Luck and the Automatic in Aristotle's Physics

In Physics II.4-6 Aristotle confronts a puzzle that seems at first to threaten the entire coherence of his natural philosophy, which is the existence of chance. If everything in nature happens either always or for the most part and much of it for the sake of something, then where does chance fit in? Events like coming into the market and finding someone you wanted to see but didn't expect to meet (Phys. 2.4, 196a3-5), or a falling stone that happens to strike a man (2.6, 197b30-32), don't seem to follow any of the natural regularities Aristotle relies on to explain change, but these events clearly happen and often enough that they demand some philosophical explanation. (Hardie and Gaye translate luck as "chance" and the automatic as "spontaneity," so in the quotations below "chance" means luck.) Aristotle's answer is not to deny the existence of luck but to absorb it into his broader theory of causes, and in doing so he makes a sharp distinction between events that happen by nature and those that occur by chance. This raises an important question though, since luck can be absorbed in a bad way as well and his predecessors had already done it twice. Some said that "nothing happens by chance, but that everything which we ascribe to chance or spontaneity has some definite cause" (196a1-3), so the man went to the market to shop and that is the whole story, and "chance" just names our ignorance of the real cause. Others kept chance as a cause but made it "inscrutable to human intelligence, as being a divine thing and full of mystery" (196b5-7). The first view explains luck away and the second keeps luck but puts it beyond explanation, so neither one actually gives us an account of it.

Aristotle builds his own account out of two divisions (2.5, 196b10-22). Some things "always come to pass in the same way, and others for the most part" (196b10-11) and no one says either of these happens by chance, but there is a third class of events that happen neither always nor usually which everyone calls lucky, so chance is real (196b13-17). The second division is between things that are for the sake of something and things that aren't, and the first group covers "whatever may be done as a result of thought or of nature" (196b21-22). The two divisions cut across each other, since even among events outside the necessary and the usual some are the kind of thing that can be for the sake of something (196b19-21), and when one of these comes about incidentally we say it happened by chance (196b23-24, 196b29-30). So everything depends on what he means by "incidentally," and he gives an explanation by analogy (196b24-29). A cause can be what it is in its own right or incidentally just like a thing can, and "The housebuilding faculty is in virtue of itself the cause of a house, whereas the pale or the musical is the incidental cause" (196b25-27) because the builder just happens to be pale or musical. A per se cause is determinate but an incidental one isn't, "for the possible attributes of an individual are innumerable" (196b28-29). Still Aristotle insists that even these chance events must have causes. They are not uncaused, just not caused in an essential way, and their causes are real but accidental. This means chance events have efficient causes like anything else (Aristotle never denies this) but what makes them chance is the way the cause relates to the result.

The classic example Aristotle gives is a man collecting contributions for a feast (196b33-197a5). "He would have gone to such and such a place for the purpose of getting the money, if he had known" (196b33-34), but he went there for another reason and got his money incidentally. I think three things make this luck. For starters the result is the sort of thing he would have chosen to go for, since it "belongs to the class of things that are intentional and the result of intelligent deliberation" (197a1-2). It is also not what he actually went for, and the end of getting the money is not "a cause present in himself" (196b36-197a1), so it wasn't among the reasons that moved him. Lastly he didn't go there "as a rule or necessarily" (196b35-36), and if he always or usually went there when collecting we wouldn't call it luck (197a2-5). His going is an incidental cause of his getting paid in the same way the musician is an incidental cause of the house, which is why "chance is an incidental cause in the sphere of those actions for the sake of something which involve purpose" (197a5-6). It is tempting to read this as two purposeful chains crossing (his errand and the debtor's), but that isn't really what Aristotle describes. Instead he is looking at one action and an end that it happened to achieve, and deliberate choice only comes in by fixing which results count as lucky, which are the ones a choosing agent could have aimed at. In chapter 6 he makes a distinction between luck and the automatic. The automatic is the wider class and includes chance happenings in the broader natural world, so everything lucky is automatic but not the other way round (197a36-b1), whilst luck needs an agent capable of choice since "an inanimate thing or a lower animal or a child cannot do anything by chance, because it is incapable of deliberate intention" (197b6-8). For example the horse that came and was saved came automatically since "though his coming saved him, he did not come for the sake of safety," and the tripod that fell and landed upright as a seat "did not fall for the sake of that" (197b14-17). The stone that struck a man also fell automatically "because it might have fallen by the action of an agent and for the purpose of striking" (197b30-32). Both kinds belong to the moving cause, "for either some natural or some intelligent agent is always the cause" (198a2-4), and since chance results are things that "though they might result from intelligence or nature, have in fact been caused by something incidentally" and "nothing which is incidental is prior to what is per se," chance is "posterior to intelligence and nature" (198a5-10). What matters for Aristotle is that these cases are the exception and not the norm. In II.8 he uses this same distinction against Empedocles, since natural things come about "either invariably or normally," but "of not one of the results of chance or spontaneity is this true" (2.8, 198b34-36), so because chance events are rare and incidental the regularity and teleological structure of nature stay intact.

Aristotle wants to claim that chance fits naturally into his larger theory of causes but there's a tension here. If chance is a real cause then isn't it just another kind of regular cause we don't fully understand? And if it's not a real cause, why include it at all? A critic may object that if chance is just the absence of a final cause then calling it a cause is misleading. It's not just a different kind of cause but a lack of one, and that isn't an explanation so much as a gap. Aristotle would say that he built his account to meet exactly this worry, since chance is never the unqualified cause of anything and the real cause is the action, which produces the result incidentally (197a10-14). So even if this objection fails, I think the larger problem is that his account equivocates between descriptive and normative uses of teleology. He uses "for the sake of" both to describe what things do and to say when they succeed. Descriptively it picks out what a thing does always or for the most part, as when each nature tends "towards the same end, if there is no impediment" (199b17-18). Normatively it picks out success and failure, as when an action is called in vain "when A which is for the sake of B, does not result in B" (197b22-23) or when monstrosities are "failures in the purposive effort" (199b1-4). Aristotle never claims that lucky events should happen, but the condition his definition rests on is that a chance result has to be among the things that "might come to pass for the sake of something" (197a32-35), and in which sense? If the sense is descriptive then the condition rules out almost nothing since anything could be someone's aim (people do sometimes throw stones at each other), and luck shrinks to any irregular result at all, which is the II.4 view that chance only names causes we failed to track. If the sense is normative then a result counts as lucky because it is good or bad for someone (fortune is called "good" or "evil" by its result, 197a25-26), which looks like a fact about our interests and preferences and not about how the result came about. So Aristotle's account seems unstable since he wants chance to be a real phenomenon (not just a name for our ignorance), something that explains why certain things happen and something compatible with final causality. It is not clear however whether he can hold all three together. Read descriptively the condition collapses chance into the ignorance view, and read normatively it makes chance depend on whose interests are at stake so it can't explain why the result happened. Either way luck gets absorbed in one of the bad ways Aristotle rejected.

In defending Aristotle the key is to understand his use of chance as a boundary concept (a way of naming the limits of teleological explanation without abandoning it entirely). He isn't claiming that chance events are for the sake of something in themselves but that they are byproducts of a world in which other causes do act for ends, so chance presupposes an ordered world. You can't have luck unless you have a baseline of purpose driven behavior for it to deviate from. Aristotle would respond that luck is parasitic on order and only makes sense because we can recognise that something happened that wasn't supposed to happen but still did, and Physics 198a5-13 puts this in his own terms. A chance result is one that "might result from intelligence or nature" but was caused incidentally, so the condition isn't a test applied to the result on its own in either sense. It points to an actual per se cause, some nature or intelligence that produces this kind of result as its end. Getting the money is what a collector's trip is for and striking a man is what a stone does when someone throws it (197b31-32). The descriptive horn fails because not just anything has a cause like this behind it, and Aristotle says it would be absurd to say you bathed in vain because the sun was not eclipsed, "since the one was not done with a view to the other" (197b26-28). The normative horn fails because the success belongs to the per se cause and not to whoever happens to benefit. Even the two uses of "for the sake of" aren't a slip, since for Aristotle what a nature does "if there is no impediment" just is its end and missing that end is failure, so the descriptive and normative senses are two sides of one fact about per se causes. This explains why chance doesn't destroy Aristotle's view of nature and actually ends up reinforcing it. We only notice the rare event as rare because most things happen always or for the most part (198b34-36), and the system still works with luck just adding noise to the signal.

Overall Aristotle's account of luck and the automatic walks a fine line between inclusion and exception, and this defense puts all the weight on per se causes really having ends. He wants to preserve the regularity of nature without ignoring the messy reality of chance, but the claim that per se causes have ends is argued for in II.8 and not in II.4-6. Empedocles' alternative, that animal parts "organized spontaneously in a fitting way" survived while the rest perished (198b29-32), is a theory on which regularity itself comes from chance, and if it is right then the signal is made of noise and chance would not be posterior to nature after all. The question that remains is whether this balance holds up. Can a theory built on purposes also explain the purposeless? I think it can, but only if Aristotle can show that the purposes came first.

Works cited

• Aristotle, Physics, trans. R. P. Hardie and R. K. Gaye (Internet Classics Archive, MIT). Bekker references checked against the Greek text, Aristotelis Physica, ed. W. D. Ross (Oxford: Clarendon, 1960), digital edition by the First Thousand Years of Greek project.