1 Introduction

We have twin starting points. We wish to examine the idea that actions of moral value are caused by virtues cultivated by a person over time. This is a notion that not only has relevance to popular thinking about virtues but also, within academic philosophy, has currency given the renaissance this century and earlier of the more central position that virtues in the Aristotelian tradition have played in morality (see, for example, Van Zyl, 2019 and Hursthouse, 1999).

Alongside that theme concerning virtues, we wish to work with the idea that a person’s self-narrative constitutes the self and causes action. The notion of the narrative self appears in various locations, including psychological work such as Sacks (1985, 105) who says, ‘Each of us is a singular narrative.’ The notion also appears in a variety of philosophical work (cf. Ricoeur, 1992, Dennett, 1992, Schechtman, 1996). A notable view relating to both virtue and narrative and the connection between them is Alisdair MacIntyre’s (1984). He says that human actions in general are enacted narratives (MacIntyre, 1984, 210–211). In this way, narratives can be seen as being continuous with life (see Carr, 1986, 16 and Lumsden 2013, 8). That is not our emphasis here as our focus is on self-narratives as internally represented constructs that cause actions, similar to how we are regarding virtues. Our focus is on how virtues and self-narratives, while apparently very different from each other, have something of a common nature so that each can each be said to cause actions, in a parallel way.

Our interest in virtues and self-narratives as causes of action may be located in a broader issue of how thinking about and understanding people and their thoughts and behaviour is related to scientific descriptions and explanations of human cognition and behaviour. One way of articulating the broad division between the two approaches is to be found in Wilfrid Sellars’ distinction between the scientific and the manifest image (Sellars, 1963, passim). At the centre of the manifest image (‘of the man-in-the-world’) is the notion of a person who acts for reasons and whose actions can be evaluated, the way that we normally think of ourselves and others.Footnote 1 In contrast, the scientific image is composed of theoretical constructions that can involve the introduction of new basic items to explain the data such as the concepts one employs in physiology or neuroscience (cf. Thompson, 2008, especially Parts I and II).

The virtues are clearly part of the manifest image and self-narratives, while perhaps less part of common parlance, have a kinship with the folk understanding that is characteristic of the manifest image. We believe it is natural to conceive of virtues and self-narratives as part of the causal explanation of action but it is often assumed the proper location of the notion of causation lies in the scientific image. To effect a reconciliation between the two concepts, we draw on the leading class of theories of action, the causal theories of action. In this approach, reasons are shown to be causes of action in a way that respects the principle that a particular causal relationship is underpinned by a scientific law. It may be, though, as Chomsky suggests, that such laws are destined to be forever elusive. We shall begin in Section 2 with an exploration of the main features of Donald Davidson’s causal theory of action in the context of alternative approaches to mind and action. Then, we shall proceed in Section 3 to briefly outline Noam Chomsky’s distinction between ‘problems’ and ‘mysteries’ and his view that the explanation of individual actions is liable to be a mystery. This puts pressure on the causal theory with which we are familiar and casts a new light on our project. In Section 4, we say a little more about virtues and self-narratives, relating them to habits and relating them to each other. Adopting a mild realist approach to virtues and self-narratives in the style of Dennett (1987, 1991) helps the position (cf. Lumsden and Ulatowski 2017, 2019, 2021, Forthcoming; Ulatowski, 2021). In Section 5, we explain how full conscious awareness of virtues and self-narratives is not generally required for them to be causally significant. Next, in Section 6, we sharpen our focus on why we should accept the causation of actions by virtues and narratives, even taking into account Chomskian mysterianism. Finally, Section 7 is a brief conclusion.

2 Theories of Mind and the Causal Theory of Action

One position on the mind–body relationship in which reasons can be seen as straightforwardly capable of causing actions is the position known as ‘the identity theory’ or ‘reductive materialism’ (Lewis, 1966; Place, 1956; Smart, 1959). On this view, mental state types are thought to be identical with neuroscientific or physical state types. Thus, beliefs, desires, reasons, and goals are all thought to be ultimately identified as brain states of specifiable kinds on the basis of an intertheoretic reduction on the model of a reduction of heat to molecular kinetic energy. In this view, there is no firm divide between what the manifest image tells us and what the scientific image tells us. The causation of action, thus, will be explicable in terms of the laws of nature at a scientific level. But the claim that neuroscientific categories will be found as the physical realisations of the mental categories embodies an optimism that is not shared by all materialist positions.

A second position to take with respect to the mind–body relationship is a more extreme version of physicalism than type-identity theory, which simply denies the significance of the manifest image in the explanation of action. Here, Paul Churchland (1988, 43) articulates this position, eliminative materialism, in an uncompromising fashion.

[O]ur common-sense psychological framework will not enjoy an intertheoretic reduction, because our common-sense psychological framework is a false and radically misleading conception of the causes of human behavior and the nature of cognitive activity.

This view of the mind and the causation of behaviour does not rely on the kind of optimism about the future of neuroscience inherent in reductive materialism. No longer do we rely on the future discovery of identities between mental and neurophysiological kinds. The only optimism required is that the causation of behaviour will be ultimately discoverable by neurophysiology. The cost is the complete relegation of the concepts of the manifest image to the waste bin of false and misleading concepts.

In his seminal paper, ‘Actions, Reasons, and Causes,’ Donald Davidson (1963) developed a position, commonly described as the causal theory of action, that dispenses with the substantial optimism of reductive materialism and, while putting confidence in science to an extent similar to the eliminative materialist, does so while preserving a place for the folk psychological concepts with which we are familiar.

This is Davidson’s proposal.Footnote 2 In his view, we do act from reasons and that is what makes us rational. And indeed, those reasons are the causes of our action. To provide a little more detail, a reason is composed of a belief and a desire, specifically the belief that the action will bring about the fulfilment of one’s desire. The rational person has a goal in mind and belief about what they will need to do to achieve it, which explains why they act that way. Thus, the causal theory of action specifies the connection between the reason and bodily movement when the latter is an intentional action that aims at the satisfaction of the person’s desire.Footnote 3

Now, consider how that reason can be the cause of action in Davidson’s view. For a reductive materialist, Davidson’s foil, there is no problem; the language of belief and desire is going to be mirrored in neuroscience. Therefore, a reason will have a neuroscientific description and can be covered by scientific laws that underpin the causal relationship. Davidson’s view, in contrast, is anomalous monism. It is anomalous in the sense that there are no exceptionless psychological laws. It is monist in that there is nothing over and above the physical. Those two things are reconciled by each psychological state token being identical with a physical state token. For Davidson, that identity at the level of tokens does not require identity at the level of types. We do not require the reduction of psychological laws to physical laws.

On Davidson’s account, mental events cause physical events, e.g., decisions cause actions which cause things to move around out in the world, and physical events cause mental events, e.g., when you touch a hot stove you feel pain. Particular causal claims commit one to the existence of an underlying strict deterministic causal law.Footnote 4 But there aren’t any such laws pertaining to one’s psychological life and, moreover, he rejects the notion of ‘bridge laws’.Footnote 5 This is Davidson’s ‘undeniable fact’ of starting from ‘the assumption that both the causal dependence, and the anomalousness, of mental events’ (Davidson, 2001b, 207). The strict deterministic laws will be physical laws, but, since there’s no translatability, those laws won’t also be psychological laws. This permits mental events to cause one another (by virtue of the physical laws underwriting them), and it permits mental events (under their physical descriptions) to cause and be caused by physical events.

Suppose that Smith wants to attract Jones’ attention and Smith believes that waving at Jones will attract his attention. Smith’s reason for waving at Jones consists of the desire to attract Jones’ attention and Smith’s belief that waving her hand in the air will attract Jones’ attention. Clearly, there could be various ways for Smith to achieve the same end. If Smith had believed that lighting a flare would attract Jones’ attention, then the combination of her belief that lighting the flare attracts another person’s attention and her desire to attract Jones’ attention is her reason for lighting the flare to attract his attention. The causal theory allows us to distinguish between such an intentional action and a mere tic or indeed involuntary bodily functions such as breathing (in normal circumstances) or digestion. Tics, breathing, and digestion are also caused internally, but not by a reason, so on Davidson’s account they are not intentional actions.

The waving example is an extremely simple example in which we would not normally bother to provide a reason. But there are more complex cases where we seek the real reason for an action. Why did the President’s press secretary resign? Various reasons may be offered by the person in question, for health reasons or to spend more time with family, and various others suggested by commentators, involving a conflict with a co-worker or as a result of pressure from the President. The causal theory allows us to ask the question: what really made him or her do it?

Thus, Davidson’s causal theory of action acknowledges the awkward nature of the interface between scientific and manifest images, and it provides one kind of model for how a reason, something clearly part of the manifest image, can also be located within the scientific image, which explains its causal power.

Davidson’s account of the relationship between mental and the physical takes as its basis events, allowing a mental event to also have a physical description. Now, events can be short or long; the second world war can be regarded as an event in the broad sweep of history. Moreover, beliefs, which enter into Davidson’s account, can be long standing, as indeed some desires can be. Even so, it could be argued that virtues and self-narratives are, by their very nature, to be contrasted with the more passing perceptions, thoughts and actions to which they relate. The nature of a virtue is something relatively stable in the story of a person’s psyche. Similarly, a self-narrative is something that persists in the to and fro of a person’s life.

While this is an apparent limitation in our use of the Davidsonian model to account for virtues and self-narratives, we can be reassured by the role we have in mind for that model. What we are looking for is a way of describing the relationship between the manifest image and scientific image with regard to virtues and self-narratives. Science need not be described in terms of events. The language of cause, it is true, can draw our attention to a particular event as being especially responsible for a particular effect. What caused the bridge to fall down? The storm contained unprecedented wind gusts that broke those struts. But the full explanation will bring in the engineering design of the bridge. It was designed to resist forces up to a certain level. Indeed, we can ask why the bridge stays up in normal circumstances! The explanation takes us to the design of the bridge, maybe in relation to the geology of the setting. Virtues and self-narratives are comparable to the engineering design of the bridge. They are standing features in the background that contribute to explanation of particular actions in different contexts.

Daniel Dennett (1971, 1987, 1991) has a position on the nature of mental states that has points of contact with Davidson’s; as Dennett (1991, 30) says, they are fellow Quinians. Later on, when we say more about how we should regard virtues and self-narratives as causes of action, we shall turn to Dennett to articulate our position.

3 Problems and Mysteries

It is worthwhile to consider the possible barriers to unifying the scientific and manifest images. In the study of language and mind, Noam Chomsky has distinguished between ‘problems’ and ‘mysteries’ (1975, chapter 4; 2009). While what are there called ‘problems’ stretch our ingenuity in theorising to its practical limits, the ‘mysteries’ are intractable for the conceptual resources that we currently possess, or even for those resources that could be developed. What we judge to be mysteries can change over time, so that the line of the distinction is not given in any a priori way but is sensitive to the progress of scientific investigation.

Chomsky suggests that the source of individual utterances or indeed of human behaviour generally lies in the area of mystery (Chomsky, 1975, 138, 156). For the remainder of the paper, we will call this the ‘mysterian view’.Footnote 6 This is related to his views about the limits of human science-forming capacity. He says that these limits flow from its internal structure which, as the other side of the coin, allow for successful theory creation in some domains (Chomsky, 2009, 184). The conceptual tool kit that is capable of performing well in some domains is likely, on the basis of those very potentialities, to be ill suited for other domains. The plumber’s tool box is not going to work for the electrician. Thus, there is no general reason to assume that we will be able to theorise successfully concerning all topics that excite our interest.

Notice that it is just blind luck if the human science-forming capacity, a particular component of the human biological endowment, happens to yield a result that conforms more or less to truth about the world. (Chomsky, 1988, 157–158)

Chomsky (2009, 184) cites Peirce (1957, 235–255) as a source of the view that we need such an innate endowment, which puts a limit upon our theories, in order to generate knowledge, but without following his assumption that such an innate endowment is the result of natural selection. To draw attention to this view of Chomsky’s is not to abandon interest in the causal theory of action, but it presents a certain kind of challenge to our thinking about the causes of action. While we may have assumed that causal theories of action are a product of Anglo-analytic philosophy and Davidson in particular, such theories of action have a long history. Chomsky draws our attention to the view of the eighteenth-century philosopher Joseph Priestley:

[Since] reasons, whatever they may be, do ultimately move matter, there is certainly much less difficulty in conceiving that they may do this in consequence of their being the affection of some material substance, than upon the hypothesis of their belonging to a substance that has no common property with matter. (Priestley, 1777/1965 in Chomsky, 2009, 193)

Chomsky’s position is the least optimistic of the ones we have been considering. We certainly do not have the intertheoretic reduction of mental categories to neuroscientific categories that reductive materialism promises. Nor do we have the lesser optimism reflected in both the eliminative materialist and the Davidsonian anomalous monism positions that there will be in principle neuroscientific explanations of the causation of behaviour.

4 Causation by Virtues and Self-Narratives

Our focus in this paper is on the way that virtues and narratives play a role in human action. Without attempting to provide a comprehensive account of either, we wish to draw out some features of both that draw a parallel between the two as causes of action. For example, both are related to character: virtues and vices are part of a person’s character, while a person’s self-narrative will reflect, or to some extent embody, their character. They are standing features of the person that we wish to highlight on occasion as a salient source of an action, even though the particularities of the situation will need to contribute to a full explanation. A pair of examples will help to explain our approach.

First, let us consider a case where we plausibly need to point to a virtue as a cause. Think of the soldier who, under fire, rescues a wounded comrade. We need to be able to point to his courage as a causal contributor. But a full causal explanation of the action is likely to be obscure. In part, it is because there will be many parameters in the conditions at the time. The soldier’s heroism may have been partly generated by his relationship with the wounded soldier, by his sense of honour for his country, or (perhaps less meritoriously) by his not wanting to witness someone else’s death. Indeed, one might talk of various character traits such as loyalty, resolution, and so forth that may be highlighted in an explanation. And, even if we focus on courage alone, it can be obscure how he managed to summon up the courage at the precise moment.

This obscurity can be described simply in terms of complexity. Chomsky’s mysterianism brings us a step beyond mere complexity. Explaining and predicting human actions in such circumstances may be forever beyond human science forming capacities. Even so, our knowledge of how a person behaves under a range of challenging circumstances may allow us to attribute the virtue of courage to them. Such a judgement will not necessarily be easy to make and there is room for dispute but we do understand the broad nature of that judgement. In a range of situations, the person recognises the danger and takes the personal risk nonetheless. Our position is that we need not, and indeed should not, think that we have merely described a person in dispositional terms: in these circumstances, they will indeed act courageously, but, for the attribution of the virtue to be genuinely explanatory, we need to treat it as an inner characteristic that makes a significant causal contribution to them acting in that way. While that characteristic is not going to be specifiable in scientific terms, it is something that we need to appeal to in order to understand others and should properly be thought of as having causal power.

A similar account applies to self-narratives as causes of actions. Any example is likely to be simplistic as it will be hard to convey a sufficiently rich flavour of the self-narrative in a short space. But consider a person, Habib, who thinks of himself as talented but a victim of discrimination in various ways. Let us suppose that he is musically talented but has experienced discrimination on ethnic and social grounds. His self-narrative has a prominent focus on unrecognised talent and discrimination. When stopped and searched by a police officer outside a club where he has been performing, he assaults the officer. We may or may not suppose that the police action itself was guided by discriminatory attitudes. Our question is, what brought about Habib’s aggressive action? The full explanation is going to remain a mystery, but it is reasonable to point to the self-narrative as a major causal contributor to the action.

An explanation of an action on the basis of its origin in a person’s self-narrative is bound to be complex and subtle given the richness that we would normally find in a person’s self-narrative. We only sketched Habib’s self-narrative in a crude, simplistic way. The general approach is that a self-narrative is not merely a bare chronicle of a life to date but can incorporate a sense of belonging or identification and emotional colour carried on the scaffold of a bare sequence of events. Moreover, it can be forward looking as well as backward looking, as past actions are to be understood as incorporating a sense of purpose, which need not be short-term. Sitting in the narrative will be various aspirations and goals. A doctoral student has built into their narrative something of their goal in undertaking that qualification, a purpose and ambition with respect to their specific topic and so forth. Thus, the self-narrative can be properly viewed as the basis for subsequent actions. Again, a brief description of a situation like this glosses over the richness of a real life situation. What are the family circumstances and expectations? What are the plans of friends and acquaintances including perhaps a boyfriend or girlfriend? There are further complexities to be found in the nature of self-narratives, which we, amongst others, argue are multi-stranded (see Lumsden and Ulatowski 2017, 2019, 2021, Forthcoming; Ulatowski 2021; Ulatowski and Lumsden 2023).

Clearly, an approach to the philosophy of action that emphasises the forward-looking dimension is the teleological approach, following Aristotle, where actions are to be understood in terms of goals. Even the causal theory has a forward as well as a backward-looking dimension as a desire, which is an essential component of a reason, is forward looking. Interestingly, Deniz Kaya (2021) connects a narrative explanation of action with a teleological approach. He takes to heart Lamarque’s (2004, 394) minimal interpretation of a narrative, which is something told that portrays at least two temporally connected events in some relation. Kaya observes that teleological explanations of action meet these minimal conditions. While that does serve to relate a narrative with an explanation of action, we think that the focus on minimal conditions for being a narrative distracts from the significance of self-narratives in human life. A self-narrative plays the role it does on the basis of a life story imbued with a rich amalgam of identifications, loyalties, emotions, aspirations, and goals. For a self-narrative to be a cause of an action, it must draw on a record of the past but one containing materials that look forward to the future.

We are trying to circumvent the presumption that the only proper framework for capturing causal relationships is a scientific one. It will be helpful to move on from Davidson’s causal theory and consider Daniel Dennett’s approach to the nature of beliefs and desires, which brings to the fore the question of whether they are real. His view of the nature of belief and desire has been described as instrumentalist (Lycan, 1990) and that view comes into focus when we consider how he treated beliefs and desires as part of an intentional system. Something is an intentional system ‘only in relation to the strategies of someone trying to explain and predict its behaviour’ (Dennett, 1971, 3–4). That strategy is to take the intentional stance towards that individual, attributing intentional states such as beliefs and desires to them and on that basis constructing explanations and hence predictions of their behaviour. Successful prediction, most of the time, clearly represents success in the strategy. But the label ‘instrumentalist’ suggests that attributions of beliefs and desires are mere instruments in the predictive process such that we shouldn’t take beliefs and desires seriously as entities in the fabric of the mind. The same reasoning could apply to virtues and self-narratives.

In later work, Dennett (1981, 1991) has pushed back against the label ‘instrumentalist’ and has firmed up his position on the reality of intentional states as being ‘mild realist’, somewhere between, as he says, ‘industrial strength’ realism of the manner of Jerry Fodor (1987), in which there is only a real intentional state if it is to be found in the language of thought realised in the brain, and the eliminative materialism of Paul Churchland, mentioned above. He captures this mild realist position by saying, ‘while belief is a perfectly objective phenomenon … it can be discerned only from the point of view of one who adopts a certain predictive strategy’ (Dennett, 1981/1985, 15). He also treats the status of intentional states as ‘real patterns’ that can be discerned from the intentional stance, using the analogy of seeing a common pattern subjected to different amounts of noise (Dennett, 1991).

Our interest is in virtues and self-narratives, but we can see that questions concerning their reality may mirror those concerning typical intentional states such as beliefs and desires, which are central to the intentional stance. Would a mild realist status for virtues and self-narratives be sufficient to attribute causal power to them? Dennett rejects the accusation that he must be treating intentional states as being epiphenomenal, having no causal power, saying, ‘If one finds a predictive pattern of the sort just described one has ipso facto discovered a causal power’ (Dennett, 1991, 43fn22). Dennett (1992) endorses a mild realist view of the narrative self, describing it as a centre of narrative gravity, where a centre of gravity is one of the examples he uses to suggest a mild realist metaphysical standing. He does say (1992, 103) that a centre of gravity ‘can also figure in explanations that appear to be causal explanations of some sort’, for they, ‘can compete with explanations that are clearly causal’, in explaining, for example, why something didn’t fall over. Thus, on the strength of that analogy, he extends his view about the causal power of intentional states to include the causal power of the narrative self, in this case with a note of reservation.

We have suggested that both virtues and self-narratives are connected to a person’s character and this makes us take a step away from Davidson’s causal theory of action. His focus is on reasons as causes but virtues and self-narratives have a nature that is rather different from reasons. David Velleman (1989) claims that, while human actions are often explained by motivation, they are sometimes explained by habit, by the agent’s character, or by emotion, and those other explanations do not need to always reduce to motivation. This is highly relevant to the influence of virtues and self-narratives. Annette Baier (2009) elaborates on that theme with particular reference to character. While David Hume is typically seen as an early source of the kind of causal theory of action that we have discussed, Baier argues that his discussions of character are in tune with Velleman’s claim that the source of action need not be restricted to motivation. She points us to occasions where the motivations or beliefs themselves demand explanation.

People may have a certain disposition and tend to act in accordance with how they have cultivated that disposition over time. She puts it nicely when she writes, in the context of Hume’s discussion of Charles I’s actions, ‘The kind of person one is helps determine both the sort of desires one has and acts on, and the way one selects advisors and forms one’s beliefs’ (Baier, 2009, 246). This is not to regard character as fundamental in the explanation of behaviour. For one thing, she allows that character can also be explained, and indeed people will sometimes act out of character, which directs our attention back to particular beliefs and desires. Certainly, she sees a person’s character, while always present, as not the sole source of the action, even where the person is acting in character. A common human nature is also always present, such that often it does not need to be remarked upon to provide an explanation. Also, it is the different situations that people find themselves in that play a major role in producing different life stories (p. 250).

When we seek to explain actions as flowing from virtues or from narratives, we need to explore a space following neither the Davidsonian account of the rationalisation of actions nor the purely scientific, neural level, description. Dennett’s mild realism can provide guidance about the metaphysical status of what we are describing.

5 Conscious Awareness in Virtues and Self-Narratives

A virtue is a source of the action in a way that reaches down below the level of immediate awareness. What enters consciousness may simply be a gut feeling about what should be done. While virtues are graspable by the human mind, that is not what is effective in acting virtuously. So, while virtues are part of the manifest image, we need to dive down below that conscious level in thinking about acting virtuously. Thus, we are drawing a contrast with those actions that require conscious planning and execution. I need to decide what I want to eat tonight and for the next few days and I decide where to purchase my supplies and how I am going to get there. Admittedly, the details of my actions are also going to be guided by habit, but at a broad level I am consciously aware of what I am doing and why.

When it comes to acting virtuously, conscious thinking about acting virtuously seems at least inessential and possibly a distraction. Acting in a virtuous way is distinct from following a rule, such as in the deontological approach, in which I consciously follow some moral rule. McDowell (1979) develops an Aristotelian virtue ethics, capturing the belief that how one should live is not codifiable. This makes itself manifest in situations where there are various concerns and a person lacks a rule telling them which they should act upon. He says (p. 345) ‘It is by virtue of his seeing this particular fact rather than that one as the salient fact about the situation that he is moved to act by this concern rather than that one’. He goes on to say (p. 346), ‘The relevant notion of salience cannot be understood except in terms of seeing something as a reason for action that silences all others’. To modify one of McDowell’s examples, you know a friend is troubled and in need of comfort but you have an obligation to attend a family gathering at that time. What do you do? That is determined by which fact is salient: your friend’s need for comfort or your place at the family gathering. If the former fact is salient, then you go to the friend and the reason to attend the family gathering is silenced. He is there considering a clash between moral claims, but the notion of silencing could be extended to the very way that virtues lead to right action and away from selfish concerns. The virtuous person who finds a lost wallet will be struck by the fact that a person is missing their property, which should be returned to them. We might say that any inclination to keep the contents is silenced.

Acting virtuously without specifically being consciously aware one is doing so can be viewed as acting from habit. In this case, one acts not from a sense of being compelled to do something or even having specific reasons for performing the action; rather, the action flows from one being habituated into doing it. Finding a lost wallet without yielding to the temptation of stealing its contents is something that has been ingrained in one’s mind. Or perhaps a more widely applicable example would be that one doesn’t ever feel compelled to rob a bank as one passes it. Some habits can be morally bad or morally neutral, but these are virtuous. To steal lost property or to rob a bank would not only be acting against one’s better interest but also be acting viciously. The person would have to part company from the ways in which she has been habituated to act. The action of returning property to its rightful owner is one of a mindless routine. Aristotle, (1990), in Nicomachean Ethics, for example, writes.

[I]t makes no small difference to be habituated this way or that way straight from childhood, but an enormous difference, or rather all the difference. (1103b, 23–25)

The causal influence of a virtue may indeed have the nature of a habit, generally speaking, but it is important to recall that the habit is making itself felt in a complex mental system which includes various kinds of conscious awareness. Thus, the nature and effectiveness of a virtue must be seen against that background.

To turn our attention now to narratives, they would normally be regarded as situated in the manifest image. Indeed, one might think that narratives are quintessentially located there. The capacity to narrativise is characteristic of the way that humans can see themselves and others as persons. But self-narratives are tricky, particularly when seen as sources of action. One significant aspect of narratives is that they often involve emotional colouring, though the nature of that involvement can be complex. To take the example of anger, someone contemplating their own life story may always or typically feel angry about the events as they see them. There also can be a meta-level awareness that they are angry about those matters, which may or may not lead to a calming of the anger. When the angry self-narrative leads to a certain action, it is unlikely that it is the more detached meta-level awareness that is doing the work. Rather, the action just springs from that accumulated anger, much the way that the virtuous action just springs from virtue. And, further, one can act out of anger without being aware of that source.

A narrative theory of the self has the tendency to operate in an overly simple picture of the mind as a source of action. The importance of emotional colour in the explanation of action has already been mentioned. We need to avoid a too bland conception of narrative as a kind of merely descriptive chronology. That would be to miss the explanatory power that is latent in the narrative approach. Also, there is a tendency to assume that a human life is governed by a singular narrative. The better model is as a bundle of narrative threads. This is a view that we have argued for elsewhere, but we are not alone in seeing some kind of complexity along those lines (see Lumsden and Ulatowski, 2017, 2019, 2021).

6 Causation in the Face of Mysteries

Supposing we are persuaded by Chomsky’s mysterian position on the causation of behaviour, we will still be inclined to speak in some way of how actions are caused. Even if the way the brain works to produce particular actions is going to be forever beyond us, we can still think of there being physical processes in the brain that underlie the things that we speak of in a pre-scientific way as motivating actions. Thus, we could regard the Chomskian view as supporting a ‘mystery’ variant of the causal theory. In this variant, virtues or self-narratives causing actions are not fundamentally more problematic than the causation of actions in general. While the mysterian position provides an additional edge to the causal connections we are interested in, people in their everyday lives happily accept that causal relationships may be unexplained, including ones that don’t concern mental causation. A wine glass is dropped on a moderately hard surface and smashes while a sibling glass is dropped on the same surface a moment later and survives. We do not throw our hands up in horror and appeal to occult forces or seek to overthrow fundamental and well-established physical principles including those in materials science. We just accept that there must be subtle differences between the point of impact and microstructure of the glass in each case. We are not necessarily in the area of mystery, for maybe this is just a problem that is intractable at a merely practical level, but this at least demonstrates an ordinary person’s acceptance of cause and effect without a grasp of underlying law.

Let us move up to a case where the focus is on human action and capabilities. Consider a fast bowler in cricket, Tim Southey, who manages to bowl out an opposition batsman, Moeen Ali. Perhaps Southey was bowling well and took a number of other wickets, but what was the explanation of the success on that very ball? It will be a combination of Southee’s skill and temperament, Ali’s skill and temperament plus atmospheric and lighting conditions, and even the support of the crowd. Explaining that particular success will be elusive, to say the least, unless Ali made an uncharacteristic batting error, or the umpires made a particularly bad call. We are a long way from causal relationships being grounded in underlying supporting laws, but this does not prevent commentators from pointing to aspects of the delivery and batsman’s shot choice as contributing causes of the loss of the wicket.

There is no pretence that we have arrived at the cause of the loss of the wicket in any comprehensive way, although a commentator’s rhetoric might, on occasion, get away from them. The singular relationship we are interested in is unlikely to reflect a general pattern (cf. Hume, 1738/2007, §105). If we single out Southey’s skill at fast outswing cricket deliveries (in favourable atmospheric conditions) as the cause of the loss of the wicket, that may single out a salient contributing cause, without that as such being supported by an underlying law. The full scientific account of the event may be intractable in its complexity. Going a step further, it may be a mystery in Chomsky’s sense of being beyond the limitations of human science-forming capacities. In either case, we are accustomed to the full explanation being elusive, which does not undermine the interest in, and viability of, a partial explanation such as cricket commentators (or, in another setting, political pundits) can provide. With the source of action in virtues and in narratives, we are in much the same situation.

Our question here is whether we can really accommodate virtues and self-narratives in a causal picture given mysterianism. Our answer rests on Dennett’s mild realism about intentional states described earlier, which we are applying to virtues and self-narratives. Where such things have explanatory and predictive bite, it is appropriate to grant them causal power. We thus drop an explanatory net over the person and attribute to them persisting states, the virtues of kindness or courage, which we consider produces, or causally contributes to the behaviour we notice. We do not expect there to be any comprehensive scientific model in the offing that will explain and predict all the behaviour. But we have a picture that works well enough. To withhold the attribution of a cause is to resist a legitimate and pragmatic attempt to make sense of the world. We can take a similar view when we see someone act in a way that we can understand as consistent with their self-narrative, which will embody: their family, ethnicity, class, professional role, religion, sexuality, musical allegiances, sports allegiances, and so on. An ordinary person need not have a carefully expressed notion of a self-narrative but would want to attribute their actions to who they are in that kind of way, which can be thought of as something within them. Once again, it is proper to view such a thing as having a causal role in conjunction with more immediate features of the context.

7 Conclusion

The central theme of this paper is how virtues and self-narratives might be thought to cause actions. Our first move was to make use of Davidson’s causal theory of action where reasons are considered to be causes of action. He rejects the view that folk psychological concepts are in principle reducible to neuroscientific concepts. Davidson, instead, proposes that a reason token is identical to some neuroscientific event, such that the causal power of a reason is explicable by there being exceptionless scientific laws that apply to it under its neurological description. We extend that thinking to take account of the differences between reasons, on the one hand, and virtues and self-narratives on the other, which have more the nature of background structures. The kind of neurological explanation that this approach promises is, at present at least, intractable. But worse, following Chomsky’s mysterian view concerning the limits of human science forming capacities, such neurological explanations of human behaviour may be forever beyond our grasp. Even so, it is still possible to understand action as originating from virtues and self-narratives. We apply Dennett’s mild realist account of intentional states to virtues and self-narratives. Their standing is justified by their usefulness in explaining and predicting a person’s behaviour. In order to understand each other, we need to treat those inner characteristics sufficiently seriously so that they can be treated as causes of action.