Aristotle and Contemporary Theories of Luck

Abstract

Contemporary theories of luck face problems when it comes to moral luck, that is, luck that nevertheless partially determines moral responsibility. Either they conceive of luck as chancy or modally fragile, which is too narrow and excludes cases such as choosing to do something that is unlikely for you to do or that you do not do in many nearby possible worlds. Others see luck as primarily a matter of lack of control, which is too broad and includes things like the sun’s rising, which is outside of our control, but certainly not a matter of luck. Some try to rescue the moral luck phenomenon by positing hybrid accounts or denying that moral luck is a species of luck. Very little has been written about how Aristotle’s conceptions of luck fit into modern conceptions. Yet, Aristotle has sophisticated accounts of luck and good fortune that shed light on certain problems. I will show how Aristotle fares compared to contemporary theories and what we can learn from his approach to luck and fortune when it comes to how lack of control, modal robustness, and probability factor into luck, the difference between luck and good fortune, and whether moral luck is a species of luck.

1. Introduction

Much of our lives are impacted in ways that we call lucky or unlucky. We may be lucky to narrowly escape a car accident because we happened to turn down the wrong street. We may be lucky that a bad choice of ours did not have disastrous consequences that were likely to have resulted from it. We think we have enjoyed good luck when we have natural virtue or virtuous parents, without which we would have easily gone astray. Each of these affects our choices, moral responsibility, and happiness. Most of the time, we make judgments of luck without thinking about the nature of luck itself. And in fact, the philosophical literature on moral luck—that is, luck that partially determines moral responsibility—did the same thing for decades before some philosophers suggested we given attention to a general theory of luck. What followed was much discussion on what a theory of luck would look like. Some argue that luck is primarily a matter of what’s improbable [1]. Others appeal to possible worlds, arguing that something is lucky if it happens in the actual world, but does not happen in nearby possible worlds [2,3]. Still others think that moral luck is a placeholder for that which is outside of a person’s control [4,5]. Lastly, some have developed hybrid views that combine these elements in different ways [6,7,8,9].

But each theory of luck faces problems. Theories that hold that luck is something primarily chancy or modally fragile are too narrow and exclude cases such as choosing to do something that is unlikely for you to do or that you do not do in many nearby possible worlds. Theories that hold luck is primarily a matter of lack of control are too broad and include things like the sun’s rising, which is outside of our control, but certainly not a matter of luck. Hybrid accounts have the same problems as probability or modal theories since they include either improbability or modal fragility as a condition. All accounts except the control account cannot explain why we call “lucky” the good (or bad) natural dispositions we inherit or other forms of good fortune that constitute stable features of our environment. Some thus deny that moral luck is a species of luck, claiming that the problem arises independently of any general theory of luck (See [4] (pp. 27–29), [10]).

Very little has been written about how Aristotle’s conceptions of luck fit into modern conceptions1. Yet, Aristotle has sophisticated accounts of luck and good fortune that have interesting connections to the contemporary debate and shed light on certain problems. I will show how Aristotle fares compared to contemporary theories and what we can learn from his approach to luck and fortune when it comes to how lack of control, modal robustness, and probability factor into luck, the difference between luck and good fortune, and whether moral luck is a species of luck.

2. Theories of Luck

“Moral luck” was first introduced in a pair of essays by Thomas Nagel and Bernard Williams who thoughtfully considered a paradox in our intuitions about moral responsibility [14,16]. On the one hand, moral responsibility seems to involve only that which is in our control. On the other hand, there are times we seem morally responsible for something despite its having been a matter of luck and outside of our control. For instance, two drunk drivers are headed home. One makes it home without incident while the other is unfortunate to meet with a pedestrian and strikes and kills him. It appears the second drunk driver is more to blame because he is responsible not only for choosing to drive drunkenly but also for killing a pedestrian. Yet he did not control the fact that a pedestrian happened in his path, and what he did control was identical to the less blameworthy drunk driver who, by chance, did not meet with anyone on his way home. This is a classic example of what Nagel calls “resultant luck,” where luck impacts the results of one’s choices and, subsequently, one’s moral status [14] (p. 60).

For much time after Nagel’s and Williams’ essays, there was relatively little written on the topic, but all who followed them presumed what is now referred to as the “lack of control” or “control” account of luck [4,17,18,19,20,21]. On this account, moral luck refers to anything outside an agent’s control that partially determines her praise-or blameworthiness [4] (p. 23).2 Yet in 2003 and 2005, Andrew Latus and Duncan Pritchard questioned this presumption. They both argued, first, that a general theory of luck needed to be developed in order to understand moral luck, and, second, that lack of control is not sufficient for such a theory [2,23]. For, many events3 like the sun’s rising or the law of gravity are outside of our control and are not matters of luck. Aside from the control account of moral luck, there are three other theories of luck: probability, modal robustness, and hybrid accounts. On the probability view of luck, luck is anything that is improbable. On the modal account, luck concerns that which happens in the actual world, but does not occur in nearby possible worlds. Some hybrid views combine lack of control with modal fragility and others combine it with improbability. Much of the literature is focused on presenting and responding to counterexamples to each view. So, in what follows, I explain each view and the problems with it by pointing to counterexamples that are most challenging.

The first theory is the probability account of luck developed by Nicholas Rescher. He describes his view as follows.

Luck is a matter of the positive (or negative) outcome for someone of a contingently chancy development that can yield positivities or negativities for this individual. One is lucky when, in a situation of uncertain outcome, one fares better than one has a right to expect, and is unlucky when one fares worse. And this matter of an outcome-antecedent “right to expect” is inherently probabilistic in nature [1] (p. 136).

Two conditions constitute Rescher’s probability account of luck. An event is lucky if it (1) has low probability and (2) is important to the agent. This means that a somewhat improbable event that is very important can be just as lucky as a very improbable, but less important, event [24] (p. 211)4. The strength of this view is that it captures the intuition that luck has to do with what is unlikely or unexpected.

There are nonetheless problems with the probability view. For instance, Pritchard argues that luck should not be understood strictly in terms of probabilities but in modal robustness (or lack thereof). He defines luck modally, claiming,

If an event is lucky, then it is an event that occurs in the actual world but which does not occur in a wide class of the nearest possible worlds where the relevant initial conditions for that event are the same as in the actual world [2] (p. 128).

He argues that there may be events that have a low probability but are nonetheless modally close and are thus matters of luck when they do not occur. For instance, in a close possible world, one might be killed despite its being improbable (say if one’s death is the result of a lottery draw, an event Pritchard claims has low probability but is nonetheless modally close; the balls just need to drop in a slightly different way). Pritchard explains that because in many nearby possible worlds, the person is killed, she is lucky to be alive in the actual world. In this case, one is lucky to be alive despite the low probability of dying [3] (p. 116), which seems to give the modal account an edge over the probability account. It can account for cases that have low probability but are intuitively lucky. There is another type of case that probability views cannot handle and that includes cases where an agent does what it is improbable that he do. For instance, given my character and general tendencies, it may be unlikely that I agree to go skydiving with a friend. But I decide nonetheless that I will go. Because going skydiving is the result of a decision I made, it is not a matter of luck despite being improbable5.

The modal account cannot explain some cases of luck. For instance, Jennifer Lackey shows that an agent can be lucky despite an event’s modal robustness. The counterexample involves a case where, given several contingent factors, an agent is guaranteed to find buried treasure (the spot he wishes to dig in order to plant a rosebush is precisely the spot that met other conditions for another person to bury her treasure). Finding the treasure appears to be a matter of luck for him despite the fact that he would have found it in many nearby possible worlds [19] (pp. 261–262)6. Since this is a case of luck that has modal robustness, it challenges the modal account of luck. In addition, Lackey suggests that some decisions based on a whim are modally fragile, but since they are conscious choices, they are not matters of luck [19] (p. 264). This captures another serious problem with modal accounts since many decisions may lack modal robustness but are not matters of luck. Take especially decisions where we are torn between two options. We may have made other decisions in close possible worlds, but making a conscious choice based on a person’s values and rational consideration is not a matter of luck. Thus, there are many times an event can be lucky despite being modally close or is not lucky despite its modal fragility and this undermines the modal account of luck.

Perhaps a more developed control account can avoid the problems that arise with the probability and modal accounts of luck without introducing other problems such as presuming the sun’s rising is a matter of luck. If so, then this account would be preferred to the other accounts. Classic control accounts hold that luck is whatever is outside of one’s control, understood in terms of direct or indirect control that is exercised in bringing something about. But, some theorists develop the control account by including a tracking or exploitation condition. For instance, Fernando Broncano-Berrocal argues that control involves both effective elements (which involve the power to intentionally bring something about) and epistemic elements, which he calls “tracking control.” To possess tracking control, an agent can monitor an event and use that knowledge to start, stop, or continue an action [29] (pp. 19–20). Likewise, Wayne Riggs recognizes that some events are outside of our control (like the sun’s rising), but not a matter of luck because an agent can exploit the event for some intended end [22] (p. 218). In either case, the sun’s rising seems to be dealt with since an agent can reliably make plans in relation to it. Thus, while it is outside of an agent’s control, it is not a matter of luck. Further, this control account can deal with events the probability and modal accounts cannot, such as unexpectedly finding treasure despite being guaranteed to find it, given that finding the treasure was outside of one’s control. It can also deal with cases where a person does what it is unlikely she do and whimsical decisions since doing what you decide to do is an exercise of control.

Yet, the refined account has also been met with criticisms. Pritchard challenges Riggs’ refined understanding of control with two arguments, which also have application to Broncano-Berrocal’s account. First, he claims that even if an agent does not exploit an event due to ignorance, it does not mean the event is lucky. For, he does not think we should take into account what an agent knows or happens to expect. Rather, if something is bound to happen, it is not a matter of luck [3] (p. 121). Further, he argues that the sun’s rising will still often be considered a matter of luck since most of the time, no one exploits this event [3] (p. 122). In Section 4, I explain why Pritchard’s insistence that we do not take into account what the agent knows is mistaken (against his first argument). In response to Pritchard’s second argument, I explain how the sun’s rising is excluded on Aristotle’s accounts of automaton and tuchē, which gives an advantage to his account over refined control accounts.

Given the foregoing problems with each theory of luck, some have argued for hybrid or mixed accounts. Neil Levy and Rik Peels, for instance, argue that luck will involve some sort of lack of control, modal fragility, and significance (See [6] (ch. 2), [7] (p. 60), [8] (p. 77), and [9] (p. 151)). Andrew Latus argues that luck is a matter of lack of control, low probability, and significance [23]. The benefit of hybrid views is that they intuitively include the lack of control condition, and by adding a modal or probability condition, they can account for why the sun’s rising or the law of gravity is not a matter of luck. But, because they include modal or probability conditions, they fall prey to the same problems as those views; they cannot explain events that are modally robust or probable, yet lucky, such as unexpectedly finding treasure a person is guaranteed to find. Further, where all accounts besides the lack of control account fail is that they exclude many instances of constitutive moral luck. It is common to think that the way we are naturally constituted is a matter of good or bad luck for us. Certain dispositions make virtue development easier or more difficult. Yet, one’s identity is modally robust and not a matter of chance. In fact, it seems to make little sense to say that if I were differently constituted, things would be otherwise, since there is no “I” that wins the natural lottery or that exists independently of how one is constituted. In nearby possible worlds, I am the same person. It is not by chance that I inherited certain tendencies.

One way outside of this difficulty is to distinguish between chancy and non-chancy types of luck, sometimes referred to as “luck” and “fortune”, respectively. For we commonly say that good things that befall us are matters of luck both when they were unlikely to occur (such as winning the lottery) and also when they involve stable, dependable features of our environment or selves, such as having good parents or being born with good natural dispositions. Yet, there is something different about these two, which may suggest there are two phenomena to be distinguished. Consider the following ways of making this distinction.

An event or state of affairs occurring in the actual world … is non-chancy lucky for an agent if (i) that event or state of affairs is significant for that agent; (ii) the agent lacks direct control over that event or state of affairs; (iii) events or states of affairs of that kind vary across the relevant reference group, and (iv) in a large enough proportion of cases that event or state of affairs fails to occur or be instantiated in the reference group in the way in which it occurred or was instantiated in the actual case [6] (p. 36).

If S is fortunate that Σ is actualized, it is of significance for S and she fails to exercise control over it, and there is a substantial number of possible worlds in which Σ fails to be actualized, whether or not they are close [9] (p. 156).

Luck is a matter of having something good or bad happen that lies outside the horizon of effective foreseeability. There is thus a significant difference between luck and fortune. You are fortunate if something good happens for you in the natural course of things … fate and fortune relate to the conditions and circumstances of our lives generally, luck to the specifically chancy goods and evils that befall us [24] (p. 28).

You can be fortunate with respect to an event whose occurrence was extremely likely, whereas an event is lucky for you only if there was a significant chance the event wouldn’t occur [30] (p. 392); see also [26] (p. 502).

What each theory shares is the idea that “luck” should be reserved for what’s unexpected or rare in some way (understood in terms of probability or modal fragility), and “fortune” picks out something more stable. Levy and Peels uniquely suggest that some analog to modal fragility is operative in fortune where an event is modally fragile compared to some reference group rather than oneself. For example, a genetic disorder is modally robust when considering who I am in nearby possible worlds but fragile in terms of the number possible worlds where it is instantiated. Aristotle similarly has two notions of luck that capture the intuition behind a “chancy” sort of luck and a stable one. Before seeing this distinction and comparing it to modern views, let us look first at his accounts of spontaneity (automaton), luck (tuchē), and good fortune (eutuchia) and see how they have an advantage over the four competing views of luck (probability, modal, control, and hybrid).

3. Aristotle on Luck

For Aristotle, luck and fortune impact many features of our lives and are important because of their relation to happiness, virtue, and choice. He often wants to exclude luck from discussion when his focus is on what is within our control or attributable to us as agents (Nicomachean Ethics (hereon EN) III.3.1112a29–31; see also Metaphysics I.1.981a5)7. Yet he also recognizes that some luck is needed to have a good life. For happiness is not just virtuous activity (EN I.7.1098a16–18); it also includes external goods (EN I.8.1099a28) which are caused by luck (Politics VII.1.1323b27–29). Eudemian Ethics VIII.2 (hereon EE) focuses on good fortune (eutuchia) and begins its discussion by observing that the fortunate seem to fare well (eupragia) just as those with practical wisdom do (EE VIII.2.1247a1). It is clear that luck and fortune play a dynamic and important role in his ethics and thus warrant attention.

There are two texts in Aristotle’s corpus that investigate the nature of luck (tuchē) and good fortune (eutuchia). The first is Physics (hereon Ph.) II.4–6 where Aristotle defines both spontaneity (automaton) and luck (tuchē). The latter is EE, which treats good fortune (eutuchia). These accounts line up with the “chancy” sort of luck. While EE is ultimately consistent with Ph. II.4–6, it suggests there are non-chancy types of luck, and there are many examples of these in EN, Rhetoric, and Politics. I will stick to the Greek words for the remainder of the paper given the fine distinctions between these ideas. Let us start with the “chancy” type of luck.

Aristotle states that automaton and tuchē are “the incidental production of some… result by a cause that took its place in the causal chain incidentally, and without the result in question being contemplated” (Ph. II.5.196b32–35). For Aristotle, automaton and tuchē involve incidental causation and what occurs “for the sake of something” (Ph. II.5.196b22, 28–29). Incidental causation is when a cause produces an effect that does not always or for the most part follow from it (Ph. II.5.196b10–11). Because of the irregularity, that which occurs by automaton or tuchē is contrasted to that which happens by nature or necessity (Ph. II.5.196b13–14; II.8.198b36). Aristotle further explains that lucky events are “for the sake of something” in the narrower sense that they could have been intended but were not (Ph. II.5.196b35–37)8. While being “for the sake of something” generally describes purposive behavior (either because it was intended by an agent or the result of a natural process), when the purpose is accomplished without having been intended, it is by luck that it is “for the sake of something.” Thus, “for the sake of something” is not identical to what “could have been intended but was not” except in cases of luck. Aristotle here is simply saying that what happens by automaton and tuchē appears to be intended in some way but was not.

Aristotle claims that tuchē is a narrower instance of automaton; cases of tuchē are cases of automaton that involve human choice (Ph. II.6.197a36–197b1). For example, running into someone who owes you money at the market is a matter of tuchē since it involves human choice. It meets the two general conditions since being in a position to collect debts does not follow always or for the most part from going to the market and, had you known the debtor would be there, you would have intended to go to the market for the purpose of collecting the debt (Ph. II.5.196b35–197a5). By contrast, automaton applies to animals and inanimate objects. If a horse is saved because it happened to run to a safe place or if a stone falls and hits someone, these are matters of automaton since they do not involve human choice and the effects occurred without being intended (Ph. II.5.197b14–19, 31–32)9. Aristotle also describes automaton and tuchē as indefinite in number and indeterminate (Ph. II.5.197a8, 16).

In addition to Ph. II.4–6, Aristotle discusses good fortune (hereon eutuchia, i.e., “good tuchē”) at EE VIII.210. Scholars have debated whether this account is consistent with the account given at Ph. II.4–611. For it seems that Aristotle mentions instances that do not share the features of the Ph.’s account of automaton and tuchē, such as success due to good impulses one has by nature (EE VIII.2.1247b32) or being divinely inspired in a way that makes one’s good fortune continuous (VIII.2.1248b7). But, he rules out the former instance since that which is caused by tuchē is contrary to reason, and natural impulses are not contrary to it; one can give an account by appealing to the impulse (EE VIII.2.1248a7–11). And the latter indeed appears to be a matter of automaton or tuchē since it involves the possession of a rarely instantiated nature; humans do not always or for the most part have a sort of divine inspiration that allows them to choose rightly without deliberation [47] (p. 93). Many scholars, by contrast, identify the divine type of good fortune at the end of EE VIII.2 with natural good fortune, i.e., having a good nature12. I have argued elsewhere that this is incorrect [47]. To briefly summarize, Aristotle leaves aside the case of natural good fortune since it is not caused by tuchē (EE VIII 2.1248a7–10), but he concludes the chapter by stating that divine inspiration is one type of eutuchia (EE VIII.2.1248b4–6). It would be odd for him to include this case in his final division, especially given that the division occurs just after the discussion ruling it out. Further, he notes that this person is successful due to divine inspiration rather than citing the impulse (EE VIII 2.1248a34–38). And while his conclusion states that the success is in accordance with impulse (kata tēn hormēn) (EE VIII.2.1248b6), he earlier described the success due to a good nature as coming from the good impulse (apo tēs hormēs) (EE VIII.2.1247b29). This suggests that the divinely inspired is a different case than those who are successful due to their good nature13.

Aristotle argues that genuine cases of eutuchia are not rational (alogos) and may accord with one’s impulses or not. He identifies being divinely inspired as a type that accords with impulses and claims it is continuous. The other genuine case involves success despite a contrary impulse (EE VIII.2.1248b4–7). Each involves incidental causation and is for the sake of something. The rarely instantiated nature of the divinely inspired person and success despite a contrary impulse do not occur always or for the most part. And since both involve success, they are both for the sake of some good [47] (p. 93). So it does seem as though he only has in mind the “chancy” cases of luck at EE VIII.2.

4. Aristotle and Contemporary Views

So far there are a few connections to contemporary theories we could make. First, it is notable that Aristotle never speaks in terms of possible worlds. But when he calls automaton and tuchē “indefinite,” his view allows for epistemological fragility14. The success enjoyed because of a lucky encounter could have been caused in a variety of ways that led the agent to the place in which the encounter occurred. Second, we can see the chancy aspect of luck from his claim that automaton and tuchē involve incidental causation—that which does not occur “always or for the most part.” This is not understood in terms of probability, but in terms of what purposive acts tend to produce. Further, refined accounts of control that exclude the sun’s rising when it was exploited for some end err in the implication that it is sometimes a matter of luck when it is not exploited. Instead, Aristotle’s account excludes the sun’s rising as a matter of luck on the condition that it happens “always or for the most part.” Third, his addition of “for the sake of something” is novel and insightful. It highlights the intuition about lucky events that they are beyond what the agent intended or could expect. Yet, it goes further to identify luck as what could have been the result of purposive action. This reveals the importance of the agent’s awareness in explaining lucky events. Further, and fourth, this shows affinity to control accounts, but with an advantage since it clarifies that it is not mere control but what could have been intended that matters.

Still, that which could have been intended may line up with Peels’ account of control. He claims that it is not de re control that matters but de dicto [9] (pp. 153–154). The agent in our market example had de re control of going to the market to collect a debt, but given the limitation of what he knew, he did not have de dicto control. What one knows or intends plays a role in the control one exercises15. Rescher asks whether we understand an agent’s “right to expect” in purely objective terms or in what the agent happens to expect [1] (p. 139). Aristotle’s view takes the latter approach, while Rescher suggests that we merely distinguish between the two types of luck and set the latter aside as a separate issue [1] (pp. 139–140). For Aristotle, it matters what the agent actually intended, what she could have instead intended, and what she could reasonably be expected to predict even if the odds of running into the debtor were high (or modally robust), unbeknownst to her. So it is not just de dicto control that matters, but how that compares to what could have been intended, which is sometimes captured by de re control (it could have been intended because the agent could have exercised direct or indirect control). But Aristotle’s concept of automaton extends more widely to include that which humans may not have de re control over, yet can be called lucky in relation to, such as if a tripod falls on its feet so that it makes for a seat. It did not fall with that aim of benefiting a person, yet it is a matter of automaton that it did since it conferred some benefit (Ph. II.6.197b17–18)16. Further, Rescher’s explaining the “right to expect” in terms of what is improbable runs into problems that are not resolved by appealing to modal fragility or lack of control. For something can be unintended and unexpected even if it is probable or modally robust or within control. This is the case when an agent lacks knowledge or awareness that something is probable, modally robust, or in her control17.

Aristotle’s account has a large advantage over the modal and probability accounts. Take for instance the treasure example. An agent may be guaranteed to find treasure on account of modal robustness (or high probability), yet finding it is a matter of luck. For Aristotle, what makes this outcome lucky is (1) the coincidence of two cause and effect sequences such that one cause results in an effect that does not always or for the most part follow from it and (2) the fact that the agent could have intended the outcome but did not. The two sequences are, first, the initial burying of treasure in a spot that met desirable conditions and, second, the planting of a rosebush in the (same) spot because it met other desirable conditions. They coincide such that planting the rosebush has the result of finding the treasure, which does not always or for the most part occur when planting rosebushes. Further, had the agent known treasure was buried there, he could have intended to discover it by digging. But he did not know of the treasure, and thus he did not intend to find it.

In an attempt to rescue the modal account of luck from this counterexample, Pritchard suggests that if the field with proper soil were large enough or if the treasure was buried deep enough (which is likely), then finding it would not occur in nearby possible worlds and thus would be lucky [3] (pp. 119–120). But the finding of treasure is indeed lucky, whether or not these additional details are added, and on Aristotle’s account we can see why. His account does not require that an agent possessed all knowledge necessary for intending at a prior time. Rather, his examples indicate that it is only what could have been intended had the agent possessed such knowledge that matters. It also does not happen always or for the most part that when one digs in order to plant rosebushes that one finds treasure. Digging for gardening purposes is an incidental cause of finding treasure in this case. And, had the agent known treasure was buried there, he would have intentionally dug in that spot. But he did not know, so his finding of the treasure is lucky.

Unlike Rescher and Peels, Pritchard denies that we should take into account what the agent knows about the situation to determine whether it is lucky. He claims that if someone is guaranteed to find the treasure, then it is intuitively not lucky for him to have found it, despite the fact that he is unaware of all relevant facts and finding the treasure was unexpected [3] (p. 120). Aristotle follows a line of reasoning that responds to Pritchard’s intuition about what we should not build into an account of luck. For one, he claims that some people question whether chance exists, arguing that despite thinking something is lucky, we can nonetheless trace it back to some cause (Ph. II.4.196a6–8). Yet Aristotle denies that this is a plausible view and suggests instead that whether or not we can find a cause, we nevertheless (correctly) call things a matter of luck when they are irregular (Ph. II.4.196b13–14) or could have been the result of intentional action but were not (Ph. II.4.196b18–19)18. Despite contingent facts in this case that guarantee that digging to plant a rose bush will result in finding treasure, digging for such a purpose is generally not carried out with the intention of finding treasure nor is finding treasure the proper goal of digging to plant rosebushes19. So in one sense, we can ignore that the agent’s knowledge was limited and explain the lucky event in terms of what generally follows—or rather, does not follow—from an intentional act. But, given that automaton and tuchē also involve what could be intended for Aristotle, it would be a mistake to abstract from what the agent knows. For it would not be lucky for someone to decide to plant a rosebush where he knows a treasure is buried in order to obtain the treasure whilst gardening. It is not directly his knowledge but what his knowledge contributes to his intentional action that matters in distinguishing this unlucky event from a luckier one.

Aristotle’s view of automaton and tuchē can easily accommodate instances of luck where something is outside a person’s control, improbable, or modally fragile since that which does not follow always or for the most part will often be outside of control, improbable, and modally fragile. Yet, his account still recognizes other clear cases of luck that are not improbable or modally fragile (such as finding treasure or intentionally doing something that is improbable that you do) and cases that are outside of an agent’s control, but not lucky (such as the sun’s rising). Ultimately, contemporary views fail because they do not ground their understanding of luck in a teleological understanding of nature and purposive action. Unnatural events tend to be improbable, modally fragile, and uncontrollable, and this is why those accounts (or hybrid accounts including any of them as necessary conditions) often get things right20. Yet, if we take as central to our account of luck that it contrasts to that which occurs by necessity or nature, we can see what is more deeply operative in intuitive examples of luck (and counterexamples that meet conditions laid out in contemporary views). Rather than the probability or modal robustness of an event’s occurring given actual conditions, or whether it was in an agent’s effective or exploitative control, we ought to consider the natural goals of actions (or natural processes) and (in cases of tuchē) what the agent intended or (in cases of automaton) what appears as though it were intended.

5. Aristotle and Good Fortune

Despite its advantage when it comes to handling cases like the sun’s rising and events that are modally robust, probable, or (unknowingly) within control yet lucky, Aristotle’s accounts of automaton and tuchē similarly do not address all cases of constitutive moral luck. Since automaton and tuchē are contrasted to nature, our natural constitution cannot be a matter of luck. Still, Aristotle claims that not all matters of eutuchia are caused by tuchē (EE VIII.2.1248a13–15). And elsewhere in his corpus we are met with examples of such cases. For instance, he says that those who are born with a good nature (i.e., who have good constitutive luck) are fortunate (eutucheis) (EN X.9.1179b21–23). In the Rhetoric, he says that one’s beauty and stature are matters of eutuchia and contrasts these with chancy goods of fortune caused by tuchē such as being the only beautiful sibling (I.5.1362a4–8). In the Politics, Aristotle gives his infamous ethnography, suggesting that certain natural temperaments due to environmental factors like climate are important for developing virtue (VII.7.1327b30–37). So, it is fair to say that he does have a non-chancy conception of eutuchia in addition to the chancy conceptions found in Physics and EE21.

The non-chancy conception that emerges seems to have two features. It involves both what is outside of an agent’s control and what contributes to her happiness22. This account of good fortune is remarkably similar to control accounts of luck, which typically include a control condition and a significance condition. Having the chancy conception thus gives Aristotle an edge on control accounts since it can exclude other cases that the lack of control account does not, such as the sun’s rising or the law of gravity.

I noted earlier that the modal, probability, and hybrid accounts have trouble explaining constitutive moral luck—luck that impacts our natural dispositions in a way that at least partially determines our moral responsibility. It is not a matter of chance we are born the way we are, and our identity is modally robust so any account that requires either of these conditions cannot explain constitutive moral luck. Aristotle’s accounts of tuchē and automaton similarly cannot explain it. While it meets the “for the sake of something” condition—one could have intended to habituate the naturally good impulses but did not—it is not a matter of incidental causation. Nature involves that which occurs always or for the most part23.

Levy does make some progress on this front by suggesting that we consider the number of possible worlds where something obtains rather than the nearness. He states, “Rather than relativizing worlds to the individual, to understand attributions of constitutive luck we should understand the relevant worlds as those in which human beings like us exist: I am (non-chancy) constitutively lucky in those traits and dispositions that vary significantly in human experience” [6] (p. 33). This account in fact applies to the continuous form of eutuchia mentioned at the end of EE VIII.2 (i.e., the divinely inspired person), though Aristotle does not talk about it in terms of modal robustness. Recall that that was an instance of eutuchia caused by tuchē, and so it is, contra Levy, a chancy form of luck. Still, this leaves out many paradigmatic cases of constitutive luck. The intuition is that even if my natural disposition is not rare relative to any population set, it is still a matter of fortune that I have it since I cannot control it and it impacts my character and moral decisions.

What this ultimately points to is that moral luck is a phenomenon that cannot be wholly comprehended by abstracting from cases and building a general theory of luck. For not all cases of moral luck are caused by luck, as Aristotle holds and some contemporary theorists have argued ([4] (pp. 20–23), [10]). The initial push to articulate a general theory of luck, then, will only get one so far. Sometimes what impacts our moral lives are chance events and other times we recognize how fortunate (or unfortunate) we are with respect to stable features of our inherited constitution, development, or environment. Only the former is caused by luck. But because the latter picks out a real feature of our lives that impacts our moral judgments, it is an important phenomenon to investigate, whether we call it “luck” or “fortune” and whether it is a species of luck or not24.

There is still a final worry that any control account of luck or fortune includes too much. The sun’s rising is significant and impacts my happiness and is outside of my control. So even if it is not a matter of automaton or tuchē for Aristotle, it fits the criteria for being a matter of non-chancy eutuchia. One passage in the Physics seems to address this very worry. Aristotle says,

If a man has his head shaved for some special treatment, and afterwards goes out for some reason indifferent to his cure, and the air and sun cure him, is the exposure to air and sun (which is the efficient cause of his cure) incident to his going out? Or must we go further back and say that it is incident to his being shaved, without reference to going out, his going out itself being reference to the cure? The only answer to such questions is that, here as elsewhere, ‘incidence’ may be more or less proximate or remote (Ph. II.5.197a22–25).

Though Aristotle is discussing incidental causation, his point about having “more or less proximate or remote” causes can be applied to cases of eutuchia that pick out stable features of our environment. To some extent, it makes sense to us to say that we are fortunate that the sun rose today precisely because it is outside of our control and benefits us. It is not a matter of luck that the sun rises, but a person with a grateful attitude begins to appreciate the numerous ways that she is benefitted from factors she did not choose and could not control. Yet we can still say that these causes are more or less proximate or remote. I am especially fortunate for parents, teachers, relatives, or natural dispositions that contributed positively to my moral development as these are more proximate causes of good development. That we had daylight to facilitate the circumstances in which I learned or played is another positive factor, but it is a more remote cause. That the control account and Aristotle’s view of eutuchia include so much is in fact a strength. As Martha Nussbaum explains, Aristotle retrieves the sentiment of the tragedians that human goodness is frail and, indeed, plant-like [51] (p. 237). The truth of the matter is that our lives are greatly impacted in many ways that make us fortunate. This adds value to human life and ought to be appreciated.

6. Conclusions

I have shown how Aristotle’s conceptions of automaton, tuchē, and eutuchia address problems that come up in contemporary debates about luck and moral luck. His view also shows that not all instances of moral luck are a species of luck. Some forms of luck, such as certain cases of eutuchia, do not meet the same conditions he sets forth for his general theory of luck (automaton or tuchē).

His view is importantly different from four leading views: probability, modal, control, and hybrid. Defining automaton and tuchē in terms of incidental causation involving that which is done “for the sake of something” can explain each of the key examples of luck offered by the competing theories without falling into the pitfalls that each theory is vulnerable to. These key examples are the sun’s rising, choosing to do something modally fragile or improbable, or luck despite modal robustness or high probability. For instance, it does a better job explaining why the sun’s rising is not a matter of luck, even on a more refined version of the control theory. On the refined theory, it is not a matter of luck when an agent exploits it for a particular purpose. Yet, this has the result of being lucky when it is not exploited. Aristotle’s account of luck that requires incidental causation can explain that given that the sun’s rising happens “always or for the most part,” it is never a matter of luck. The modal, probability, and hybrid accounts requiring one of these as a condition cannot account for whimsical decisions or decisions to do something that it improbable that you do, such as making a choice to go skydiving when it is unlikely, given your character, that you would do this. Further, there are lucky events that are modally robust or probable, where an agent is guaranteed to enjoy success given contingent factors but was not aware of this, such as the treasure example. Yet, the modal, probability, and hybrid theories that include either of these conditions cannot account for why these events are lucky.

Lastly, the account of eutuchia explains the non-chancy types of luck—which can only be accounted for on the control account of luck. This addition does not present a unique advantage over contemporary theories since the control account of luck can also explain constitutive luck, and other modern views similarly adopt this strategy of distinguishing between chancy and non-chancy types of luck. But, because his chancy theory of luck has great advantages over all theories attempting to provide a general theory of luck, his views add much value to contemporary discussions of luck.