Abstract:

A longstanding debate in the philosophy of action opposes causalists to anti-causalists. Causalists claim the authority of Davidson, who offered powerful arguments to the effect that intentional explanations must be causal explanations. Anti-causalists claim the authority of Wittgenstein, who offered equally powerful arguments to the effect that reasons cannot be causes. My aim in this paper is to achieve a rapprochement between Davidsonian causalists and Wittgensteinian anti-causalists by showing how both sides can agree that reasons are not causes, but that intentional explanations are causal explanations. To this end, I first defuse Davidson’s Challenge, an argument purporting to show that intentional explanations are best made sense of as being explanatory because reasons are causes. I argue that Wittgenstein furnishes anti-causalists with the means to resist this conclusion. I then argue that this leaves the Master Argument for the claim that intentional explanations are causal explanations, but that by distinguishing between a narrow and a wide conception of causal explanation, we can resolve the stalemate between Wittgensteinian anti-causalists impressed by the thought that reasons cannot be causes and Davidsonian causalists impressed by the thought that intentional explanations must be causal explanations.

1. Davidson’s Challenge

Causalists and anti-causalists in the philosophy of action have been opposing each other for decades.[1] One side has claimed the authority of Donald Davidson, who offered powerful arguments to the effect that intentional explanations must be causal explanations—a view that has achieved the status of “a central dogma in many circles within philosophy of mind” (Sehon 2016: 4).[2] The other side has claimed the authority of Ludwig Wittgenstein, who offered equally powerful arguments to the effect that reasons cannot be causes, because reasons are subject to first-person authority in a way in which causes are not; reasons are normally neither physiological processes nor (onsets of) mental states, but what our mental states are about, which makes them unlikely candidates as causes of action; and reasons justify where causes could not.[3]

My aim in this paper is to achieve a rapprochement between Davidsonian causalists and Wittgensteinian anti-causalists by showing how both sides can agree that reasons are not causes, but that intentional explanations are causal explanations. To this end, I begin by defusing what I call Davidson’s Challenge, an argument purporting to show that intentional explanations are best made sense of as being explanatory because reasons can be causes.[4] I argue that Wittgenstein furnishes anti-causalists with the means to resist this conclusion (Section 2). I then argue that this leaves what I call the Master Argument for the claim that intentional explanations are causal explanations, but that by distinguishing between a narrow and a wide conception of causal explanation, we can resolve the stalemate between Wittgensteinian anti-causalists who are impressed by the thought that reasons cannot be causes and Davidsonian causalists who are impressed by the thought that intentional explanations must be causal explanations (Section 3).

Chief among the arguments which Davidson advanced in “Actions, Reasons, and Causes” (2001a), the essay which shattered a longstanding Wittgenstein-inspired consensus on anti-causalism, is an inference to the best explanation: when we ask why someone acted as they did, we do not just want to know a reason to perform the action, but we want to know which reason they actually acted on; and the best account of the difference between these two kinds of reasons is that the latter was causally efficacious in bringing about the action. Davidson thus throws down the gauntlet against anti-causalist accounts of action explanation by challenging them to differentiate between the following two cases:

(1) A has a reason to φ and φ-s for precisely that reason.

(2) A has a reason to φ and φ-s, but does not φ for that reason.

If reasons are to explain actions, it will not suffice to observe that A φ-d and had a reason to φ:

… for a person can have a reason for an action, and perform the action, yet this reason not be the reason why he did it. Central to the relation between a reason and an action it explains is the idea that the agent performed the action because he had the reason. (Davidson 2001a: 9)

What could enable us to differentiate between (1) and (2), Davidson asks, if not the fact that the reason in question was causally efficacious in bringing about A’s φ-ing? He concludes that “failing a satisfactory alternative, the best argument for a [causal] scheme … is that it alone promises to give an account of the ‘mysterious connection’ between reasons and actions” (Davidson 2001a: 11). This is Davidson’s Challenge. It purports to show that intentional explanations are causal explanations because the reasons for which we act are the causes of our actions.

However, Davidson’s later writings suggest that his position is perhaps not as far removed from the Wittgensteinian view of intentional explanation as these historical dialects suggest. Davidson leaves ample room for the idea that intentional explanations are of a markedly different kind from those given in the natural sciences. In “Mental Events” (Davidson 2001d), he makes it clear that no conceptual reduction of intentional to physical explanation is to be had.[5] Elsewhere, he insists that “there is an irreducible difference between psychological explanations that involve the propositional attitudes and explanation in sciences like physics and physiology” (Davidson 2004: 101), and that “the methodology of history (or, for that matter, of any of the social sciences that treat individual human behaviour) differs markedly from the methodology of the natural sciences” (Davidson 2005: 285). Davidson’s granting that “beliefs and intentions are not little entities lodged in the brain” (1999: 654) also indicates his agreement with Wittgenstein in rejecting the hypostatisation of reasons for action as parts of “a hidden machine, say, a machine in [the] brain” (Waismann 1965: 122). Wittgenstein’s anti-causalism forms part of his crusade against what he regards as a misguided tendency to hypostatise psychological phenomena, and while he is adamant that reasons are not causes, he has little to say on whether intentional explanations are a species of causal explanation; and one can deny that reasons are causes, in the strict sense of figuring as relata in causal relations, while endorsing the weaker claim that intentional explanations are a species of causal explanation.[6]

This indicates that a rapprochement between Wittgensteinian anti-causalism about reasons and Davidsonian causalism about intentional explanation may be in the offing. And indeed, I shall argue that by distinguishing a narrow from a wide conception of causal explanation, we can maintain that the reasons for which we act are not causes, but that intentional explanations are causal explanations in virtue of citing features of persons associated with events standing in causal relations to actions. But paving the way for such a rapprochement requires defusing Davidson’s Challenge, since that argument purports to tie the intelligibility of intentional explanation to the causal efficacy of the reasons for which we act. It is therefore to the task of defusing Davidson’s Challenge using Wittgenstein’s anti-causalism that we first turn.

2. A Wittgensteinian Response to Davidson’s Challenge

Davidson’s Challenge to differentiate between a reason to φ and the reason one φ-ed is really two challenges: an epistemic challenge to tell the difference between a reason and the reason, and a metaphysical challenge to answer the constitutive question of what makes a reason the reason for which one acted.[7] In this section, I argue that Davidson’s Challenge fails to settle the argument between the causalist and the anti-causalist, and thus fails to tie the intelligibility of intentional explanation to the causal efficacy of the reasons for which we act, because (a) Wittgenstein provides anti-causalists with the means to resist the challenge, and (b) causalists themselves struggle to spell out how reasons must cause actions if the agent is to act for that reason without presupposing the notion of acting for reasons.

A Wittgensteinian response to Davidson’s Challenge might start by noting, first, that explanations in terms of reasons serve to interpret some bodily movement as constituting a particular kind of action. The paradigmatic situation in which such explanations have a point and are called for is when a deed becomes a riddle to us (Wittgenstein 2009: §79). To be given an explanation in terms of reasons is to be given the “trains of thought” that lead to an action, which enables one to understand the action (Wittgenstein 1982: §§92–94). An intentional explanation thus serves an interpretive function. It “gives the attitudinal conditions in terms of which to derive the understanding of the agent’s behaviour as the act that he performed” (Stoutland 1976: 302). Providing an explanation for an action, and even denying an explanation for an action—“I’m leaving the room, but not because you tell me to” (Wittgenstein 2009: §487)—elucidates the action’s meaning in light of the circumstances of the action and of how it fits within the wider pattern of the agent’s behaviour.

Second, for Wittgenstein, identifying the reason on which someone acted is a matter of the “circumstantial evidence” (2009: §488), of the context in which the action occurs and of how well a given reason helps us make sense of the overall pattern of the agent’s verbal and non-verbal behaviour. This circumstantial evidence includes:

  • what the agent said or did at the time;
  • what she said or did earlier or later, particularly when asked for reasons;
  • what reasons previously weighed with her in comparable circumstances;
  • what we take to be in or out of character for her;
  • what she would have said and done had she been asked what her reasons were;
  • her abilities (such as her ability to speak a language).

Just as context is crucial to ascribing reasons to persons in general, it is crucial to determining the reason on which someone actually acted. Wittgenstein illustrates this in one of his lectures by imagining himself and one of his students, James Taylor, walking along a river, when suddenly, Taylor stretches out his arm and pushes Wittgenstein into the water. Puzzled, Wittgenstein asks Taylor “why he did this” (1966: 22), and two explanations for Taylor’s action are proposed: “(1) He subconsciously hated the other man”, and “(2) He was pointing at something” (1966: 23). To decide between the two explanations, Wittgenstein goes on to argue, we look to the context of the action, that is, to what Taylor himself said was the reason that weighed with him in his deliberation, to how truthful Taylor was known to be about such things, to Taylor’s past behaviour, to the circumstances of the action, and to Taylor’s character; we would declare the second explanation to be correct if “he had never shown any unfriendly feelings, [if] a church-steeple and I were in his field of vision, and Taylor was known to be truthful” (1966: 23).[8] Interpretation is required to get us from mere motion to meaningful movement, and even when we have one description of such a movement (“he stretched out his arm”), that interpretation may not suffice to make sense of the action in the wider context. This is where we ask for explanation, and an intentional explanation provides a redescription of the action by presenting it as the expression of particular purposes and intentions. The function of intentional explanation, on this account, is to interpret deeds, and the correct intentional explanation is that which best makes sense of the overall patterns of the agent’s behaviour and utterances.

Within these patterns, a special role is played on Wittgenstein’s view by what the agent says his or her reasons were. Normally, the agent “cannot be mistaken in specifying his reason … we call the reason that which he gives as his reason” (Wittgenstein & Waismann 2003: 110):

“Why are you turning out the light in your room?” I say: “Because I want to go to sleep.” He asks: “Are you sure?” And I reply: “I must surely know why I am doing it.” This certainty indicates that specifying a reason is the criterion for having this reason. (Wittgenstein & Waismann 2003: 31)

The certainty in question does not reflect one’s privileged epistemic access to one’s practical deliberation, but rather one’s (defeasible) first-personal authority in stating what one’s reason is. As Wittgenstein puts it in the Blue Book, “In order to know the reason which you had for making a certain statement, for acting in a particular way, etc., no number of agreeing experiences is necessary” (1958: 15). In giving one’s reason, one makes an expressive rather than a descriptive move in the language game: one does not report a connection between a reason and the action it is a reason for, or “read it off from some other process which took place then” (Wittgenstein 2009: §637), “on grounds of self-observation,” but rather voices one’s mind and thereby tells others something about oneself “which goes beyond what happened at that time” (2009: §659), namely how the action fits within the larger pattern of one’s thought and behaviour. In this sense, one does not describe, but makes a connection between what one did and why one did it (Wittgenstein 2009: §486, §§682–683).[9]

Of course, one may have reasons for what one did and still be silent as to what these reasons were, or make disingenuous, insincere, or self-deceived statements about what they were. But what then makes it true that one still had a reason for acting as one did, or that one is insincere, disingenuous, or self-deceived in specifying one’s reason, is the wider context of one’s verbal and non-verbal behaviour.[10] These cases are derivative on the basic case in which the reason is “nothing more than just the one [the agent] gives when asked” (Wittgenstein 1979: 5).

This furnishes the anti-causalist with ample means to resists Davidson’s Challenge in both its epistemic and its metaphysical key.[11] Taking up the terminology proposed by Glock (2014: 43), we can distinguish two different strands in this Wittgensteinian response to Davidson’s Challenge:

Deliberationism: what distinguishes the reason on which someone acted from other reasons to act this way is that it is the reason that weighed the most in the agent’s deliberation.

Patternalism: what distinguishes the reason on which someone acted from other reasons to act this way is that it is the reason that best makes sense of the overall pattern of the agent’s verbal and non-verbal behaviour.

On a Wittgensteinian view, these two strategies are of course linked, for one way of spelling out the deliberationist strategy is in terms of patternalism: which reasons weighed the most in the agent’s deliberation is normally, though not necessarily or indefeasibly, determined by the agent’s truthful declarations or avowals of what these reasons were and how these fit into the wider patterns of the agent’s behaviour.

It is straightforward to see how this allows anti-causalists to respond to Davidson’s Challenge in its epistemic key. One finds out what the agent’s reason was by looking at what she said or did at the time, at what she said or did earlier or later, particularly when asked for reasons, and at what reasons previously weighed with her in comparable circumstances, and assesses these things against the background of what one knows to be in or out of character for this person.

As for how anti-causalism fares with Davidson’s Challenge in its metaphysical key, the question is whether the difference between a reason there is for A to φ and the reason A φ-s can only be understood in terms of there being a causal connection between reason and action. The challenge is that of saying what the difference consists in rather than that of telling the difference—to spell out what makes true a claim of the form “A acted for this reason and not for that one.” The point of the challenge, from the causalist’s perspective, is that if we are to distinguish metaphysically between the reason for which one does something and a reason to do something, which in practice we are perfectly able to do, there must be a difference between the two; and this difference can, according to the causalist, lie only in the idea that one is causally efficacious while the other is not.

Yet on a Wittgensteinian understanding of what we are at when we give reasons, this is a conclusion the anti-causalist can resist. To give the reason for which one acted is to indicate a deliberative route, but not in the sense of specifying the process by which, causally, one in fact got to the action in virtue of the reason’s causal efficacy; rather, it shows how, rationally, one gets to the action in virtue of the rational connections the consideration cited bears to other considerations. It is only if reasons were mental states that both rationally and causally accounted for an action that giving one’s reasons for a past action would be equivalent to describing past psychological events or states of the agent. But as Wittgenstein insists, it is only in “some cases” that giving a reason means “telling the way which one has gone oneself” (1958: 14); crucially, it can also mean “describing a way which leads there and is in accordance with certain accepted rules” (1958: 14), but which one has not gone oneself. In these latter cases, giving reasons does not consist in reporting or recapitulating some psychological process of reasoning one has gone through. There are cases in which we act without consciously reflecting on our reasons, but where we can nevertheless be said to act for reasons. Hence, what distinguishes the reason for which one acts from other reasons need not be its position in a network of causes. It might instead be its salient position in the pattern of the agent’s reasons as a whole. The agent’s reason stands out as the reason for the action not in the mechanical sense in which it causally accounts for what is done, but in the normative sense in which in her view, it trumps other reasons, and thereby rationally accounts for the action as what, on her judgment and all things considered, is to be done. The relativisation to the agent’s perspective is required if agents are not to be conceived of as unrealistically rational. If what reasons have weight from the agent’s perspective is determined by what the agent, in practice, implicitly treats as a reason and by what reasons she would give if asked (rather than by what reasons became explicit as occurrent thoughts in the agent’s practical deliberation), this relativisation does not conflict with Wittgenstein’s claim that the agent need not have her reasons in mind in order to act on them.[12]

The conclusion one reaches is that Wittgenstein gives anti-causalists the means to answer Davidson’s Challenge in both its epistemic and its metaphysical key. No appeal to causation is required either to determine that someone acted for one reason and not another, or to spell out what makes it the case that this is so.

What this Wittgensteinian response to Davidson’s Challenge brings out is that the two camps confronting each other in the causalism/anti-causalism debate are opposed in both temperament and method. Where the causalist’s guiding intuition is to explain action by zooming in on the agent and the mechanisms underlying her movements, the anti-causalist’s intuition is to zoom out to bring the larger patterns of behaviour into view. Where the causalist seeks to uncover the action’s triggers by analysing it into particular events, the anti-causalist seeks to explain it by striving for a synthetic understanding of the amalgam of human interaction.[13] With a nod to Wittgenstein, Davidson in fact concedes that when we explain an action by reference to reasons, we redescribe the action and place it in a pattern. Yet he insists that there is no conflict between the two strategies. He points out that “events are often redescribed in terms of their causes” (Davidson 2001a: 10). If someone was injured, for example, we could “redescribe this event ‘in terms of a cause’ by saying he was burned” (Davidson 2001a: 10). In other words, a description which elucidates the meaning of an action can do so by reference to the causes of the action. Davidson’s success in reinstating causalism may well derive in part from his conciliatory assent to most of what anti-causalists have to say about the interpretive function of intentional explanation. In what looks like a synthesis, he takes anti-causalist insights on board while insisting that they form no obstacle to intentional explanations’ being causal explanations.

A possible line of objection to this Wittgensteinian response to Davidson’s Challenge is indicated by William Child:

The fact that we can effect a distinction by using a set of criteria which are not explicitly causal is clearly compatible with the distinction’s being a distinction between things with a causal history of one sort and those with a causal history of another. For example, I can tell the difference between the song of a garden warbler and the song of a blackcap by applying a test which is not explicitly causal; how sustained is the song, and what is its pitch? But the distinction thus effected is a distinction between songs with different aetiologies; the one is produced by garden warblers, the other by blackcaps. (1994: 96–97)

Child adds that it is open to the causalist to insist that the considerations used to draw the distinction are in fact causal, so that the argument begs the question against the causal view in any case (1994: 97). This may well be the case, but the causalist begs the question in just the same measure by arguing merely that our effecting the distinction without causal criteria is compatible with the distinction’s being one between things with different aetiologies. Child’s argument might establish that different aetiologies are possible despite the fact that we draw the distinction without resorting to them, but what the causalist needs to demonstrate is that drawing the distinction requires the aetiologies to be different. As long as both causalist and anti-causalist interpretations are compatible with the data, but not entailed by it, Davidson’s Challenge fails to decide the issue.

It would in any case be imprudent to pin too much of one’s faith in causalism on the possibility of differentiating between a reason to φ and the reason one φ-ed in causal terms. A decade after the publication of “Actions, Reasons, and Causes,” Davidson himself admitted that the causalist account led him to “despair of spelling out … the way in which attitudes must cause actions if they are to rationalize the action” (2001c: 79). The problem, originally pointed out by Anthony Kenny (1975), and which Davidson deemed “insurmountable,” lies in the possibility of “wayward causal chains,” which invalidate the conclusion that “if attitudes that would rationalize x cause an agent to do x, then he does x intentionally” (Davidson 2001c: 79). In other words, the causalist cannot distinguish between the following two cases:

(3) A has a reason to φ and φ-s for that reason.

(4) A has a reason to φ and φ-s because of that reason.

(3) is the well-behaved case, in which the attitudes that rationalise φ-ing cause the agent to φ, and the agent φ-s intentionally, that is, for that reason. But the causalist account is powerless to distinguish such right causal chains from wrong causal chains as typified by (4). In these wayward cases, the combination of beliefs and desires (what Davidson calls a “primary reason”) both rationalises and causes the action, and yet the action is not done intentionally, that is, it is done only because of that reason, but not for that reason. Davidson provides the following example:

A climber might want to rid himself of the weight and danger of holding another man on a rope, and he might know that by loosening his hold on the rope he could rid himself of the weight and danger. This belief and want might so unnerve him as to cause him to loosen his hold, and yet it might be the case that he never chose to loosen his hold, nor did he do it intentionally. (2001c: 79)

The conclusion many, including Davidson himself, have drawn from this is that there is no way of spelling out the “right way” in which reasons must cause actions if the agent is to act for that reason without presupposing the notion of acting for reasons.[14] The promise of a causalist “account of the ‘mysterious connection’ between reason and action” (Davidson 2001a: 11) so far remains unfulfilled.

Moreover, as Wittgenstein also remarks, there may be two reasons which could, with equal justification, be said to be the ones on which the agent acted, and “both explanations may be correct” (1966: 23). Wittgenstein gives the example of someone whose action could be rationalised either by a conscious reason or by an unconscious one, even if the two explanations came into conflict with one another: “The explanations could in a sense be contradictory and yet both be correct … One could be love and one could be hatred” (1966: 23). In view of these reminders of the way we normally go on, and which philosophy tends to dispose us to forget, the insistence that there needs to be a single, causally efficacious reason appears as a theoretical artefact, generated perhaps by the pressure on the causalist account to identify the salient cause of the action. When we see our actions in a way uncorrupted by the theoretical oversimplifications of philosophy, Wittgenstein contends, there is no issue with the fact that there may be language games which function by deploying two utterly different but equally operative motives (1966: 23).[15]

The upshot is perhaps best summed up by saying that Davidson’s Challenge is effective primarily as a way to preach to the converted, but that it fails to tie the intelligibility of intentional explanation to the causal efficacy of the reasons for which we act, because Wittgenstein provides anti-causalists with the means to resist the challenge, and causalists themselves struggle to spell out how reasons must cause actions if the agent is to act for that reason without presupposing the notion of acting for reasons.

3. A Rapprochement between Davidson and Wittgenstein

While Wittgenstein’s writings contain the seeds of arguments to resist Davidson’s Challenge itself, one should not lose sight of the fact that they do little to answer what might be called the Master Argument for causalism which provides the background rationale for Davidson’s advocacy of causalism about intentional explanation. The linchpin for this argument is that intentional explanations are still explanations of why something happened—as opposed to explanations of why something counts as something else or has a particular property, such as why parking in such-and-such a way is illegal (Child 1994: 91).[16] The Master Argument for a causalist conception of intentional explanation then runs as follows:

An intentional explanation is an explanation of why something happened.

No non-causal explanation can explain why something happened.

Therefore, intentional explanations must be causal explanations.

In order to resist the conclusion, anti-causalists about intentional explanation might deny either of the premises. If they choose to deny the first premise, the burden falls on them to show what the explanatory import of intentional explanations is: intentional explanations may help us grasp the meaning or the significance of an action, but how can intentional explanations really explain why A φ-d if they fail to explain why the action occurred? If anti-causalists deny the Master Argument’s second premise, they can reverse the argument and reject it as follows: intentional explanations are not causal; they do explain why something happened; therefore, some explanations of why something happened are not causal. But in response, causalists can insist that for any supposedly non-causal explanation of why something happened, however informative, we can still ask why the particular event in question occurred when and how it did, and if the event has a cause, the answer will, if only indirectly, have to provide information about its causal history, and will therefore be a causal explanation.[17] If anti-causalists want to deny that intentional explanations are causal, therefore, it seems that they must deny that intentional explanations are explanations of why something happened.

If anti-causalism about intentional explanation comes at the cost of denying this prima facie plausible assumption, it is worth asking whether there is a textual basis to the conclusion that Wittgenstein himself goes down this path, or whether he is best seen as taking a rather more noncommittal position which leaves room for a conciliatory view of intentional explanation. Wittgenstein is clearly rightly claimed as an authority by anti-causalists about reasons, who maintain that reasons are not a species of causes.[18] But this leaves open the question whether he is also rightly claimed as an authority by anti-causalists about intentional explanation. As Alvarez (2007: 105) notes, one can deny that reasons are causes, in the strict sense of figuring themselves as relata in causal relations, while endorsing the weaker claim that intentional explanations are a species of causal explanation.

There can be little doubt that Wittgenstein’s work brings out a profound difference between intentional explanations and the causal explanations of physics. He emphasises that to regard bodily movements not under their mechanical aspect, but under their aspect as reason-guided actions, involves viewing them in the context of rule-governed practices and institutions (Waismann 1965: 124; Wittgenstein 2009: §§197–199). This emerges from the combination of the following two ideas: on the one hand, as we have seen, Wittgenstein contends that the reasons that explain and determine the meaning of an action are given, and therefore constrained, by the way in which the action fits into the larger pattern of the agent’s character and history: they must cohere with the context of the action, with the agent’s personality and with her past and subsequent behaviour (on pain of either being unintelligible or of casting doubt on her sincerity or self-knowledge). On the other hand, he takes the norms of reasoning which delineate what rational connections a reason stands in to be context-bound: they depend on the language game, that is, on the interactive and rule-guided complex of activities and language-use within which something is adduced as a reason: “Not until there is a language game are there reasons” (Wittgenstein 1980: §689), and a “reason is a reason only inside the game” (Wittgenstein 1979: 4). Taken together, these two ideas imply that if what action a movement constitutes depends on the reasons that explain it, and if what reasons can explain it in turn depends on our practices and institutions, then the characterisation of movements as actions will be as context-bound as the characterisation of propositions as reasons. If we want to say, as Stoutland notably does, that “what constitutes an agent’s bodily movements as intentional under a description is their being explained by the agent’s reasons for acting under that description” (2010: 56),[19] then the horizon of possible actions will in part be set by the horizon of possible reasons for action, which is in turn set by the context of human customs and institutions:

An intention is embedded in a setting, in human customs and institutions. If the technique of the game of chess did not exist, I could not intend to play a game of chess. (Wittgenstein 2009: §337)

The institution of chess is required in order to get from the mere movement to the move within the game. Likewise, making a mark on a piece of paper can be described physically or physiologically, but whether making a mark on a piece of paper is writing one’s name, and whether writing one’s name is signing a letter, a cheque, a contract, or a will, depends on the social customs and institutions it is embedded in. Absent these customs and institutions, nothing would count as performing these actions, no matter what went on in the mind or brain of the agent (Hacker 2011: 69). Thus, a key characteristic of intentional explanations which is alien to the causal explanations of physics is that it refers to rule-governed practices and institutions, and thus to the agents’ local perspectives and idiosyncrasies. By contrast, the hard sciences’ causal explanations of the behaviour of nature are not context-bound in this way. They strive for a representation of the world “which is to the largest possible extent independent of the local perspectives or idiosyncrasies of enquirers” (Williams 2006: 184).

Yet even granted this difference, it does not follow that intentional explanations are not a kind of causal explanation. Whether they are or not depends largely on how wide a notion of causal explanation one has in mind; and this brings into view the possibility of holding on to the anti-causalist insights suggesting that intentional explanations are sui generis while broadening and shoring up one’s conception of causal explanation to make room for a sui generis mode of causal explanation. In the remainder of this paper, I sketch out in barest outline how such a rapprochement between Wittgensteinian anti-causalism and Davidsonian causalism can be achieved.

Following Strawson (1992b) and Davidson (2001b), we can distinguish between (i) the network of events or circumstances standing in extensional, natural relations of cause and effect, and (ii) the network of descriptions standing in intensional, non-natural relations of explanans and explanandum. While causal relations are extensional and hold between events independently of how they are described, relations of explanation (causal or otherwise) hold between events as described, which means that an event described in one way may provide an explanation for another, while the same event described in another way fails to do so. It may be that a certain event e1 caused a certain event e2, but it is the fact that e1 occurred which explains the fact that e2 occurred.[20] Consider the following example:

(5) Turing died because he bit into an apple.

(6) Turing died because he ingested cyanide.

While (5) and (6) refer to the same two causally related events, only (6) causally explains Turing’s death, while (5) does not. The explanatory force of causal explanation depends on the right description of explanans and explanandum. Moreover, the explanatory power of causal explanations derives from and depends on the obtaining of causal relations between events: it is because the opening of the fridge door caused the ice to melt that the fact that the fridge door was opened explains the fact that the ice melted and thus forms the reason why it happened. Behind every causal explanation there are two causally related events.

But this still allows us to distinguish between a narrow and a wide conception of causal explanation, and using this distinction, we can resolve the stalemate between anti-causalists who are impressed by the thought that reasons cannot be causes and causalists who are equally impressed by the thought that intentional explanations must be causal explanations.

On the narrow conception, an explanation is a causal explanation just in case it explains why something happened and the fact acting as explanans either directly names or points to the event that stands in a causal relation to the explanandum. For example, if the fact that the temperature dropped explains the fact that the ice melted, the two causally related events—the drop and the melting—are explicitly mentioned. In other cases, the explanation might mention a preceding step in the causal chain, such as the fridge door’s being opened (which points to the drop in temperature by reference to its cause). But as long as the notion of causal explanation involves the requirement that it more or less directly—we can think of the distinction as graded rather than sharp—name the causal relata at issue, denying that the reasons for which we act can themselves figure as such relata will bar one from thinking of intentional explanations as causal explanations.

On the wide conception, by contrast, an explanation will still count as a causal explanation if it explains why something happened and the fact acting as explanans only mentions features of the situation which are relevant by being associated with the cause. This wider conception of causal explanation thus covers what Child calls “feature-citing explanations” (1994: 103). For example, the fact that the driver was drunk explains the crash, and does so even if one has no knowledge of the proximate cause of the crash. Similarly, mentioning a property of something can figure in a causal explanation without our having to think of the property itself as a cause:

Suppose I strike this glass a smart blow, and it breaks. We can explain its breaking by saying that it was fragile. The cause of the glass’s breaking was its being struck. “It broke because it was fragile” is a causal explanation, whose truth depends on the obtaining of a causal relation—the causal relation between the striking and the breaking. (Child 1994: 125)

In this example, the explanans does not explicitly mention an event, but a property of the glass—its fragility. The item which is causally related to the action and the item which causally explains it fall apart. Yet one still explains why something happened—why the glass broke—by making it intelligible why its being struck caused its breaking.

It seems to me, therefore, that one must refrain from concluding, as some have done (Hacker 2011: 69), that Wittgenstein denies that intentional explanations are a species of causal explanations. If one allows for the fact that the item which is causally related to the action and the item which causally explains it can fall apart, then intentional explanations can still be a form of causal explanation. Intentional explanations may be sui generis, as Wittgenstein intimated; but one way in which they can be sui generis is by being a sui generis mode of causal explanation. Nothing Wittgenstein says bars intentional explanations from being a species of causal explanations.

On this wider conception of causal explanation, the possibility of a rapprochement between Wittgenstein and Davidson emerges, achievable by combining the view that reasons are not causes with the view that intentional explanations are causal explanations. We can maintain, with Wittgenstein, that the reasons for which we act are not causes, because they are the contents of our propositional attitudes—what we believe or what we desire (Glock 2014; Queloz 2017); but if feature-citing explanations can be causal explanations without our having to think of the features themselves as causes, then so can explanations in terms of propositional attitudes, and we can therefore also maintain, with Davidson, that intentional explanations are causal explanations. Facts about propositional attitudes—not about their content, but about the attitudes of believing or desiring themselves, such as that A believes that p or that A desires that q—can figure in causal explanations of actions on the assumption that they are associated with events standing in causal relations to the action to be explained.

On this conciliatory view, intentional explanations still count as a form of causal explanation because facts about propositional attitudes have causal import. Mentioning features or properties of persons, such as the fact that Taylor believes that the steeple would please Wittgenstein and desires that he be pleased, would still causally explain the fact that Taylor pointed to it by rendering it intelligible why his noticing the steeple caused him to stretch out his arm.

In mentioning propositional attitudes as parts of a causal explanation, we need not take ourselves to be “limning the units that a future neuroscience will use in identifying the causes of behaviour” (Robinson 1990: 51). As Davidson writes, “it is changes in the attitudes, which are events, which are the often unmentioned causes” (1993: 288). But he adds that beliefs and desires should not be thought of as “little entities lodged in the brain,” and therefore,

since beliefs and desires aren’t entities, it is a metaphor to speak of their changing, and hence an extension of that metaphor to speak of them as causes and effects. What happens is that the descriptions of the agent change over time. The relevant entity that changes is the person, and there seems no difficulty in supposing that these changes have a physical description. (Davidson 1999: 654)

We can still acknowledge the profound differences between intentional explanations on the one hand and causal explanations of the kind advanced in neurophysiology on the other. Indeed, it would be surprising if there were no such differences, since they reflect the fact that intentional explanations are tailored to our needs in everyday social interaction, and that they pick out the features of persons and their behaviour which are of most interest to us in these contexts. In another context, such as a neurophysiological investigation, other features are of interest, and it makes sense that we should advance causal explanations of an entirely different kind there. In each case, we turn our inquiries around “on the pivot of our real need” (Wittgenstein 2009: §108). But acknowledging these differences in kinds of explanations and the needs they answer to falls short of denying that intentional explanations are explanations of why something happened. There seems to be room, therefore, to bring together Davidsonian causalism about intentional explanation and Wittgensteinian anti-causalism about reasons in a way that does justice to insights on both sides.

4. Conclusion

I have been arguing for three claims in this paper. First, Wittgenstein furnishes anti-causalists with the means required to answer Davidson’s Challenge and to resist the idea that the intelligibility of intentional explanation depends on reasons being causes. Second, while Wittgenstein offers us reasons to think that intentional explanations are sui generis, these reasons fall short of entailing that intentional explanations cannot be a species of causal explanation. And third, in response to the Master Argument to the effect that intentional explanations must be causal explanations of some kind, we can resolve the stalemate between Wittgensteinian anti-causalists and Davidsonian causalists by distinguishing between a narrow and a wide conception of causal explanation.

On the wide conception of causal explanation, a rapprochement between Wittgenstein and Davidson can be achieved by maintaining that the reasons for which we act are not causes, but that intentional explanations are causal explanations in virtue of citing features of persons associated with events standing in causal relations to actions. This, to be sure, is no more than a sketch, and a far more detailed treatment of Wittgenstein’s and Davidson’s positions than I have room for here would be required to flesh out the precise extent to which they are compatible. What even a cursory treatment of these two figureheads of causalism and anti-causalism reveals, however, is that they yield much less of a basis for disagreement than historical dialectics suggest—indeed, that causalism and anti-causalism are profitably seen as complementing rather than contradicting one another.

Acknowledgements

Thanks to the editors of Ergo and to three anonymous referees for helpful comments on this paper. Thanks also to Hans-Johann Glock, Joachim Schulte, Damian Cueni, and Adrian Häfliger. This paper was written with the support of grant P0BSP1_162025 of the Swiss National Science Foundation.

References

  • Aguilar, Jesús, H. and Andrei A. Buckareff (Eds.) (2010). Causing Human Actions: New Perspectives on the Causal Theory of Action. MIT Press. https://doi.org/10.7551/mitpress/9780262014564.001.0001
  • Alvarez, Maria (2007). The Causalist/Anti-Causalist Debate in the Theory of Action: What It Is and Why It Matters. In Anton Leist (Ed.), Action in Context (103–123). De Gruyter. https://doi.org/10.1515/9783110898798.103
  • Alvarez, Maria (2010). Kinds of Reasons: An Essay in the Philosophy of Action. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199550005.001.0001
  • Bishop, John (1989). Natural Agency: An Essay on the Causal Theory of Action. Cambridge University Press.
  • Blackburn, Simon (1998). Ruling Passions. Oxford University Press.
  • Candlish, Stewart and Nic Damnjanovic (2013). Reasons, Actions, and the Will: The Fall and Rise of Causalism. In Michael Beaney (Ed.), The Oxford Handbook of the History of Analytic Philosophy (689–708). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780199238842.013.0023
  • Child, William (1994). Causality, Interpretation, and the Mind. Clarendon Press, Oxford University Press.
  • D'Oro, Giuseppina and Constantine Sandis (Eds.) (2013). Reasons and Causes: Causalism and Anti-Causalism in the Philosophy of Action. Palgrave Macmillan.
  • Davidson, Donald (1993). Replies. In Ralf Stoecker (Ed.), Reflecting Davidson: Donald Davidson Responding to an International Forum of Philosophers. De Gruyter.
  • Davidson, Donald (1999). Replies. In Lewis E. Hahn (Ed.), The Philosophy of Donald Davidson. Open Court.
  • Davidson, Donald (2001a). Actions, Reasons, and Causes. In Essays on Actions and Events (3–20). Clarendon Press.
  • Davidson, Donald (2001b). Causal Relations. In Essays on Actions and Events (149–162). Clarendon Press. https://doi.org/10.1093/0199246270.003.0007
  • Davidson, Donald (2001c). Freedom to Act. In Essays on Actions and Events (73–82). Clarendon Press. https://doi.org/10.1093/0199246270.003.0004
  • Davidson, Donald (2001d). Mental Events. In Essays on Actions and Events (205–224). Clarendon Press. https://doi.org/10.1093/0199246270.003.0011
  • Davidson, Donald (2001e). Psychology as Philosophy. In Essays on Actions and Events (229–244). Clarendon Press. https://doi.org/10.1093/0199246270.003.0012
  • Davidson, Donald (2004). Problems of Rationality. Clarendon Press. https://doi.org/10.1093/0198237545.001.0001
  • Davidson, Donald (2005). Truth, Language, and History. Clarendon Press. https://doi.org/10.1093/019823757X.001.0001
  • Glock, Hans-Johann (2014). Reasons for Action: Wittgensteinian and Davidsonian Perspectives in Historical, Meta-Philosophical and Philosophical Context. Nordic Wittgenstein Review, 3(1), 7–46.
  • Hacker, Peter Michael Stephan (2009). Agential Reasons and the Explanation of Human Behaviour. In Constantine Sandis (Ed.), New Essays on the Explanation of Action (75–93). Palgrave Macmillan. https://doi.org/10.1057/9780230582972_6
  • Hacker, Peter Michael Stephan (2011). Wittgenstein and the Autonomy of Humanistic Understanding. In Richard Allen and Malcolm Turvey (Eds.), Wittgenstein, Theory and the Arts (39–74). Routledge & Kegan Paul.
  • Hornsby, Jennifer (2004). Agency and Actions. In John Hyman and Helen Steward (Eds.), Agency and Action (1–24). Cambridge University Press. https://doi.org/10.1017/CBO9780511550843.002
  • Hurley, Susan Lynn (1989). Natural Reasons: Personality and Polity. Oxford University Press.
  • Kenny, Anthony John Patrick (1975). Will, Freedom, and Power. Blackwell.
  • Lewis, David (1986). Causal Explanation. In Philosophical Papers, Vol. 2 (214–240). Oxford University Press.
  • Mele, Alfred (1992). Springs of Action: Understanding Intentional Behavior. Oxford University Press.
  • Mele, Alfred (2003). Motivation and Agency. Oxford University Press. https://doi.org/10.1093/019515617X.001.0001
  • Queloz, Matthieu (2016). Wittgenstein on the Chain of Reasons. Wittgenstein-Studien, 7(1), 105–130. https://doi.org/10.1515/witt-2016-0108
  • Queloz, Matthieu (2017). Two Orders of Things: Wittgenstein on Reasons and Causes. Philosophy, 92(3), 369–397. https://doi.org/10.1017/S0031819117000055
  • Robinson, William (1990). States and Beliefs. Mind, 99(393), 33–51. https://doi.org/10.1093/mind/XCIX.393.33
  • Sandis, Constantine (2011). The Things We Do and Why We Do Them. Palgrave Macmillan.
  • Sandis, Constantine (Ed.) (2009). New Essays on the Explanation of Action. Palgrave Macmillan. https://doi.org/10.1057/9780230582972
  • Schroeder, Severin (2001). Are Reasons Causes? A Wittgensteinian Response to Davidson. In Severin Schroeder (Ed.), Wittgenstein and Contemporary Philosophy of Mind (150–170). Palgrave.
  • Sehon, Scott (2016). Free Will and Action Explanation: A Non-Causal, Compatibilist Account. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198758495.001.0001
  • Stoutland, Frederick (1976). The Causation of Behaviour. Acta Philosophica Fennica, 28(1–3), 286–325.
  • Stoutland, Frederick (2010). Reasons and Causes. In Pasquale Frascolla, Diego Marconi, and Alberto Voltolini (Eds.), Wittgenstein: Mind, Meaning and Metaphilosophy (46–66). Palgrave Macmillan.
  • Strawson, Peter Frederick (1992a). Analysis and Metaphysics: An Introduction to Philosophy. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198751182.001.0001
  • Strawson, Peter Frederick (1992b). Causation and Explanation. In Analysis and Metaphysics: An Introduction to Philosophy (109–132). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198751182.003.0009
  • Tripodi, Paolo (2015). Davidson and the Wittgensteinians on Reasons and Causes. In Danièle Moyal-Sharrock, Volker Munz, and Annalisa Coliva (Eds.), Mind, Language & Action: 36th International Wittgenstein Symposium (137–156). De Gruyter. https://doi.org/10.1515/9783110378795.137
  • Waismann, Friedrich (1965). The Principles of Linguistic Philosophy. Macmillan.
  • Williams, Bernard (2002). Truth and Truthfulness: An Essay in Genealogy. Princeton University Press.
  • Williams, Bernard (2006). Philosophy as a Humanistic Discipline. In Adrian William Moore (Ed.), Philosophy as a Humanistic Discipline (180–199). Princeton University Press.
  • Wittgenstein, Ludwig (1958). The Blue and Brown Books: Preliminary Studies for the Philosophical Investigations. Blackwell.
  • Wittgenstein, Ludwig (1966). Lectures and Conversations on Aesthetics, Psychology and Religious Beliefs: Compiled from Notes Taken by Yorick Smythies, Rush Rhees and James Taylor. Blackwell.
  • Wittgenstein, Ludwig (1979). Wittgenstein's Lectures: Cambridge 1932-35. (from the Notes of Alice Ambrose and Margaret Macdonald). Blackwell.
  • Wittgenstein, Ludwig (1980). Remarks on the Philosophy of Psychology (Vol. 2). Blackwell.
  • Wittgenstein, Ludwig (1982). Last Writings on the Philosophy of Psychology (Vol. 1). Blackwell.
  • Wittgenstein, Ludwig (1998). Culture and Value (2nd ed.). Blackwell.
  • Wittgenstein, Ludwig (2009). Philosophical Investigations. Wiley-Blackwell.
  • Wittgenstein, Ludwig, and Friedrich Waismann (2003). The Voices of Wittgenstein: The Vienna Circle. Routledge.

Notes

    1. See Alvarez (2007), Sandis (2009), Candlish and Damnjanovic (2013), and D’Oro and Sandis (2013) for overviews of the debate.return to text

    2. See Mele (1992; 2003), Bishop (1989), and the essays in Aguilar and Buckareff (2010) for nuanced versions of causalist views.return to text

    3. Queloz (2017); see also Schroeder (2001), Hacker (2009), and Glock (2014).return to text

    4. I thereby pursue the line of thought which Glock (2014: 42–43), at the end of a comparative study of Wittgenstein and Davidson on reasons for action, identifies as requiring further investigation. Tripodi (2015) pursues a similarly reconciliatory strategy from the other direction, as it were, by arguing that Davidson’s conception of intentional explanation is misleadingly described as a causalist one.return to text

    5. This reflects Davidson’s belief in the anomalism of the mental: while causal relations are nomological in character, i. e., fall under strict laws, there are no strict laws relating mental and the physical events. In “Mental Events,” he aims to show how this is compatible with the idea that at least some mental events interact with physical events.return to text

    6. As Alvarez (2007: 105) and Glock (2014) have noted.return to text

    7. One may doubt whether this distinction is available to interpretationists like Davidson who take the best guide to the ontology or nature of the mental to be the epistemology of the mental. Yet even on an interpretationist account of the mental, we can still distinguish between how we know someone is in a mental state and what it is to be in that state.return to text

    8. Wittgenstein also remarks that it is in view of the larger context that an explanation in terms of unconscious reasons would be corroborated: namely if it “often happened” that when one person “was obviously pointing out something and pushed the other in the river,” “the person pushed in had a similarity with the father of the other person” (1966: 22–23).return to text

    9. For a more extensive discussion of Wittgenstein’s views on agents’ first-personal authority in giving reasons, see Queloz (2017).return to text

    10. It is true that on this kind of patternalist account, there would be a point at which, if the agent were shifty and changeable enough—in the way that the eponymous character in Diderot’s Rameau’s Nephew is, for example (Williams 2002: Chapter 8)—the larger patterns of the agent’s behaviour would cease to cohere sufficiently for there to be a fact of the matter as to what the agent’s reason was. But the same is true on Davidson’s view: “if we are intelligibly to attribute attitudes and beliefs, or usefully to describe motions as behaviour, then we are committed to finding, in the pattern of behaviour, belief and desire, a large degree of rationality and consistency” (2001e: 237).return to text

    11. A conclusion endorsed also by Schroeder (2001).return to text

    12. I am grateful to an anonymous reviewer for pressing me on this.return to text

    13. The Wittgenstein-inspired expressivist approach to value is described by Simon Blackburn along these lines: “To understand the value of a piece of money it is no good staring at it. It is necessary to understand the processes of human economic behaviour. You need to approach the token not with a microscope and a scalpel, but with an eye for large patterns of human interactions” (1998: 50).return to text

    14. See Davidson (2001e: 232–233) and Child (1994: 98).return to text

    15. As Wittgenstein writes in Culture and Value, the only way “for us to avoid prejudice or vacuity in our claims” is to take any single model in terms of which we seek to understand a phenomenon “as an object of comparison—a measuring rod as it were—within our way of looking at things,” and “not as a preconception to which everything must conform” (1998: 30). To do otherwise is to fall victim to “the dogmatism into which philosophy can so easily degenerate” (Wittgenstein 1998: 30).return to text

    16. Which is not to say that everybody agrees that intentional explanations are explanations of why something happened. Dissenters include Hornsby (2004) and Sandis (2011).return to text

    17. See Lewis (1986) and Child (1994: 91–93) for further defence of the view that all explanations of events are causal explanations. I am indebted to an anonymous reviewer for comments on this.return to text

    18. Queloz (2016; 2017), Schroeder (2001), Hacker (2009), and Glock (2014).return to text

    19. See also Hurley (1989: 97–98).return to text

    20. Strawson (1992a: 112–113); Alvarez (2010: 29–30); Davidson (2001b: 151).return to text