Monday 9 September 2013

On causation

This is a redaction of a conversation ongoing in the comment thread on a post of Ed Feser here.

Feel free to join in!

23 comments:

  1. If Aristotelians want to extend the common sense notion of cause into our understanding of the world of atoms and particles, then it seems that a fight with 20th century physics (or whatever metaphysic supports it, if you will) is inevitable, as the present discussion and an earlier one with physicist Robert Oerter would suggest. Indeed, I think one can make the case that physics abandoned the notion of cause in favour of description back in the 19th century with Maxwell's equations. The idea holds on in macroscopic physics with the Einsteinian notion that causal effects propagate no faster than the speed of light but now it seems that in quantum physics 'spooky action at a distance' is a reality. What we are to make of this I don't know. Maybe the fault lies within us and there is ineradicable indeterminacy of explanation: complete but shallow versus deep but incomplete, never deep and complete.

    ReplyDelete
  2. Brandon said

    A few points:

    (1) Your comment at September 3, 2013 at 5:30 AM is interestingly inconsistent, and shows some of the problems of talking in this area without precision (which, certainly, is difficult in commments boxes). You say in three consecutive sentences (a) that talking about causes in the context of atoms and particles will require picking a fight with 20th century physics, which requires an inconsistency between the two; (b) that the case can be made that physics abandoned the concept of cause in favor of description in the nineteenth century, which requires that any talk of causes is simply in a different logical domain from anything physics is talking about; and (c) that macrophysically modern physics holds that causal effects propagate no faster than the speed of light and that quantum physics involves causation in the sense of 'action at a distance'. Each of these is a plausible thing to say, and one can certainly find physicists who will say them; and thus we have all the tell-tale signs of equivocation right here: the word 'cause' is simply not being used consistently across the board. There is no possible way to make any progress in any direction without clarifying this first; until equivocation is eliminated, all talk on the subject is hopeless.

    (2) An Aristotelian account of causation is not an "extension of the common sense notion of cause" into anything; the Aristotelian account of causation recognizes the common sense notion of cause as genuine, but it is a logical error to assume that one can therefore move in the opposite direction. Aristotle is not common sense; it just follows on Aristotelian principles that common sense is not wrong as far as it goes, not that common sense goes as far as any explanation can go.

    (3) It's a mistake to assume that realism directly implies 'classical realism', which is a specific kind of assumption about physical theories. The collapse of classical realism does not imply the collapse of the entire family of realisms. (Aristotelianism, however, does not directly imply anything on this question for this particular topic; it's just that without considering these issues, it's premature to assume that anyone has the definitive interpretation of the discoveries in question, which is necessary before we can say anything about what metaphysics they require.)

    (4) It is entirely possible that an Aristotelian account would need revision in light of findings of physics; Aristotelian principles themselves don't rule it out. This has to be established rather than simply asserted or assumed from the beginning, however, and for obvious reasons Aristotelians are reluctant to simply take appearances at face value in these matters, having a centuries-long history of being attacked on the basis of apparently contrary phenomena that turned out to be no more than apparently contrary. It is an error, however, to assume that even inconsistency with physical theory is a "fight with physics", which is an atavistic anthropomorphism; all that results is a physical puzzle, which either may be soluble with further research or may require modification of the account. Either way, it needs to be established on the specifics, not simply jumped to on the basis of one's general sense of what things imply.

    That said, it's good to raise the questions

    ReplyDelete
  3. Urban Jean said

    Thank you, Brandon, for treating my somewhat slapdash remarks seriously. It would have been clearer had I forthrightly claimed that all of physics can be formulated without reference to cause. There is nothing in the differential equations of classical electromagnetism, say, or general relativity, that looks like a cause. In discussing the solutions to the equations one might describe what's going on in the familiar language of common sense efficient causality: moving charges cause the radiation of electromagnetic waves which cause remote charges to accelerate. The equations merely say that moving charges and radiation co-occur in certain ways. Quantum indeterminacy makes causal language quite inappropriate even in an informal presentation---particles are not understood as pieces of macroscopic matter writ very small. Quantum entanglement---Einstein's spooky action at a distance---is not understood in terms of causal effects travelling faster than light but as unexpected correlations between widely separated measurements. Upshot: physics is not formulated in causal terms at all.

    You lose me, I'm afraid, with your paragraph (2). I confess I get my Aristotle second hand. In Natural theology, natural science, and the philosophy of nature, section (4), Ed says

    The Scholastic tradition had worked out a complex and sophisticated theory of causation . . . Now the moderns gradually chucked out almost all of this nuance -- which, despite its complexity, is really just a systematic articulation of common sense thinking about causation -- as they unpacked the implications of their anti-Aristotelian revolution.

    We are left with effiicient causation which Hume then rather presciently deconstructs.

    I can't usefully add to Dianelos's excellent presentation on micro-world acausation except to offer a reference: the SEP entry on Bell's Theorem

    ReplyDelete
  4. Scott said


    "It would have been clearer had I forthrightly claimed that all of physics can be formulated without reference to cause."

    It would also have been more correct.

    On the other hand, it would also have deprived the claim that some quantum events are "uncaused" of any interest it ever might have had.

    ------------------------------------------------------

    In particular, it appears impossible to square your statement:

    "[P]physics is not formulated in causal terms at all."

    with Dianelos's:

    "[P]hysics does not prove that there are uncaused events; it does only suggests that there are uncaused events. But that suggestion is much stronger than many realize."

    How a science that isn't "formulated in causal terms at all" could "suggest" anything whatsoever about "uncaused events" is a mystery I confess myself unable to solve; indeed it seems to leave metaphysics right where it was all along. Are you sure you're agreeing with Dianelos here?

    -------------------------------------------------------

    "We are left with effiicient causation which Hume then rather presciently deconstructs."

    In what sense was this deconstruction prescient? Do you mean that he anticipated modern views?

    ReplyDelete
  5. Brandon said

    I don't think Ed's claim is correct.

    The fundamental problem with taking a positivist interpretation of physics in general is that it guarantees that physics is irrelevant to causal discussions; you can't even rule something out unless you're speaking a language that can formulate it. In general when people do appeal to physics on these matters that do assume that the equations describe causes and also that (1) causes must be deterministic, (2) there are no constituent causes, and (3) all causes are of the same kind and have the same account; Aristotelians reject all three of the latter, though.

    A minor matter, that's neither here nor there -- the efficient cause was for all practical purposes scrapped prior to Hume; a significant portion of Hume's arguments are derived directly from Malebranche, who argued that only God could possibly be an efficient cause, and thus that everything else simply operated according to contingent general laws. Hume's primary innovation was making the general laws features of the mind rather than of reality. Of course he is also the one people looked back to when reviving the view under positivism, so it's not surprising that his name comes up.

    ReplyDelete
  6. Urban Jean said

    Thanks, Brandon. I appreciate that you know a great deal more about early modern thought than I'm ever likely too. I'll have to leave the historical issues to you and Ed. What you say about Hume making general laws features of the mind chimes somewhat with what I would want to say. The obvious interpretation of the success of physics is that its equations do indeed capture the regularities we believe exist in nature. Where does this leave causation? Well, if brains are good at detecting correlations within themselves---perhaps a baby learns that the fist in its visual field is its fist because its brain detects correlations between neural activity in its visual cortex with activity in its motor cortex---then perhaps our concept of causation can be traced back to our appetite for correlation. The physicist would not deny that there are correlations in nature, even if he is sceptical of causes. However, if a sophisticated Aristotelianism can accommodate non-deterministic causes, as you suggest, then there is clearly room for an alternative account.

    ReplyDelete
  7. Scott said

    "Where does this leave causation?"

    Right where it was.

    "[P]erhaps our concept of causation can be traced back to our appetite for correlation."

    (1) One of your unstated premises here is that something about modern science shows the concept of causation to be wrong (or at least reducible to something else). But according to your own view, you're not entitled to that premise. If modern physics is silent on causation, then it's silent on causation.

    (2) The concept of "causation" you're implicitly relying on here is entirely modern; Aristotelian causation isn't based on "correlations" between activities or events. So even if you were successful in arguing against the modern conception, A-T causation would be completely untouched. In fact Thomists would agree in principle with your conclusion (though not your way of arriving at it).

    -----------------------------------------------------

    "However, if a sophisticated Aristotelianism can accommodate non-deterministic causes . . . "

    There's no special sophistication required; Aristotelian-Thomist causation has never been "deterministic" in the first place. But non-deterministic causation is still causation, not lack of a cause altogether.

    ReplyDelete
  8. Brandon said

    I actually think that's probably where the naturalist has to go in these matters. One of the difficulties with it is that all the obvious reasons why one would think that scientific theories have anything to do with reality are causal, because causal reasoning is the only kind of reasoning to the existence of something unknown that has been historically regarded even as reasonably accurate. I do a lot of work with the history of the problem of the external world -- how we can even know that there is anything other than our own minds -- and the only accounts that are neither obviously untenable (like solipsism) nor strongly causal are at least minimally idealist, holding that we at least don't know that anything other than minds exist. The reason is that when we switch from causation to correlation we're switching from a kind of reasoning that by its nature goes from known to unknown to something that merely deals with the coherence of things known -- and getting only so far as the coherence of our own thoughts and sensations means we don't actually ever deal with any reality other than our own minds. A second difficulty is that none of our most plausible accounts of evidence -- what makes something evidence, what counts as evidence and why, etc. -- can't be disentangled from causal reasoning. Since correlations are promiscuous we even usually judge the evidence-value of correlations by how useful or significant they are for causal reasoning. Even when scientific theories don't mention causal concepts, it's at least extraordinarily difficult to make sense of scientific experiments without causal terms, for precisely the same reason: we need a ways to distinguish genuine experiments from pseudo-experiments, for instance, but historically they've always been distinguished causally.

    Tearing out causal reasoning, in other words, is not a minor matter -- it leaves us pretty much marooned, rationally speaking, and having to build everything again literally from scratch. I do think it would be interesting to see how far one could get along such lines (Hume, as you note, did some interesting work in this direction, as did the early positivists -- but they also gave up on scientific theories being any more than useful classifications of experiences) -- it's just that it's a major project rather than a minor revision.

    ReplyDelete
  9. Urban Jean said

    Scott says that physics is no guide to our understanding of 'causation'. I beg to differ. Physics is in the business of offering explanations of certain natural phenomena. Suppose we heat some gas in a closed container. Its pressure rises. Why? The explanation involves an understanding of what heat and pressure are in terms of the motions of the particles we understand constitute the gas. How does this fit with the Aristotelian taxonomy of explanation? I don't think it fits at all. It's not that the Aristotelian scheme is wrong. It's more that it is irrelevant or even inapplicable. It's troubling that a system that lays claim to universality falls short in this regard. But neither does the explanation fall back on the common sense notion of 'cause' that some say is a spavined descendent of the Aristotelian efficient cause. True, my remarks to date have been directed at this modern conception of cause. Everywhere we are offered 'causal' explanations of natural phenomena I believe physics offers a better explanation that makes no reference to causes. I won't say causes can be reduced because in the end I doubt they are there in any fundamental sense. Further, in so far as the germ of the modern concept lies within the Aristotelian efficient cause then to that extent reality eludes the earlier concept too. But my main point is that the form of explanation we find in physics and the natural sciences generally lies outside the Aristotelian framework.

    ReplyDelete
  10. Scott said

    "Scott says that physics is no guide to our understanding of 'causation'. I beg to differ."

    Then you're differing with yourself, as you've already made quite a lot of hay out of the fact/claim that physics can be entirely formulated without reference to causation.

    "Physics is in the business of offering explanations of certain natural phenomena."

    It is? What happened to your earlier view that "physics abandoned the notion of cause in favour of description back in the 19th century"? Description isn't explanation.

    "Suppose we heat some gas in a closed container. Its pressure rises. Why? The explanation involves an understanding of what heat and pressure are in terms of the motions of the particles we understand constitute the gas. How does this fit with the Aristotelian taxonomy of explanation? I don't think it fits at all."

    I'm more than a little bemused by your claim that an explanation in terms of "motion" doesn't "fit[] at all" with Aristotelian metaphysics.

    ReplyDelete
  11. Urban Jean said

    Slow down a bit Scott. That's just the opening sentence! The rest of the piece elaborates on why I think the kind of explanation we find in physics extends causation understood as the Aristotelian taxonomy of explanations.

    Do you think that the kinetic theory, which is largely descriptive of what gases are, has no explanatory value?

    If you do think the kinetic theory explanatory then which of the four Aristotelian kinds of explanation does it fall under do you think?

    ReplyDelete
  12. Scott said

    "Slow down a bit Scott. That's just the opening sentence!"

    I dealt with what followed in the rest of my post.

    "If you do think the kinetic theory explanatory then which of the four Aristotelian kinds of explanation does it fall under do you think?"

    All four of them in one way or another, but primarily efficient causation. A gas doesn't qualify as a "substance" according to A-T metaphysics; all of its properties reduce to the (local) motion of its constituent particles, and the potential for motion of any one particle is actualized by the other particles as efficient causes.

    Of course this brief reply isn't intended to be complete, and others may add to it or correct it.

    ReplyDelete
  13. Urban Jean said

    Hello Brandon, if still reading. Agree entirely about the ubiquity of causal reasoning. Not so sure that "causal reasoning is the only kind of reasoning to the existence of something unknown that has been historically regarded even as reasonably accurate". How does this sit with the hypothetical-deductive reasoning in physics? Regarding causal reasoning in the interpretation of experiments, perhaps we could say something like this: suppose ordinary causal reasoning is in some sense approximate, but good enough to enable us to climb up to the acausally expressed laws of physics. We then kick away the causal ladder and take the postulates and laws of physics as axiomatic. From this foundation we explain how it comes about that the world looks causal to creatures like us and why this is a good approximation. This might be thought circular, but I don't think it is. But, as you say, it's a major project.

    ReplyDelete
  14. Brandon said

    How does this sit with the hypothetical-deductive reasoning in physics?

    It's a complicated matter, but to simplify a bit, it doesn't extend beyond already known bounds; that is, it's a process of establishing that a known phenomenon fits into a known pattern or regularity. Or to put it in other words, it isn't a matter of saying that, given a known A, some as-yet-unknown thing should also be there; it only tells you that, given a known A, you already know that B should also be there. This is a much weaker sort of reasoning. To get to new territory with hypothetico-deductive reasoning people usually use analogy, which is not generally regarded as reliable, or causal reasoning, or both.

    The scaffolding approach is interesting, but it looks very much like it requires that causal claims in fact be true -- that they are true only to a degree of approximation doesn't make them any different from almost any other true claims that we make, and it seems that they would have to be true, rather than just heuristics, to be trustworthy enough to the point that we could take the laws we're getting as true themselves rather than just as useful for calculating certain kinds of problems. I doubt, too, that there's any point at which we can actually round the curve -- i.e., given the nature of physics research there will never be a point at which we can take the laws of nature as axiomatic rather than as just being true to at least the best approximations we know on the basis of these experiments that we explain and justify causally. Scaffolding that stays forever is really just an architectural feature. It would be a different matter if a noncausal general theory of evidence could be developed; it's the lack of this that's the real problem, since we're never going to get to a point where evidence is not an essential part of experimental reasoning.

    ReplyDelete
  15. Urban Jean said

    Hi Brandon. The sense in which I think ordinary causal thought is 'approximate' is this. Physics, which I take as my starting point, gives us a very fine-grained description of the world. A macroscopic lump of matter that we can see and feel and heft contains billions of billions of elementary parts. Any lump of matter, but especially living things, undergoes a process of continual exchange of energy and parts with its surroundings. To carve the world into such lumps and to explain what happens in terms of interactions between the lumps is to constrict oneself to an extremely coarse-grained description, which is thus necessarily approximate. For most of us, for most of the time, this is perfectly adequate. Given what we are, our ability to manipulate the world is extremely coarse (in terms of elementary parts) and the communication bandwidth between individuals so low that to be more precise would incur significant costs in time. It just doesn't matter. Unless, say, one is engaged in some extremely delicate scientific or engineering endeavour. To say that striking a match causes a flame leaves so much out: dryness of match, force of strike, roughness of surface, presence of oxygen, and so on. Nevertheless, we have got to the point where a fine-grained description is possible, and we can look back and see why the coarse-grained is sometimes inadequate. And we can take this step, I think, even if we regard the hypothetical fundamental parts as purely instrumental---we can predict how approximate the coarse-grained picture actually is. Someone earlier asked "has physics been reduced to a game where men make the rules rather than trying to discover them? " My reply: physics is a game in which we guess the rules rather than deducing them. It's the best we can do given what we are.

    ReplyDelete
  16. Scott said

    "To carve the world into such lumps and to explain what happens in terms of interactions between the lumps is to constrict oneself to an extremely coarse-grained description, which is thus necessarily approximate."

    Where the underlying reality is fine-grained, this is of course true, and no Aristotelian or Thomist would deny it. But a good deal of the world divides naturally into "lumps" (substances), and in those instances higher-level explanations are not only possible but necessary.

    The paradigmatic example is of course a living organism, of which you could conceivably know the states and motions of every subatomic particle and still not know what it was doing in the relevant sense. For example, if I'm swinging my fist into someone's face, nothing in physics is going to tell you whether I'm hitting him deliberately or accidentally, or acting courageously or recklessly.

    If you have a disagreement here, it's going to be about this sort of anti-reductionism, not about the A-T account of causation itself.

    ReplyDelete
  17. Urban Jean

    No, Scott, I really do have a beef with Aristotelian (efficient) causation, as far as I understand it. My alternative account may lead me to a reductive view of living things and you would no doubt regard that as a reductio ad absurdum of my account of causation. But at the moment I'm just trying to home in on what I find deficient not just in Aristotelian causation but present-day common sense causation too.

    ReplyDelete
  18. Scott said

    Then I don't understand your objection. What you've written is this:

    "[When] one is engaged in some extremely delicate scientific or engineering endeavour[, t]o say that striking a match causes a flame leaves so much out: dryness of match, force of strike, roughness of surface, presence of oxygen, and so on. Nevertheless, we have got to the point where a fine-grained description is possible, and we can look back and see why the coarse-grained is sometimes inadequate."

    Of course. Again, why would you expect an Aristotelian or Thomist to disagree, and how does any of this conflict with Aristotle's fourfold account of causation?

    No A-T theorist would claim that merely knowing the four sorts of cause is sufficient for the purposes of science; metaphysics isn't physics, and nobody has said otherwise. "Striking a match causes flame" is an illustration of a metaphysical principle, not a scientific theory in its own right.

    But the fact that a scientific account is much more detailed doesn't in and of itself mean that the more detailed account doesn't fit the Aristotelian scheme, and I have yet to see any clear and coherent claim that it doesn't. Indeed, the evidence so far suggests that you may simply be insufficiently familiar with that scheme, as indicated by what appeared to be your earlier failure to appreciate the central role that motion (not limited to, but most definitely including, local motion) plays in Aristotelian metaphysics.

    ReplyDelete
  19. OK, Scott, on we go.

    I imagine metaphysical principles are of value and interest by virtue of the inferences we can draw from them. From there is a struck match can we infer there is a flame? No. From there is a flame can we infer there is a struck match? Again, No. Just what inferences can we draw from striking a match causes a flame? I'm not sure there are any, and this suggests to me that the central term 'causes' has no meaning. We do more or less understand 'match', 'strike', and 'flame'. So I have a fairly radical beef with 'causation'. I don't think it captures anything real in the world.

    The clear and I hope coherent claim that the physics account doesn't fit with Aristotelian or common sense causation is the observation that it contains no sentences of the form a causes b.

    ReplyDelete
  20. Scott said
    "I imagine metaphysical principles are of value and interest by virtue of the inferences we can draw from them."

    Not really—and not at all, if you're specifically referring to inferences about physical facts. In the case of the match and the flame, all the fourfold account of causation does is tell us what an explanation will look like or how it will be structured. (And of course not all four types of cause need be present in every instance.)

    Moreover, no one has ever seriously proposed that striking a match causes a flame is true under all conditions or none (except in the loose sense of "causes" as "tends to bring about if other conditions are right"). Obviously some other things have to be the case in order for a flame to result; some of those things may count properly as "conditions," but some of them may well be other causes (or, perhaps equivalently, parts of the cause). A fully detailed description of how a match causes a flame will even include intermediate causal steps, as the physical process itself surely doesn't just go directly from "match" to "flame" but includes some chemical and thermodynamic stuff in between.

    But that clearly doesn't mean the term "causes" has no meaning at all, or even that it's used incorrectly in the statement striking a match causes a flame. The meaning of "causes" is what it's always been: something making something else happen. And a full physical account positively helps us to see how striking a match "makes" there be a flame, giving the term "causes" the very precision you correctly note that it lacks when we use it broadly.

    "The clear and I hope coherent claim that the physics account doesn't fit with Aristotelian or common sense causation is the observation that it contains no sentences of the form a causes b."

    The account may not contain any such sentences, but that doesn't make it not fit with Aristotelian or common-sense causation. It doesn't contain any sentences about music, either. Does that mean physics is incompatible with music? (In fact a physical account of sound goes a long way toward helping us to understand music, whether it contains any sentences about "music" or not. So I'd hardly say they don't fit together.)

    In order to show that there's a problem here, you need to identify some positive mismatch between the physics and the metaphysics. I'm not seeing one.

    ReplyDelete
  21. Scott, I understand that 'all the fourfold account of causation does is tell us what an explanation will look like or how it will be structured.' If an explanation has an efficient cause component then this will presumably contain assertions of the form x causes y. These themselves may be supplemented with or elaborated into further statements of this form, but you appear to be saying that an efficient causal explanation has to be 'causes all the way down'. The claim that an explanation must have this form is obviously in some tension with our acceptance as explanations of physics accounts in which causes do not appear. This is the mismatch I'm try to get you to see, and why I said much earlier that cause-free physics accounts are a form of explanation not encompassed by Aristotle's doctrine, and hence the latter must be considered inadequate

    ReplyDelete
  22. Scott said

    "This is . . . why I said much earlier that cause-free physics accounts are a form of explanation not encompassed by Aristotle's doctrine, and hence the latter must be considered inadequate."

    I'd say the inadequacy, if there is one, is rather the other way around—at least if such accounts are offered as explanations rather than descriptions.

    Then again, if I recall correctly, Aristotle himself didn't use the term cause for any of the four causes either; it was the Scholastic philosophers who turned (for example) Aristotle's ὑλη into causa materialis—"material cause" rather than just "material." (Not that Aristotle didn't use the term αἴτιον at all, of course, but he didn't require it to appear in an explanation.)

    At any rate I'm not seeing anything in a mere cause-free description that Aristotle can't account for. Nor do I see any reason to think that just because even an (attempted) explanation may not include statements of the form x causes y, it doesn't comport with Aristotelian metaphysics. For example, as I mentioned in another post, if physical laws are really supposed to be laws, then they include final causation whether that's expressly stated or not.

    ReplyDelete
  23. Can we get clear on this, Scott? In your view the accounts that physics gives for natural phenomena have no explanatory value?

    ReplyDelete