1.1 Anscombe’s Definition

Anscombe’s idea that there is a special question ‘Why?’ reserved for intentional action draws on a passage in Wittgenstein’s Blue Book, in which he distinguishes between two ‘uses’ of the word ‘why’. At the time when Wittgenstein composed this text, the two dominant theories in the philosophy of action both held that an intentional act is a movement of the body with a distinctive kind of mental cause, either something like a conscious decision or choice (generally called a ‘volition’, ‘act of will’ or ‘intention’) or alternatively an idea of a movement or a kinesthetic image.Footnote ⁴

Wittgenstein attacked both these theories, claiming that intentional acts are normally explained by motives or reasons, neither of which is any kind of cause: so, not a volition or act of will, nor a mental image, but something supposedly quite different. And corresponding to these two supposedly different kinds of explanatory factor – causal and rational – are two kinds of explanation, both of which we can seek with the word ‘why’. We evidently do ask the question ‘why?’ when we are interested in knowing what caused someone to behave in a particular way. Anscombe mentions a nice example:

But according to Wittgenstein, we can also ask ‘why’ a person did a certain act in a quite different sense, because we are interested in knowing their intention or motive or reason for doing it, and this question is not answered by giving a cause. Instead, ‘Giving a reason is like giving a calculation by which you have arrived at a certain result’ (Wittgenstein, Reference Wittgenstein1958a, p. 15).

Wittgenstein did not attempt to define an intentional act in the Blue Book or the Philosophical Investigations. But Anscombe seized on his claim that the word ‘why’ has a ‘double use’, and his distinction between cause and reason, because she saw the opportunity to employ it for this purpose. Here is her first statement of the idea:

Anscombe’s proposal that we define intentional action in terms of the concept of a reason was highly influential, not least because Davidson adopted it, despite the substantial differences between his conception of reasons and reasoning and hers. However, Davidson suggests that we can simply define an intentional action as one done for a reason (Davidson Reference Davidson1980, p. 6). This is a bolder definition than Anscombe’s because it implies that every intentional act is done for a reason, even if the reason is too obvious or too insignificant to mention. Anscombe, by contrast, accepts that we sometimes do an act intentionally – quite literally – for no reason at all. She writes:

This is why Anscombe uses the formulation ‘actions to which a certain sense of the question “Why?” is given application’, rather than defining an intentional action simply as one that is done for a reason, as Davidson does. Anscombe accepts that if we ask someone to explain their action, using the special question ‘Why?’, the answer may be ‘positive’ or ‘negative’, by which she means that it may give a reason or it may deny that there was a reason, but she maintains that the denial does not imply that the question was out of place.

For example, suppose you are asked why you are tapping your foot. You may say ‘It had gone to sleep, I was trying to get the feeling back’ or ‘It’s just something I do occasionally, for no reason’. The first answer is positive, in Anscombe’s sense, whereas the second is negative. But Anscombe thinks of both these answers as granting the legitimacy or pertinence, or as she puts it the ‘applicability’ of the question.Footnote ⁶

Now suppose you reply instead ‘I didn’t realise I was tapping my foot’, or ‘I can’t help it, I have a tremor in my leg’. These answers, she claims, reject the question. There is no point in asking for your reason if the movement is one you were unaware of making, or if it was caused by a tremor, so ‘I didn’t realise I was tapping my foot’ implies that the special question ‘Why are you tapping your foot?’ was out of place or ‘inapplicable’, whereas ‘It’s just something I do occasionally, for no reason’, does not imply this. The question is apposite, but the answer just happens to be ‘No reason at all’.

Anscombe makes one further comment in support of her distinction between the ‘For no particular reason’ answer, which accepts the question ‘Why?’, and the ‘I didn’t realise’ or ‘I have a tremor’ answer, which rejects it. She compares saying ‘For no particular reason’ to answering the question ‘How much money do you have in your pocket?’ by saying ‘None’:

Anscombe’s thought is that the question ‘How much money do you have in your pocket?’ is ‘refused application’ if one does not have a pocket, but not if one has an empty pocket. In both cases, then, the question (‘Why?’ or ‘How much?’) can be given application or refused application, and if it is given application, the answer can be positive (some reason/some money) or negative (no reason/no money).

The meaning of the first passage in which Anscombe presents her definition of intentional action should now be clear: intentional actions are ones to which the special question ‘Why?’ is given application; and what is special about this ‘Why?’ is that ‘the answer, if positive, gives a reason for acting’. We have not attempted to gloss or explain the idea of a ‘reason for acting’. For the moment we shall assume that the sense of the question ‘What is/was your reason for doing that?’ is sufficiently clear. But as we shall see, its precise meaning matters a good deal.

1.2 Criticism of Anscombe’s Definition

Anscombe’s definition of intentional action has been interpreted, criticized, and defended in an extensive literature, which we do not propose to examine or add to here. What interests us is whether Anscombe’s special question ‘Why?’ can be used to define a much larger category of acts than intentional ones, namely, those for which the agent is responsible. However, we shall approach this question by considering a series of objections that can be advanced against Anscombe’s definition of intentional action, because it seems to us that these objections cease to bite, and the doubts that motivate them are allayed, if the special question is repurposed in the theory of responsibility in the manner we propose.

Four objections are pertinent to our argument. The first is that the class of acts to which the special question is ‘given application’ can reasonably be thought to be either larger or smaller than Anscombe suggests it is. For what is wrong with saying that if I am tapping my foot because I have a tremor in my leg, the special question is still given application because tapping one’s foot is the kind of movement that can be made for a reason – unlike, say, a sneeze? One might add that the reply ‘I have a tremor’ does not amount to a rejection of the question in the way that it would if I said, ‘None of your business!’ or ‘Leave me alone!’ Alternatively, what is wrong with agreeing that ‘I have a tremor’ rejects the question but insisting that ‘No reason at all!’ does as well, because it implies that the question’s presupposition, that I was tapping my foot for a reason, is mistaken? We shall argue in due course that the scope of the question is in fact much larger than Anscombe claims. But what matters for the moment is that Anscombe does not offer a principled reason for defining its scope in the way she does.

The difficulty here is that a question may be said to be ‘refused application’ if it depends on a presupposition that is false. But it is not obvious what the presuppositions of the special question ‘Why?’ are. The question ‘How much money do you have in your pocket?’ does indeed presuppose that the person addressed has a pocket, at least in most cases, the presupposition being triggered by the definite description ‘your pocket’. Similarly, the question ‘Is the king of France bald?’ presupposes that France has a king. But what reason is there to accept that Anscombe’s special question depends on the presupposition that the act was either done for a reason or for no reason, rather than carrying one of the presuppositions suggested in the previous paragraph? None is given in Intention, except for the analogy with the ‘How much money?’ question. But although this example suggests that a presupposition can have the ‘some or none’ disjunctive form, it does not follow that the special question does. Besides, it is also arguable that the ‘How much money?’ question carries a ‘soft’ or cancellable presupposition that the person addressed has some money in their pocket, at least in some cases.Footnote ⁷

The second objection is this. Suppose the special question is, as Anscombe claims, ‘given application’ to all and only intentional acts. Even if this is assumed, Anscombe’s definition is still unsatisfactory. Her own comparison between the ‘Why?’ question and the ‘How much money?’ question brings out why. Consider a pocket. It may be true that a pocket is the part of a piece of clothing of which we can sensibly ask ‘How much money do you have in it?’, but if so, presumably the reason for this is that a pocket is, by definition, a small bag or pouch sewn into clothing, and money can be kept in a small bag or pouch. Nobody would define a pocket – or explain what a pocket is – by stating that it is the part of a piece of clothing that the question ‘How much money …?’ applies to. On the contrary, the fact that the question applies to it is explained by the definition: we can ask how much money someone has in their pocket because a pocket is a small bag or pouch sewn into clothing, and money can be kept in a small bag or pouch. In the same way, if it were true that an intentional act is the kind of act about which we can sensibly ask the special question, the definition of an intentional act would have to explain why.Footnote ⁸

Third, Anscombe’s claim that the special question is ‘given application’ to all and only intentional acts appears to be not merely doubtful and unproven, as per the first objection, but demonstrably false, because there is a kind of behaviour that plays an ineliminable part in human life, which is not intentional but often is done for a reason, namely, the unintentional spontaneous expression of an emotion such as amusement, gratitude, anger, grief or joy.Footnote ⁹ Consider laughter. We sometimes laugh intentionally, to put someone at ease, or to contribute to a jolly atmosphere. But much laughter is not intentional, for example, nervous laughter, laughter caused by tickling, and (for the most part) spontaneous laughter at a joke.Footnote ¹⁰ Yet we may be laughing for a reason, that is to say, the kind of reason that qualifies as a positive answer to the special question ‘Why?’

Consider laughing at a joke, the kind of laughter that bubbles out of us without being intentional. There may be laughter that is both spontaneous and intentional, but we are not interested in that now. When our laughter is not intentional, but simply a spontaneous expression of amusement, it is not the upshot of a ‘calculation’ or process of deliberation. Giving a reason is definitely not in this case like giving a calculation by which you have arrived at a certain result. This is not to say that our spontaneous laughter is always uncontrollable: we can quite often inhibit, modulate, or stop it, if we choose to. But – this is the key point for our purposes – neither does it mean that we are not laughing for a reason, or that the special question is ‘refused application’. Remember, the special question is one where a positive answer gives a reason for doing the thing in question, and we certainly can give a reason for laughing at a joke, whether we laugh intentionally or not, and however difficult it can be to explain exactly what made it funny. Of course, it may be a poor reason, because the joke was offensive, or deserved a groan rather than a laugh. But equally, it may be a perfectly good reason, which fully justifies laughing.

The reasons for thinking that the special question is ‘granted application’ when we laugh spontaneously at a joke include the following. (i) Our reasons for laughing can be considered good or bad reasons, depending on how far they justify our laughter, and if they are good reasons, they show our laughter to be not merely a predictable response but a reasonable one. (ii) We can normally state our reasons for laughing, even if we find it difficult to say more than that the joke was funny. (iii) Stating our reasons will typically have the same dual function of explaining and justifying as it does when we state our reasons for doing an intentional act–especially an act in the present or the recent past. (iv) We normally have the same immediate and ‘non-observational’ knowledge of our reasons for laughing as we have of our reasons for doing an intentional act. (Which is not to say we are infallible in either case.)

It can be replied, on Anscombe’s behalf, that the ‘reasons for acting’ that correspond to the special question ‘Why?’ are not merely justifications, but justifications of a specific sort, ‘reasons why it would be useful or attractive if the description [of an act or its outcome] came true’ (§3), as Anscombe puts it in Intention, but our reasons for the spontaneous action that expresses our emotions, like our reasons for the emotions themselves, are not justifications of this kind. For example, a joke’s being funny justifies amusement, and the laughter that expresses it, because it makes these reactions apt or fitting, not because it makes them useful or attractive.

However, this argument depends on an excessively narrow understanding of reasons for acting, because it ignores a kind of ‘reason for acting’ Anscombe herself draws attention to in Intention – a ‘backward-looking’ reason, such as ‘He killed my father’, ‘I promised I would’, or ‘The light was red’.Footnote ¹¹ These are all reasons for doing something intentionally – e.g. killing a man, preparing supper, stopping at the lights – but, like ‘The joke was funny’, they refer to a past event or circumstance rather than saying what was ‘useful or attractive’ about the conduct they explain.

Anscombe’s attitude to this kind of reason may seem elusive. On the one hand, she says that the avenger’s explanation ‘He killed my father’ is ‘surely a reason rather than a cause’ (§5) and describes it as ‘coming in the range of “reasons for acting”’ (§35). On the other hand, she claims that ‘He killed my father, so I shall kill him’ does not qualify as a ‘form of reasoning’ (§35). But her position is not difficult to explain. Evidently, ‘He killed my father’ qualifies as a ‘reason for acting’, despite not saying what positive value the agent saw in killing the man, because it allows us to ‘construct’ (as she later put it) a reason that does say this, such as ‘for the sake of vengeance’ or ‘in order to avenge his father’s death’.Footnote ¹² Similarly, ‘I promised I would’ implies that the agent saw promise-keeping as a positive value and ‘The light was red’ implies the same about following the Highway Code. So, in these cases too we can ‘construct’ a description of the agent’s end: ‘keeping one’s promises’ or ‘following the Highway Code’.

In reality forward-looking reasons do not have this primary role. A backward-looking reason’s explanatory force – the fact that it makes an act intelligible – does not depend on our being able to transform it into a forward-looking one. On the contrary, if Orestes is said to have killed Clytemnestra in order to avenge his father death, we can explain what this means by saying that he regarded his father’s death as a reason for killing her. The backward-looking reason explains the forward-looking one, not the other way around. As Anscombe herself acknowledged decades later, the idea that forward-looking reasons are always primary is a dogma (Anscombe Reference Anscombe1981, p. viii). Hence, the reply to the objection fails. Action that is not intentional, such as laughing spontaneously at a joke or weeping at sad or joyous news, can be explained by a ‘reason for acting’ no less than action that is intentional. Hence intentional action cannot be defined in terms of the special question “Why?”’

The fourth and final objection applies the same reasoning to a different kind of case. For the special question ‘Why?’ is also given application in cases involving unintended but foreseen effects of intentional action, whether these are caused intentionally or (as Anscombe maintained) not. She writes as follows:

Setting aside the controversial question of whether acts of this kind are or are not done intentionally, it is puzzling that Anscombe holds that one rejects the special question ‘Why?’ in such cases. Consider the case of harming a patient’s immune system by administering chemotherapy, which Anscombe claims is done voluntarily but not intentionally. Surely there is an intelligible answer to the special question ‘Why?’, namely, that one causes the harm for the sake of the curative effect the therapy is intended to achieve – not as a means or as an end, but as a sine qua non.Footnote ¹³ Hence, it is legitimate to put the question: it is ‘given application’.Footnote ¹⁴ The opposite view relies on the mistaken principle that the only ‘reasons for acting’ (including acting intentionally) are ‘reasons why it would be useful or attractive if the description came true’ (§3).

2.1 Previous Criticisms Disappear

We have criticised Anscombe’s use of the special question ‘Why?’ to define intentional action on four main grounds. As we shall now see, once the special question ‘Why?’ is repurposed in the way we have proposed none of these objections applies. For ease of exposition, we shall consider the second objection first. But first, a brief comment about the special question ‘Why?’

First, every question ‘Why?’ – whether special or not – is a request or demand for an explanation, but not every apposite response to it provides or offers an explanation. For example, it may offer an apology – i.e., one that does not include an explanation – instead. Second, although we shall speak freely – and correctly – about explaining conduct and about the special question ‘Why?’ being given application to an act, it should always be borne in mind that what we can be asked to explain, and what we can be responsible for, is doing or not doing such and such a kind of act, for example, giving or not giving the baby a bath. It necessarily involves a specification or description of what we are supposed to have done or to have left undone. Thus, ‘Why did you/didn’t you kill the old man at the crossroads?’ and ‘Why did you/didn’t you kill your father?’ are not the same question, even if they concern one and the same act.

Explanatory Order

Intentional action cannot be defined as action to which the special question ‘Why?’ is given application. For even if it were true that the special question is given application to every intentional act and to no act that is not intentional – which is doubtful – the definition of intentional action would, we have argued, have to explain why. However, one concept that the special question can be used to define is the concept of answerability, because an answer can only be defined as a response to a question, or in very similar terms. (See for example OED, answer, 1.a.: ‘A spoken or written reply, or some other response, to a question.’). In fact, question and answer stand in converse relations to each other; but like ‘offence’ and ‘defence’, and unlike ‘buy’ and ‘sell’ or ‘above’ and ‘below’, there is a definitional asymmetry between ‘question’ and ‘answer’, since the latter is defined in terms of the former, and not the other way round. The reason why this definitional asymmetry exists is that something counts as an answer (or a defence) only because it stands in a certain relation (e.g., being an intended response) to a presumed question (or a presumed offence), but not vice versa.

Furthermore, a specific kind of answer corresponds to a specific kind of question. Thus, suppose we are interested in knowing a person’s reasons for acting in a certain way. The question we would use to elicit this information is the question why they conducted themselves as they did, as opposed to, say, where or when. But it is not just any why.Footnote ²⁰ It is a why that seeks their justificatory reasons for their conduct (Smith, Reference Smith2015, p. 103), as opposed to non-rational causes. In other words, it is the special question ‘Why?’ But giving justificatory reasons is precisely how people answer for their conduct. Hence the answer that an answerable person gives, or ought to give, is not a response to just any question about their thoughts, feelings, or conduct; it is a response to this question: the special question ‘Why?’ That is why the special question can be repurposed in the way that we propose, within a theory of responsibility as answerability.

The ultimate reason why the special question ‘Why?’ can define answerability but not intentional action is that there is a fundamental connection between the special question ‘Why?’ and answerability, but not between this question and intentional action. The basic idea behind intentional action has been explained historically in different ways, e.g., in terms of a goal or aim or plan, in terms of wanting, willing or volition, in terms of guidance or control, and so on. But none of these explanations is properly spelled out in terms of any kind of why-question. By contrast, the same is not true for the concept of an answer, or the concept of answerability. We cannot refer to an answer without thereby implying the existence of a question, and we cannot describe a person as answerable without thereby implying the existence of the special question ‘Why?’Footnote ²¹

Thus, once we repurpose the special question ‘Why?’, the criticism about reversed explanatory order loses its force.

Applicability of the Question ‘Why?’

We also argued that Anscombe’s distinction between cases in which the special question ‘Why?’ is given application and cases in which it is refused application lacks a convincing rationale. However, once we use the special question ‘Why?’ to define responsibility, the distinction between cases does have a rationale, because the cases in which the question is refused application are those in which the agent has adequate grounds either for an exemption from responsibility or for a denial that they did the alleged act, etc., and it is given application in all other cases. The grounds for an exemption or a denial may be any of the following (stated simply here and discussed in more detail below):

1. i. the agent concerned lacks capacity or competence (a) at the time of the demand for an explanation or (arguably) (b) at the time of their alleged conduct;

2. ii. the agent lacks physical control (e.g. a tremor).

3. iii. the party requesting or demanding an explanation, or the forum or institution in which it is being requested or demanded, does not have the required standing;

4. iv. the agent did not do the act or omission, hold the belief, etc., in the first place

(i), (ii) and (iii) are generally regarded as different kinds of exemptions, whereas (iv) is termed a denial. But the distinction between exemptions and denials is not essential for our purposes. What we want to emphasise is what exemptions and denials have in common, namely, that both imply that there is nothing that the individual concerned can be legitimately required to answer for.Footnote ²² In this regard, they differ from justifications and excuses, which admit answerability but avert liability, e.g., to blame or censure. In our view, this difference explains the distinction between those cases in which the special question is given application, and those in which it is refused application.

We shall now consider the four grounds for an exemption or a denial in greater detail and argue that in each of these cases the special question ‘Why?’ is refused application, as our proposal predicts.

(i) Exemptions, unlike denials or excuses, are generally due to a lack of capacity, competence, or capability. For example, focusing for the moment on the criminal law, it is uncontroversial that a defendant can be exempt from responsibility, understood as answerability, because they are not fit to plead. And in this case the special question ‘Why?’ is not given application.Footnote ²³ But there is controversy over whether lack of competence at the time of the alleged offence, such as when the defendant meets the M’Naghten test for insanity, also provides the basis for an exemption.

The standard view in the juridical literature is that lack of competence at the time of the alleged offence does not provide the basis for an exemption from responsibility, and that instead it either provides the basis for a denial or for an excuse, depending on the details of the case.Footnote ²⁴ However, Duff argues that exemptions also apply in cases where a person is incapable of being guided by reasons at the time of the alleged offence. According to Duff, exemptions of this kind apply when the defendant was ‘incapable of functioning as a rational agent – of operating within the realm of reasons’ (Duff, Reference Duff2009, p. 286). He writes:

To illustrate the difference between Duff’s view and the standard juridical view, consider the Canadian case R v Parks ([1992] 2 S.C.R. 87). Kenneth Parks drove 14 miles to his parents-in-laws’ home, entered the house using a key they had given him and bludgeoned his mother-in-law to death. His defence team argued that he was sleepwalking – technically, in a state of ‘non-insane automatism’ – and he was acquitted. According to Duff’s view, Parks may have an exemption, on the basis that he was not ‘operating within the realm of reason’ at the time of the alleged offence. For although he struck his mother-in-law repeatedly with a tire-iron for the reason that doing so was likely to cause her death, it seems that he was not capable of being guided by the moral and legal reasons to desist. By contrast, according to the standard view, an exemption cannot be based on lack of capacity at the time of the alleged offence. Rather, Parks may have either an excuse or a denial, the latter on the basis that an attack conducted while sleepwalking does not qualify as a voluntary act, in the sense employed in the criminal law.

We shall remain neutral on Duff’s disagreement with the standard juridical view. What we insist upon, and explain further below, is that however this is decided, the distinction between these kinds of cases is aligned with the distinction between cases where the special question ‘Why?’ is refused application and cases where it is given application. (We have introduced this idea within a legal framework, but we should emphasise that according to our view, the concepts and distinctions explained here apply to moral responsibility no less than to criminal responsibility.)

(ii) A second kind of case in which a person is exempt from responsibility is also one in which they are not capable of being guided by reasons not to do the act or make the movement in question. But in this case, it is not because of any intellectual impediment or psychological disorder, but rather a lack of physical control. Typically, the lack of control in such cases is due to a disease or pathological condition, such as a tremor, an epileptic seizure, Tourette Syndrome, etc.

However, suppose a clumsy dancer steps on his partner’s toes. He is not normally exempt from responsibility for doing so, although he may have an excuse that diminishes or even eliminates his culpability, say, that the floor was slippery or the room was dark. This is because the clumsy dancer, unlike the person experiencing a seizure, is still required to conform to certain norms and standards in his dancing. As Horder explains:

By contrast, a person experiencing a seizure ‘does not stand to be judged in the light of guiding standards, in terms of what ought to have been done or not done’ (Horder, Reference Horder2007a, p. 47). A person may stand to be judged for how they deal with their epilepsy – what adjustments they make because of it – but not for the seizure itself. Thus, they are exempt from responsibility rather than excused.

The applicability of the special question ‘Why?’ tracks the distinction between this second class of exemptions (as in cases of epileptic seizure) and excuses (as in cases of clumsy dancing). The question is refused application in respect of the seizure, but it is given application when a person is asked to explain why they trod on their partner’s toes. For in this case, it is legitimate to challenge the agent to provide a justification or (more likely) an excuse for their conduct, which they might provide by showing for example that they took reasonable steps (no pun intended) to ensure that they would not harm their partner, say, by practicing carefully, not wearing steel-capped shoes, and so on.Footnote ²⁵ It is true that they might respond to the challenge without providing either a justification or an excuse, e.g., by apologising for their clumsiness, but as we have seen, this would not amount to a rejection of the question.

(iii) The third kind of exemption is one where the party requesting or demanding an explanation, or the forum or institution in which it is being requested or demanded, does not have the required standing, for example, where an acquaintance asks me an intrusive question and I tell him that it’s none of his business, implying that I am not responsible to him. The equivalent of this class of cases in the criminal law is immunity to prosecution. But whether we are concerned with legal or moral responsibility, this is another class of cases in which agents are exempt from responsibility and the special question ‘Why?’ is refused application. This reflects the fact that responsibility in the sense of answerability is not just for X but also to Y, that is, to the individual or group or institution that has the standing to request or demand an answer to the special question ‘Why?’Footnote ²⁶

Notice that lack of standing differs from lack of competence and lack of physical control inasmuch as the latter exempts an agent from responsibility absolutely whereas the former exempts them relatively, that is, relative to the individual or institution lacking standing.Footnote ²⁷ As an anonymous referee helpfully pointed out, this is because lack of standing is a fact about the one requesting or demanding an explanation whereas lack of competence and lack of physical control are facts about the one of whom it is requested or demanded. Anscombe herself did not consider standing, presumably because it has no bearing on whether a person’s conduct was intentional or not. By contrast our proposal includes it in a comprehensive theoretical framework and explains its relationship with other grounds for refusing application to the special question ‘Why?’

(iv) Finally, the fourth class of cases are denials rather than exemptions, that is, denials that one engaged in the conduct in question, where such conduct could be an act, an omission, or neither (e.g., being drunk in charge of a motor vehicle). Since responsibility is always responsibility for something, not having engaged in the conduct in question precludes being responsible for it. For instance, the question ‘Why did you steal the car?’ is rejected by the answer ‘I did not steal it’.

However, understanding this kind of case also requires a more nuanced approach to ignorance than Anscombe provides. Anscombe says that the special question is always refused application by a plea of ignorance, such as ‘I was not aware I was doing that’ (§6). In Anscombe’s own example, a man is pumping water into a house without being aware that it is poisoned. Anscombe maintains that in these circumstances, the question ‘Why are you pumping poisoned water?’ (interpreted in the special way) is refused application.

However, in our view a person who unknowingly poisons the water supply in this way can be required to answer for their conduct, and if they claim that they were not aware that the water was poisoned, this will not amount to a rejection of the question or imply that it was based on the mistaken assumption that they did so intentionally. It is true that ignorance may avert or reduce liability to blame or punishment, although it does not exclude negligence. But if it does avert liability, because it is not itself culpable to any degree, it qualifies as an excuse as opposed to a denial or an exemption and does not imply that the question was not pertinent or applicable.Footnote ²⁸

So, the special question is not always refused application by a plea of ignorance. In some cases, ignorance provides an excuse, and in others it is the basis of a denial. Which kind of case we are concerned with depends on whether knowledge or belief is a constitutive element of the conduct under consideration. The special question is refused application when it is such a constitutive element, and when ignorance can therefore form the basis of a denial.

For example, compare the water case with a case of lying. The man in the water case is not exempt from responsibility for his conduct because he did not lack the capacity for rational thought or agency, nor is he in a position to deny that he pumped poisoned water into the house. So, on our view, the special question is given application, but the man may be able to avert liability by pleading ignorance, depending on whether he was negligent. By contrast, if one does not know or believe that what one is saying is false, one does not lie. For instance, someone who answers a question incorrectly out of ignorance or by mistake is entitled to reject the question why they lied because they did no such thing. Hence, in this case ignorance can be the basis of a denial. Similarly, ignorance will provide a basis for the denial of a criminal offence if the mens rea condition for the offence is knowledge or belief, as in the case of handling stolen goods.

So, if we use the special question ‘Why?’ to define responsibility instead of intentional action, the distinction between cases in which it is given application and cases in which it is refused application has a convincing rationale.

Non-Intentional Action

Our third criticism of Anscombe’s use of the special question ‘Why?’ to define intentional action was that the question is in fact given application to conduct that is not intentional, because it is accidental, inadvertent, or automatic. Anscombe considers the question ‘Why did you knock the cup off the table?’ answered by ‘I thought I saw a face at the window and it made me jump’ (§5) and the question ‘Why did you jump back suddenly like that?’ answered by ‘The leap and loud bark of that crocodile made me jump’ (§8). In these cases, she says, the special question ‘Why?’ (‘Why did you knock the cup off the table?’ or ‘Why did you jump?’) is refused application. But even an automatic reaction to seeing a face in the window or to a sudden noise may be subject to legitimate criticism or challenge – e.g. in circumstances where a heightened level of self-control is vital to avoid detection – and this may take the form of the special question ‘Why?’Footnote ²⁹ Hence, the special question is given application in some cases of these kinds, though not in all.

However, if we employ the special question instead as part of a theory of responsibility, this objection disappears. For we are responsible not only for intentional acts and omissions but also for acts that are not intentional, and for attitudes, emotions, and even beliefs. As we saw earlier, someone who spontaneously laughs at a joke is responsible for their laughter, which may have been inappropriate, say, because the joke was tasteless or offensive; and if they are unable to provide a justification or an excuse for laughing, they may become liable to an adverse judgement or reaction. Similarly with attitudes and emotions: even if they are not intentional, they can be justified or unjustified and we can be answerable for them (Smith, Reference Smith2005). Moreover, we also believe things for reasons and it is plausible that we can be answerable for beliefs. Finally, as noted above, we can be responsible for accidental, inadvertent, or automatic behaviour too, including in situations similar to Anscombe’s example of the startled person who jumps when they see a face at a window or hear a loud bark.

In sum, repurposing the special question ‘Why?’ in the context of a theory of responsibility allows us to offer a more convincing account of when the question is refused and when it is given application. It is refused application when a person has adequate grounds either for an exemption from responsibility or for a denial. In other cases, including many acts that are accidental, inadvertent, or automatic, the special question ‘Why?’ is given application. Furthermore, the repurposed special question ‘Why?’ not only ranges across actions, emotions, and beliefs. It can also capture both moral and non-moral answerability. And this is a further virtue of the account we propose, since answerability is not confined to the moral domain (Shoemaker, Reference Shoemaker2015, p. 78; Smith, Reference Smith2015, p. 107).

Foreseen but Unintended Side-Effects

Finally, our fourth criticism of Anscombe’s use of the special question ‘Why?’ concerned the difference between foreseen and intended outcomes, or between direct and oblique intention. As we have seen, Anscombe holds both that the special question ‘Why?’ is only given application to intentional action and that foreseen side-effects are not caused intentionally. But the special question is not refused application to acts that consist in causing foreseen side-effects, regardless of whether Anscombe is right in thinking that it is refused application to acts that consist in causing unforeseen side-effects, and so these two claims cannot both be true. But once we repurpose the special question ‘Why?’ in the way we have proposed, this problem disappears. For we are normally responsible for the foreseen side-effects of our intentional conduct, regardless of whether we are responsible for causing their unforeseen side-effects. Hence, we should expect the special question ‘Why?’ to be given application in these cases.

2.2 Responsibility and Language

It is striking to discover that these four powerful objections to defining intentional action in terms of the special question ‘Why?’ all lose their bite when we use the question to define responsibility instead. But even if this is granted, doubts about our proposal may remain. First, is the question we employ to define responsibility really Anscombe’s special question ‘Why?’? Second, supposing that it is, does our approach represent a significant advance in the theory of responsibility, or is it merely a footnote, linking the concept of answerability to Anscombe’s account of intention? We shall address these questions in turn.

Regarding the first question, philosophers who find Anscombe’s theory of intention more plausible than we do may feel tempted to concede that responsibility can be defined in terms of a question ‘Why?’ while insisting that it is not the same question as Anscombe’s. After all, we are morally and legally responsible for a wide range of conduct that is not intentional – as Anscombe is well awareFootnote ³⁰ – and so our question ‘Why?’ is ‘granted application’ in a wide range of cases in which Anscombe tells us hers is not. So, are we really using Anscombe’s question? Or is our disagreement with her merely terminological, and capable of being resolved simply by recognising a third kind of question ‘Why?’?

Our reply is that our disagreement with Anscombe is not merely terminological, it is about the scope of the special question ‘Why?’ (i.e., the range of conduct to which it is ‘granted application’) and hence the philosophical purpose it can serve (i.e., the kind of conduct we can use it to define). Remember, Anscombe defines her special question ‘Why?’ in terms of a ‘reason for acting’: ‘the sense is of course that in which the answer, if positive, gives a reason for acting.’ (Anscombe, Reference Anscombe2000, §5) As we have argued, this means that the special question is in effect a request or demand for a justifying explanation of a person’s conduct. We argue that this request or demand is ‘granted application’ precisely to the conduct for which we are responsible – that is, answerable – even though we can sometimes respond with an excuse, an apology, or a confession instead. Furthermore, rethinking the scope of Anscombe’s question in this way allows us to explain the vague idea of being ‘granted application’ precisely, by means of the concepts of exemption and denial.

As we have seen, Anscombe’s belief that the scope of the special question ‘Why?’ is restricted to intentional action is partly the result of her adherence to the Aristotelian doctrine that a reason for acting must explain what is ‘useful or attractive’ about the act or activity concerned, which (she believes) excludes various kinds of behaviour for which we are responsible, such as unwittingly poisoning the inhabitants of a building, harming a patient’s immune system, or laughing spontaneously at a joke. And it is supported by her analogy with the ‘How much money?’ question. We have challenged both of these ideas. But whether or not our arguments are accepted, our disagreement with Anscombe is about the nature of ‘reasons for acting’ and the scope and philosophical use of the special question ‘Why?’

The second question was whether our approach represents a significant advance in the theory of responsibility or whether it merely links existing ideas about responsibility with Anscombe’s theory of intention. In our view, it does advance the debate about responsibility, mainly because of the way in which it frames the topic. To see this, take a step back to Anscombe’s theory of intention. Unlike simpler definitions couched in terms of reasons, such as Davidson’s, Anscombe’s definition of intentional action in terms of the special question ‘Why?’ immediately places intentional action in the context of a linguistic practice – in Wittgenstein’s terminology, a language-game. As we have seen, Anscombe summarises her theory with the remark that she has defined intentional action in terms of language – the special question ‘Why?’; preceded, a few paragraphs earlier in the same section, by the following comment:

Thus, in Anscombe’s view, the way to explain ‘what distinguishes actions which are intentional from those which are not’ is not to identify a special kind of thought or feeling that causes or accompanies intentional action; it is to describe the kind of dialogue or exchange – again, the language-game – to which this ‘form of description of events’ belongs, and the phrasal structures in which it occurs.Footnote ³¹

Our approach to the theory of responsibility retains this key feature of Anscombe’s theory of intention. So, in common with many contemporary philosophers, we believe that we need to think about responsibility in the context of a social practice – or various social practices – in order to theorise about it effectively. But the language-game approach is distinctive is several ways, of which two are especially notable, because they set our views apart from those of other philosophers who think of responsibility as answerability.

First, we depart from the widely held view that an agent is only answerable for conduct that reflects their quality of judgement or quality of will.Footnote ³² This is not to say that these kinds of attitudes play no part in the practice of assessing conduct, but they pertain to a person’s culpability rather than their answerability. Thus, ‘I didn’t mean to’, ‘I couldn’t have known’, etc. – pleas that implicitly deny a defective quality of judgement or quality of will – do not deny answerability. On the contrary, they implicitly acknowledge it, but seek to avert or reduce culpability by offering an excuse. Answerability depends on a person’s ability to participate in a language-game that includes demanding and responding to the demand for a justification of our conduct and on their relationship to an interlocuter, rather than on an attitude their conduct may or may not express. For as we have seen, the special question ‘Why?’ is given application in the absence of grounds for an exemption or a denial. The scope of answerability, as we understand it, is therefore not defined by a special kind of attitude that can cause or accompany our conduct, and it is broader than generally supposed.

Second, the responsibility language-game has different variants, including the informal, customary language-game of moral responsibility and the institutionalized and more tightly structured language-games of legal responsibility, which are played out in courts of law. But it is invariably an interpersonal language-game, because moral responsibility relates one person to another – those they are responsible to – while legal responsibility places a defendant in a similar relationship to a plaintiff, a court, or perhaps to their fellow-citizens or peers.Footnote ³³ In Lucas’s phrase, responsibility, in this sense, is a ‘web of dialectical obligation’ (1993, p. 11). Accordingly, it not only relates a person to the conduct and perhaps the attitudes or beliefs for which they are responsible, but also relates them to those others – be they individuals, institutions, or groups – who are owed that responsibility, in other words, those with the standing to request or demand an answer to the special question ‘Why?’ Gardner and Smith both deny this. Gardner defends the claim that ‘basic’ responsibility is the ability to give an account of oneself (Gardner, Reference Gardner, Cruft, Kramer and Reiff2011, Reference Gardner2003), regardless of whether anyone is in a position to demand it. Similarly, Smith claims that ‘questions of basic moral responsibility need to be distinguished from … questions about the standing of particular individuals to take up specific moral responses to others.’ (Smith, Reference Smith2015, p. 108) Perhaps this claim is motivated by the assumption that responsibility is a necessary condition for fault or culpability, since these do not depend on anybody’s standing in the same way.Footnote ³⁴ Be that as it may, our view differs from theirs because we recognise lack of standing as grounds for an exemption from responsibility.

Thus, repurposing Anscombe’s special question ‘Why?’ in the theory of responsibility and adopting Wittgenstein’s language-game methodology, which shaped her theory of intention, illuminates the subtle architecture of exemptions, justifications and excuses, and transforms our conception of responsibility in fundamental ways.Footnote ³⁵