explanationism that pertains to accounts of intentional action. In the controlled environment of the philosophy of action, where one only considers instrumental connections between ends and means, and where ethical and political questions are temporarily bracketed, it is easier to see, both, that there is an alternative, and what the alternative is.

2. Objective and Subjective Representations of Action

Let us begin by considering why it has seemed so natural, so undoubtedly correct, to address the question, ‘What is action?’ by asking, first, ‘What is an explanation of action?’ One apparent rationale begins from a general thought about philosophical method – namely, that the nature of a thing is revealed in an account of its representation.

There are countless expressions of this idea in the history of philosophy, and many within the analytic tradition. To answer the question, ‘What is a number?’ a philosopher might investigate the structure of our thought about numbers; or, to answer the question, ‘What is a cause?’ a philosopher might investigate what is involved in representing one thing as being the cause of another; or again, to answer the question, ‘What is life?’ a philosopher might investigate what it is to represent something as alive. Some of these investigations will exemplify what Strawson called ‘descriptive metaphysics’, a kind of metaphysics that, according to Strawson, was practiced by both Aristotle and Kant, and whose principal aim is to describe the ‘structure of our thought about the world’.Footnote ⁷

Not every action theorist ascribes to the general principle that the nature of a thing is revealed in an account of its representation, but many do. For those who do, explanationism can seem to be the natural way, or even perhaps the only way, to bring that principle to bear on the topic of human agency. If we answer the question ‘What is a number?’ by investigating the structure of our thought about numbers, then, it seems, likewise, we should answer the question, ‘What is an action?’ by investigating the structure of our thought about actions; and if accounting for the nature of action requires us to investigate the structure of our thought about actions, then, it seems, we ought to follow Anscombe and Davidson in focusing on the question ‘Why?’ and on the reasons for action that answer that question: for, it is in the structure of action explanation that we see what it is to think about action.

The point of articulating this seeming rationale for action explanationism is not to call into question the methodological principle that the nature of a thing is revealed in an account of its representation. Nor is the point to deny that action explanation is exactly what one should focus on insofar as one wants to understand what it is to represent action. Taking these for granted, what I mean to question is the further idea that this methodological principle dictates that we answer the question, ‘What is an action?’ as we would the question, ‘What is a number?’ – by embarking on an investigation of what it is to represent that which is in question. The difficulty lies in seeing how an account of the representation of action could be anything other than an account of what it is to represent action.

To see how it could be something else, consider that, for certain values of x, phrases like ‘the representation of x’, or ‘the consciousness of x’, or ‘the awareness of x’, or ‘the thought of x’, are ambiguous. Wherever x is a subject of representation, or of consciousness, or of awareness, or of thought, such phrases are ambiguous between an objective interpretation and a subjective interpretation, according as they are taken to employ an objective or a subjective genitive construction. On the subjective interpretation of the phrase, ‘the representation of x’, x is the subject of representation: it is what does the representing. On the objective interpretation of the very same words, an x is the object of representation: it is what gets represented. Thought belongs to a thinker in two distinct ways: there is, on the one hand, the thinking I do, and, on the other hand, the thinking I suffer; the first is thinking of which I am the subject, the second is thinking to which I am subjected when you grasp me in thought.

In some areas of philosophy, this raises a question of methodology. We have been considering the following principle:

But where x is a subject of representations, the principle is ambiguous. It might be given an ‘objective’ interpretation:

Alternatively, the principle might be given a ‘subjective’ interpretation:

Both of these interpretations might be thought inadequate, because partial and one-sided. It might be thought that a genuine understanding must comprehend the unity of subjective and objective representations:

So, in areas of philosophy where the topic of reflection is itself a thinking subject, there is a methodological question that needs to be decided.

This methodological question does not always arise because the topic of reflection is not always a thinking subject. Someone who proposes to explain what a number is by accounting for ‘the representation of a number’ faces no such question. Thought relates to numbers in only one way: there is such a thing as thinking about a number, but no such thing as thinking as a number. The representation ‘of’ a number is always ‘of’ a number in the objective sense that a number is the object of representation.

But unlike numbers, agents think. As a result, someone who proposes to explain what an agent is by accounting for ‘the representation of an agent’ faces a decision. The decision is whether to account for (1) the representation of which an agent is the object, (2) the representation of which an agent is the subject, or (3) the unity of these representations.

There is exactly the same array of theoretical options when the topic of reflection is, not an agent, but the activity of an agent, an action. The philosopher who proposes to explain what action is by accounting for ‘the representation of action’ must decide whether to account for (1) the kind of representation that is ‘of’ action in the objective sense that it is characteristic of someone who is thinking about action, (2) the kind of representation that is ‘of’ action in the subjective sense that it is characteristic of someone who is performing an action, or (3) the unity of these representations.

Consider this array of options as it presents itself to a descriptive metaphysician, whose the aim is, again, ‘to describe the structure of our thought about the world’.Footnote ⁸ Because we are agents, some of ‘our thought about the world’ is thought that we think about agents, and some of it is thought that we think as agents. The practitioner of descriptive metaphysics is therefore faced with the decision whether to account for (1) the structure of our thought about the world insofar as we are thinking about someone acting, (2) the structure of our thought about the world insofar as we are acting, or (3) the unity of these structures of thought.

Or again, suppose that, as Davidson says, ‘the concept of cause is what holds together our picture of the universe’. The philosopher who proposes to account for what holds together our picture of the universe must decide whether to account for (1) the picture that is ours insofar we have agents in view, (2) the picture that is ours insofar as we are agents, or (3) the unity of theses pictures.

It is the same for a practitioner of conceptual analysis. Agents both apply concepts and are such to be conceptualized. The analyzer of concepts must therefore decide whether to account for (1) concepts in their application to agents, (2) concepts in their application by agents, or (3) the unity of these two applications of concepts.

Given that an agent is both a subject and object of thought – both a represent-er and a represent-ee – the methodological question can be put in various ways, but it cannot be avoided, not in the philosophy of action. In their practice, philosophers of action do in fact settle the question one way or the other, even if neither they nor their readers are conscious that a question is being settled. Ryle's example is instructive. His famous discussion of ‘knowing how’ in the second chapter of The Concept of Mind opens as follows:

So, Ryle's stated aim is to account for what is going on ‘when we describe people’ – that is, when we represent others, others whom we represent as exercising qualities of mind. His principal thesis is this:

What Ryle offers is a theory of ascriptions, or descriptions, or imputations, of intelligence. When I ascribe intelligence to someone – for example, to a baker – I subsume him under concepts of mind. A theory of such ascriptions is, thus, a theory of what one person thinks about another. But the person who is thought about – the one who figures as the object of thought in an ascription of intelligence – that other person is also a thinker, and he has thoughts of his own. While my mind is on the baker, his mind is on the bread. And if there can be a logic to my thought, there must be a logic to his. One might have expected that in a philosophical treatment of ‘knowing how’ the baker's thought would be the center of attention, since, after all, he is the one with know-how: it is he, the baker, not I, the spectator, whose mind is formed in the special way that is under investigation. But in Ryle's treatment of the topic – and in the subsequent literature – the baker's thought is not in fact the center of attention: the center of attention is the thought of someone other than the baker, someone who is thinking about the baker. The baker himself is not in the business of making ascriptions of know-how. Insofar as his mind is on his work, he is concerned with many things, about which he thinks many things: concerning the dough, he may think, for example, that it needs more time to rise, or concerning the stove, that it is too hot. These are ascriptions, but not of know-how. They are imputations, but not of intelligence. They are descriptions, but not of anyone as possessing an ability. In the act of baking bread, the baker applies concepts, but the concepts he applies are not concepts of mind. It was open to Ryle to give an account of the application of concepts by knowers-how, but instead he elected to give an account of the application of concepts to knowers-how. That is a decision to theorize human agency from the perspective of someone who is thinking about an agent, rather than from the perspective of the thinking agent himself.Footnote ¹¹ Such was the general practice in twentieth-century action theory. Action explanationism is the central manifestation of that practice.

3. Explanation and Calculation

Though Anscombe was not first analytic philosopher to theorize action in terms of explanation, she may have been the first to observe that there is an alternative. On her view, the question ‘Why?’ is significant because it reveals ‘an order that is there wherever actions are done with intentions’Footnote ¹² – but, as she notices, it is not unique in doing so. The self-same order of ends and means revealed by the question ‘Why?’ is also revealed by a question ‘How?’ that an agent confronts in acting.

Anscombe's account of intentional action is built on the observation that, if I am doing A, and someone asks me why, the question may draw out that I am doing A in order to do B. Repeated application of the question ‘Why?’ may then draw out that I am doing B in order to do C. This gives rise to a series of ends, A–C, in which each action is done for the sake of the next:

1. C. replenishing the house water-supply

2. B. operating a pump

3. A. moving my arm up and down

Anscombe notes that instead of reading this order from bottom to top, as a series of ends, A–C, we can also read it from top to bottom, as a series of means, C–A. She writes: ‘if [C] is given as the answer to the question “Why?” about A, B … can make an appearance in answer to a question “How?” [about C]. When terms are related in this fashion, they constitute a series of means.’Footnote ¹³ Just as successive answers to the question ‘Why?’ expose that I am doing A in order to do B, and B in order to do C, so, also, successive answers to the question ‘How?’ reveal that I am doing C by means of doing B, and B by doing A. In drawing out this series of means, the question ‘How?’ reveals the order of the agent's thought, starting with the end to be achieved and concluding with the means to that end. That is why, near the end of Intention, Anscombe claims to have uncovered ‘the same order’ as is revealed in Aristotle's account of the practical syllogism: ‘I did not realize the identity until I had reached my results’, she says. The explanatory question ‘Why?’ and the calculative question ‘How?’ lead, respectively, up and down the same purposive scale. It is one order, not two, because the way up is the way down.Footnote ¹⁴

One might have expected that Anscombe would be alarmed by her discovery that she had approached the topic from the opposite direction of her master, Aristotle. In fact, she is undisturbed: she appears to think that it makes no difference, one way or the other, which way we go at it. That would indeed be the case if the representation of intentional action were sufficiently like the representation of a whole natural number. The arithmetic function ‘+1’ ascends the scale of natural numbers, from 1, to 2, to 3, and the inverse function, ‘–1’, descends the same scale. Just as, in action theory, a single order can be surveyed from bottom to top, as a series or ends, A–C, or, alternatively, from top to bottom, as a series of means, C–A, so, also, in arithmetic, a single order can be surveyed from bottom to top, as a series of addends, 1–3, or, alternatively, from top to bottom, as a series of minuends, 3–1. One might think, as Frege perhaps did, that the most general and fundamental truths about numbers are encoded in arithmetic functions such as these. But whatever might be learned about the nature of a number through reflection on the functions of arithmetic, it could hardly make a difference whether we chose to focus on addition or subtraction, for these two inverse functions stand in the same kind of relation to any natural number. It would be very strange to think that either of them enjoyed theoretical priority relative to the other.Footnote ¹⁵

If explanation and calculation differed from each other only in the way that addition differs from subtraction then it would be six of one and half a dozen of the other whether we accounted for intentional action via the question ‘Why?’ or via the question ‘How?’ But we are already in position to see that explanation and calculation differ in a much more radical way. We have seen that, since numbers do not think, any representation ‘of’ a number is ‘of’ a number in the objective sense that a number is the object of thought. This means that addition and subtraction are two different structures of thought about numbers. But that is not the difference between explanation and calculation. The latter do not correspond to two contrasting structures of thought about action. Action explanation reveals the structure of what one thinks insofar as one thinks about action. But its opposite, calculation, reveals the structure of what one thinks insofar as one is acting.

4. Calculation as the Subjective Representation of Action

If I propose to do something – for example, to go home – I need to find the means to my end, and I therefore need to embark on a search. The problem I face, if I want to go home, is that I cannot simply ‘go’ there: there is no such thing as ‘going’ home – or for that matter anywhere else – except in some specific way by some specific means. Thus, the first question I must settle, as someone who wants to go home, is how to get there. Am I to walk home? Or to ride my bike? Am I to take a taxi, or a train, or a boat – or perhaps some combination of these? Making this decision may or may not be difficult, but a decision must be made. If I do not make it – if I do not come down on the question how, specifically, to achieve my end – I cannot achieve my end.

Coming to such a decision is not sufficient for discovering the means to an end. If I propose to take a taxi home, I still have to find a taxi: that is, I have to identify something the taking home of which would be the taking home of a taxi. This is no trivial task. (An American visiting London for the first time might search in vain for yellow sedans and arrive at the false conclusion that the city, though full of hearses, is empty of taxi cabs.) If I cannot find a taxi, either because none is available, or because, although one is available, I fail to recognize it as that for which I am searching, then I cannot take a taxi home. In that case, the question from which I began – the question how to get home – remains unanswered.

Or, to take a less mundane example, consider the predicament of Anscombe's gardener, whose ultimate objective is to establish the kingdom of heaven on earth. The Nazis are in power and are prosecuting a murderous war, and his immediate question is how to stop it. One way to stop the war, he thinks, is to get some better leaders in power. But how? One would first need to get the Nazis out, and one can't just run a candidate, not against the Nazis. The gardener decides to assassinate their party chiefs, who happen to be living in the house where he works. But how? With a rifle? With dynamite? No, he thinks: poison. But how? Should the poison be put in their bread-flour? (No, too risky.) It occurs to him that he could poison their water-supply, so he sets about finding a slow-acting poison. He continues in this manner until eventually he is standing in front of the garden water pump, gripping the handle and moving his arm up and down, up and down. The fact that this is a hare-brained scheme is neither here nor there. What matters is its trajectory: it starts from a general end, which could be achieved in various ways, and it proceeds through specifications of that end until it is brought to bear upon the concrete particulars of agent's immediate circumstances.

That is what it is to answer the question ‘How?’.Footnote ¹⁶ In the present context, the most important thing to notice about the answers to this question is that they are not reasons acting. They do not rationalize the action that is in question. They do not justify it, or disclose its motive, or give an interpretation, or situate it in a pattern, or appeal to a ‘wider context’, or identify a cause. They are in no way explanatory of what they are about. And yet, they are the primary concern of an agent as such.

5. Explanation as the Objective Representation of Action

If what speaks to the question ‘How?’ speaks to the mind of an agent, what speaks to the question ‘Why?’ speaks to the mind of a spectator. The answer to the question ‘Why?’ can only enlighten someone whose action is not in question: for, it is a question about the action of someone who already knows the answer to it, and who therefore does not stand to learn anything from the inquiry.

It follows that the poser of the question ‘Why?’ – the one for whom it is a question – inquires about the action of another. The question may or may not be posed directly to the one whose action is in question. If I care to know why you are setting up a camera, I can try to figure it out without saying a word, and thus without asking anyone, or I can ask a third party, or I can ask you. Whether or not I ask someone, and no matter whom I ask, the question is about someone other than myself.

Anscombe's discussion of the question ‘Why?’ focuses on the scenario in which the question is posed directly to the one whose action is in question. In what Michael Thompson calls ‘the fundamental scene Anscombe is working with throughout [her] book’, there is a confrontation between two people:

The scene here is well-described. When two individuals end up in the cognitive relation marked by the use of the second person – ‘Why are you doing A?’ – it is not on the basis of their mutual consent. One of the two individuals, ‘the observer’, has imposed himself on the other, ‘the agent’. Moreover, what will transpire between them is not the kind of transaction from which both can hope to benefit: it will not be an exchange, but a gift; in answer to the question ‘Why?’ reasons will be given.

This bears emphasis. In posing the question ‘Why?’ to an agent I am asking him to explain himself to me. So, I am asking him to do something that he had not been doing before I posed my question. When I tap him on the shoulder and request that he explain himself, he will have been doing something else – setting up a camera, or whatever it was that caught my eye – and answering the question ‘Why?’ is no contribution whatsoever to what he had been doing. (Explaining why one is setting up a camera is not a step in setting up a camera.) Thus, what I am asking him to do, in asking him to explain himself, is, from his perspective, simply a distraction. He may indulge me, but if he does, that is what he is doing.

A scene like this makes vivid that the concerns of an agent are radically different from those of an observer who is trying to understand what that agent is doing. Prior to the moment of contact when the observer poses the question ‘Why?’ – there needn't be such moment, but if there is, then prior to it – the agent and the observer are each engaged in rational inquiry, though in rational inquiries of different kinds: on the one hand, the agent is trying to figure out how to set up his camera, wondering, for example, where it should be placed and whether it needs a flash; on the other hand, the observer is trying to figure out why the agent is setting a camera up, wondering, say, whether he plans to take a picture of a building, or, perhaps, of someone passing by (Marilyn Monroe is known to be in the neighborhood), and if the latter, whether it is for journalistic or for merely sensational purposes. While the calculating agent is busily searching for means, the explainer of his action is trying to discern his ends.

6. Calculation as Prior to Explanation

Earlier I claimed that if action explanation were related to calculation in the way that addition is related to subtraction, then it would not matter whether one accounted for action via the question ‘Why?’ or via the question ‘How?’ We have already seen one important difference: namely, that while the functions ‘+1’ and ‘–1’ reveal two contrasting ways of thinking about a number, the questions ‘Why?’ and ‘How?’ do not reveal contrasting ways of thinking about an action: the question ‘Why?’ reveals a way of thinking about an action, but the question ‘How?’ reveals the way an agent thinks in acting. A second important difference is that the symmetry between addition and subtraction differs from the symmetry between explanation and calculation.

To get the relevant contrast in view consider, first, the symmetry between a footprint and a foot. The parties to it are not, as it were, equals: it is the foot's print, not the print's foot. It is similar between a face and its mirror image: the image reflects the face; the face does not reflect the image. What is characteristic of such symmetry is that it is grounded in one of the two symmetrical terms. There is conformity between the terms because one of the terms is such as to conform to the other. Between a face and its mirror image, the face is the prior reality: it and its image are alike, not because the face is like the image, but because the image is like the face. A foot and its footprint have the same contour because the footprint conforms to the foot, not vice versa.

Addition and subtraction are not related in anything like that way. Neither the series of addends, 1–3, nor the series of minuends, 3–1, has any plausible claim to be the prior reality to which the other conforms. There is between them a merely formal symmetry and a symmetry of equals.

By contrast, action explanation is such as to conform to calculation. It is a common trope among action theorists that action explanation represents the calculation of the agent whose action is explained. When Wittgenstein remarks that, ‘giving a reason is like giving a calculation by which you have arrived at a certain result’,Footnote ¹⁸ he appears to suggest that explaining one's action is a matter of conveying the thought that led one to act as one did. In a passage quoted approvingly by Carl Hempel, William Dray writes that a rational explanation of action is one that offers ‘a reconstruction of the agent's calculation of means to be adopted toward his chosen end’.Footnote ¹⁹ In a similar vein, Davidson asserts that ‘the explanation of an action [is] the retracing of a course of reasoning on the part of the actor’.Footnote ²⁰ The ‘re-’ of ‘reconstruction’ and of ‘retracing’ is the ‘re-’ of ‘reverse’, ‘return’, and ‘reflect’. The prefix indicates that something has been doubled, transposed or bent back to its point of origin. If action explanation represents the calculation of the agent whose action is explained, then it represents a kind of thought that is not only distinct from, but prior to, action explanation.

In fact, calculation is prior to explanation in several different ways. First, calculation is prior in the order of normativity. Insofar as my aim is to understand the reason why you are setting up a camera, your thought, as agent, provides the standard of correctness for my thought, as observer. My thought that you are setting up a camera in order to take a picture of Marilyn Monroe is a good explanation of your action if and only if it is accurate in depicting you as having the aim of taking a picture of Marilyn Monroe, and as having decided to set up a camera as a means conducive to that end. In other words, my answer to the questions ‘Why?’ is responsible for conforming to your answer to the question ‘How?’ In general, an explanation portrays the depicted agent as thinking a certain way, and the explanation is evaluated as true or false by reference to the agent's thought – that is, according to whether it portrays the agent's thought accurately or inaccurately.Footnote ²¹ But the reverse is not the case: the agent's calculation does not portray the thought of an explainer. While explanatory thought is evaluated by reference to calculative thought, calculative thought is evaluated by reference to the task at hand.

Second, calculation is prior in the order of time. It is only possible to explain what is the case. So, if it is possible to explain why the gardener is operating the water-pump, then it must be the case that he is operating the water-pump. But if he is already operating the water-pump, then he has already chosen this action as a means to his end – his end being, say, to replenish the house water-supply (with poisoned water). Thus, the calculation that is ‘reconstructed’ or ‘retraced’ in a sound explanation of why the gardener is operating the water-pump, this calculation has already finished its work by the time that explanation begins.

Third, calculation is prior in the order of being. Calculation produces the explanandum of action explanation. Were it not for the agent's calculation, there would be no action to explain, and thus no conceivable explanation of it. The very possibility of action explanation depends on the actuality of the agent's calculation.

Finally, calculation is prior in the order of account. To rationalize an action is to portray the action as being the product of calculation. So, a philosophical account of rationalization must explain what it is to portray an action as being the product of calculation. So, it must explain what calculation is. This means that an account of rationalization must contain an account of calculation or else be incomplete as an account of rationalization. The reverse is not the case because to calculate is not in itself to portray anyone as rationalizing anything.

None of this is to say that calculation is intelligible in abstraction from explanation. It is not to deny that the subjective representation of action and the objective representation of action can each only be understood in relation to the other. It is rather to observe that, within their relation, explanation and calculation are not on an equal footing: one of them is beholden to the other.