In Early Modern Philosophy we have come to Hume’s critique of causality (Enquiry, section 4). I have always had difficulty getting the nature of his critique straight in my mind. So I sat down to try to put it clearly, and came up with the following. Nothing new or original here – just what I hope is a clear articulation of one of Hume’s great insights.
We all know what happens in the story that begins, “David Hume came to a fork in the road….” The two tines of the fork are marked “relations of ideas” and “matters of fact,” which Hume regards as “the only two objects of human reason.” What makes a true relation of ideas true is the fact that when you try to deny it, you contradict yourself. What makes a true matter of fact true is … a good question. Hume would like to find an answer to it. He observes that every purported matter of fact is founded in one way or another upon cause and effect. From what we immediately perceive we infer a cause; from our memories, we infer a past event; from our past experience of two events being always conjoined, we infer the second from evidence of the first. And so on. So, he claims, if we want to know what makes true matters of fact true, we need to know what makes causal claims true. Hume then proceeds to demonstrate that general causal claims are not relations of ideas, since if you deny a causal claim, you will not thereby contradict yourself (though you will run the risk of sounding silly). But neither are general causal claims matters of fact, since … well, since why? Remember, we are trying to figure out what makes true matters of fact true. We don’t know the answer yet.
Hmm. Well, practically, what Hume ends up assuming is that a true matter of fact is true in virtue of accurately capturing what is present in our experience. So I eat some bread, and it nourishes me. I eat some more, and it nourishes me, too. The true matter of fact issuing from this experience is that the bread I have eaten on these two occasions has nourished me. But typically we also infer something much stronger: that the bread that I will eat in the future also will nourish me, or that bread always has nourishing qualities (under similar conditions). But these claims do not accurately capture what is present in my experience, if only for the reason that the future, let alone what “always” happens, is not present in my experience.
If this is enough to disqualify general causal claims as matters of fact, then Hume has succeeded in showing that causal claims are not objects of human reason. What does that mean? It does not mean that we should be skeptical of causal claims. It means we should not take ourselves to have any real understanding of why they are true. As he writes towards the end of part I of section 4:
Elasticity, gravity, cohesion of parts, communication of motion by impulse; these are probably the ultimate causes and principles which we discover in nature; and we may esteem ourselves sufficiently happy, if, by accurate inquiry and reasoning, we can trace up the particular phenomena to, or near to, these general principles. The most perfect philosophy staves off our ignorance a little longer: as perhaps the most perfect philosophy of the moral or metaphysical kind serves only to discover to larger portions of it [namely, our ignorance]. Thus the observation of human blindness and weakness is the result of all philosophy, and meets us at every turn, in spite of our efforts to elude it or avoid it.
11 thoughts on “Hume on causal knowledge”
I tend to read Hume as being a rather harsher skeptic that what Huenemann has presented here. My interpretative clue in reading the Enquiry will be that Hume wants to destroy metaphysics (I think natural science ends up being a bit of collateral damage). To do so, Hume needs to undermine causal claims which function as a linchpin premise in almost every argument for the existence of God.
Here is my short crib of Huenemann’s reading of Hume: We aren’t debating whether or not causal claims are true. Rather what we are trying to sort out is what makes them true. As it turns out, what makes them true is beyond my capacity to understand. So I can’t ultimately justify the truth of these claims, but the causal claims themselves are plainly true (not just useful, but true). So it is not that the causal claims are not true (that I should be skeptical of them), it is that I can’t demonstrate their truth.
But if so, isn’t a fair bit of metaphysical speculation about God still alive and well? What one needs for the cosmological argument, for example, is the principle of sufficient reason. If Hume is willing to grant this as a true proposition (even if I can’t ultimately explain why it is true), can we then carry on with cosmological arguments?
If we restrict ourselves to the Enquiry (set aside for the moment the Dialogue Concerning Natural Religion), this seems like a weak reading. It seems to undermine the “customary convention” move (reflected in the Treatise concerning personal identity as a “fiction”). Instead of the PSR being a “customary convention”, a habit of the mind, and something of a projection of our subjective psychology onto the wold, this reading counts it as a genuine truth about the world. And if so, has Hume (on Huenemann’s reading) really accomplished his task of ending the practice of metaphysics?
[ By the way: I think the matters of fact (the contrary being possible) bit is somewhat misconstrued from the get-go. My complaint: Hume is failing to distinguish between our imaginative and our intellectual powers. Hume assumes that since I can imagine something (suppose I imagine something coming into existence without a cause) that thereby I can conceive it. But can you conceive of an effect without a cause? I don’t think so. ]
To Vince’s second paragraph, I don’t think Hume thinks “cause-effect always has a mysterious little black box that hides the point of contact between cause and effect”. Rather I think his problem is that induction can only be justified via induction. That believing in cause and effect requires the belief in the uniformity of nature (i.e. that the future will resemble the past). So Hume thinks the belief that induction is a sound method is an article of faith, or at least not demonstrated by “reason” (deduction). I don’t think he would deny cause-effect is demonstrated by induction.
I tend to read Hume as more of a radical skeptic, hence Kant’s being woken from his dogmatic slumber. So my more optimistic leap was not meant to be a statement of my reading of Hume, rather I took it that this was Huenemann’s reading. Huenemann’s post, as I understood it, suggests that Hume is not nearly so radical a skeptic as is often thought. Huenemann?
I’m with Vince on the “little black box” view. I think he views causality as utterly opaque to rational explanation. Yet there is causality, nonetheless. So we record the regularities, use them, and move on. (Mike, I think what you’re saying is consistent with this.)
I can’t think of any text where Hume actually shows any skepticism toward the truth of causal principles. Indeed, he bases his critique of belief in miracles on it. Where Hume is skeptical is when we try to apply our understanding of the principles regulating familiar, natural objects to things that we never experience. So the soul, the entire universe, God – we really don’t have a lot of experience to go on with these things, so it’s better to just stick to what we know, and let big question marks hang over the rest. He very tentatively endorses a kind of deism in the Dialogues, but I think he was genuinely unsure over it – hence the dialogue format.
(Note also that Hume thinks, according to our experience, causes are always in close contact with their effects, and he uses that observation to criticize Newtonian action-at-a-distance. So science indeed suffers a bit of “collateral damage,” as Kleiner calls it.)
That “stick to what you know” approach is enough to take the wind out of the sails of any dogmatic metaphysical program I know of. Aquinas, for instance, doesn’t conclude that there probably, maybe, is some sort of mindy-type creator of the universe. It even would deflate Kant’s project, since Hume wouldn’t be so sure about this “synthetic apriori” business. Are we really sure that Euclidean geometry could never even possibly be violated in our experience?
It’s been a while since I’ve read the Enquiry but what I was trying to emphasize is that Hume’s skepticism seems fairly specific to me and it won’t be resolved “as we continue to demonstrate how cause and effect works mechanically” but I see now that there are a few different issues muddled together there.
A passage from section 5, supporting my reading of Hume’s skepticism: “Nature will always maintain her rights, and prevail in the end over any abstract reasoning whatsoever. Though we should conclude, for instance, as in the foregoing section, that, in all reasonings from experience, there is a step taken by the mind which is not supported by any argument or process of the understanding; there is no danger that these reasonings [the reasonings of common life], on which almost all knowledge depends, will ever be affected by such a discovery.” So the fact that we can’t really understand how causality works should not limit our confidence in causal laws.
Does that mean we’re free to apply causal principles in such ways as, say, Aquinas does? Nope. In section 7, as Hume muses over occasionalists’ arguments that all causal powers actually belong to God, he writes that “[w]e are got into a fairy land, long ere we have reached the last steps of our theory; and there we have no reason to trust our common methods of argument, or to think that our usual analogies and probabilities have any authority. Our line is too short to fathom such immense abysses.” I think he’d say the same over each of Aquinas’s 5 ways.
From Huenemann’s first paragraph, I then see no reason why Hume should reject the use of teleology in common life and science (so long as we understand teleology properly, as immanent teleology). In fact, I would argue, immanent teleology is necessary for articulating the efficient causality Hume seems comfortable enough using, and I would argue that “almost all knowledge depends” on using immanent teleology.
To the second paragraph: Here the worry is not so much about using causal principles per se, but more generally in his “reduced empiricism.” Hume’s notion that all knowledge begins in experience is not, of course, new. Aristotle thought this, as did the medieval scholastics. I call Hume’s empiricism “reduced” because he takes this as a greater limit that the scholastics or Aristotle did. When Aquinas says that nothing is in the intellect that was not first in the senses, he did not think that meant knowledge was thereby confined to what is immediately sensed. Rather, the sense was the “instrument” through which we discerned reality. I know about the nature of horses through the sense, though the nature of horses is not itself sensible. But for Hume we are much more narrowly confined.
[I wonder, Huenemann, how this might play in to your recent – and stunning – willingness to play ball with formal cause?]
Why the narrow confinement? Here is where I want to push Huenemann again. If it is fair enough to extend beyond immediate experience in the case of science (since we have no experience of the future and cannot justify that the future will resemble the past), why is it not okay to extend beyond experience in the metaphysical as well as we do in the scientific? If we can use the principle of the uniformity of nature in science, why can’t we use it in metaphysics? I know Hume thinks metaphysics is a fairy land, but does he have any reason for thinking it is more of a fairy land than the natural sciences, given that both leap beyond the bounds of our immediate experience? What is sauce for the goose is sauce for the gander … if the principle is legitimate in the natural sciences (despite our ultimate inability to justify it), then it is legitimate in metaphysics.
Now you might say that you won’t get a very thick theology out of an immanent teleology, and I would be inclined to agree. Perhaps Hume agrees to, given his leaving the door a bit ajar in the Dialogues. But I think Hume underestimates how much mileage you can get out of those arguments. I think you would get the sort of “negative theology” you find in the first 20 or so questions in the Summa, which is as thick of a negative theology as you’ll find and it moves the ball downfield a far way.
This is why I think it is relevant that in Huenemann’s initial post he framed Hume’s view of causality as one not just of “usefulness” but one of “truth” (albeit a truth we cannot justify). If you just read Hume as thinking causal principles are useful but we have no reason for thinking them true, you more obviously put the kabosh on future metaphysical endeavors.
I think Hume’s advice is to try to get by with as little metaphysics as possible. So when you ask “Why can’t we get away with more?” his reply is, “Because the more you try to get away with, the greater likelihood that you’re wrong.” I believe he certainly would rule against scientists to the extent their theories don’t help to predict or control events. String theory, for instance. And he’d rule against science that seems to contradict, in some outrageous way, common experience (such as the claim that time isn’t real). It’s more likely that the theorists have gotten confused somewhere.
I don’t know about truth vs. usefulness. I think Hume thinks there is a truth about the world (well, I just don’t understand anyone claiming otherwise), but we can never know if the theory we’ve worked up is really, really true, but it might be, and at any rate, it’s all we can do.
It sounds like Hume (according to Professor Huenemann) would argue that metaphysical theories go out on a limb with no capacity to judge the stability of that limb whatsoever, since no one in the material world can really experience transcendent reality (if there is one). They can only ever be a shot in the dark. Scientific theories have a better idea of the girth of their limb, since they are used to explain a world that we can perceive. That’s what distinguishes the value of metaphysics from that of science. That seems a reasonable position to take, if you believe that the material world we perceive gives us no evidence of a transcendent reality. On the other hand, if you believe that in our material world we do have some evidence for the existence of some transcendent reality, there does seem to be some room for cosmological or teleological arguments. The point at which they cease to explain the world we perceive would be the point where they enter into “fairy land.” The question becomes: How much mileage can you get out of them before that point? Aquinas though he could get pretty far. Hume thought that as soon as you invoke some transcendent reality that we can’t perceive, you’ve gone off the reservation. I guess it just comes down to how far it’s logically permissible to leap.
Or Professor Kleiner is right, and Hume would argue that there is no evidence that we have any valid perception of Truth whatsoever, in which case we do what appears to be useful, but it’s all pretty tenuous. Science has the advantage of being verifiable, and producing iPhones, which makes it more useful. Metaphysics does not appear to be useful at all, since our theories are unverifiable shots in the dark that will never produce anything. That would seem to be more damning to religion in general.
I’ll try to anticipate Kleiner’s response, and reply to it. “But scientists do not make fewer metaphysical assumptions than metaphysicians do. They just aren’t aware of making them, and don’t focus on them or make them explicit.” But I think this, in a sense, might be Hume’s point.
When we figure out what metaphysics a scientific theory commits itself to, we don’t just “read off” the metaphysics from the theory. It takes a fair bit of theorizing, especially when the scientific theory is instrumentalistic (meaning, concerns itself just with predicting measurements, and going as lightly as possible with regard to ontology). So I might begin with a theory that switch X controls the light bulb. I try it out a number of times, and, yep, there seems to be a connection. What am I committed to? Well, switch X, the light bulb, and an apparent “control” relation. I could go deeper, into a theory about hidden wires, electrons, the telos of electrical flow, and on and on. But am I “committed” to these theories? It all depends on the arguments linking my superficial beliefs to them. And that is where Hume is skeptical. He thinks those arguments can easily go astray, and are more likely to get more and more lost as we get further away from measurements.
I came across an excellent example of this recently by following a link to a recent post on Edward Feser’s blog, where he chastises physicists for thinking in sloppy ways about what “nothing” is. (Basically, they make claims about it having parts, being homogeneous, and so on, and he says they don’t know what “nothing” means.) After his post follow 78 comments, and down the rabbit hole we go, until it becomes quite clear that really nobody knows what they are talking about. I am NOT calling them dumb, or foolish; I am saying that it is so easy to get confused, to draw conclusions that really don’t follow, and to get dizzy over wordplay when we can’t check our claims against any kind of experience.
I think you sell short the metaphysical assumptions of science. It isn’t just that they don’t ask a further question about “hidden” things. What people like Dennett do is assume materialism without ever making an argument for it at all. It becomes something like an unthought default position.
So isn’t your claim here just this: assuming a materialist metaphysics is not as bad as arguing yourself into a hylemorphic metaphysics?
If that is not too unfair, why in the world should I agree to that? Particularly when explaining the sort of efficient causality Hume is willing to accept as true (on your account) will require immanent teleology. And, to boot, any knowledge claim at all will have to be “Platonic” in the sense of requiring formal cause. In other words, an account of what we are doing when we make knowledge claims just begs for hylemorphism.