Monday, August 23, 2010

Edwards on infinite causal series

In a comment on a post down below, Richard Hennessey asks me to reply to the criticisms of Thomistic cosmological arguments presented by Paul Edwards in his article “A Critique of the Cosmological Argument” (criticisms Hennessey has endorsed in a recent post over at his own blog). The relevant sections of Edwards’ essay are in part III, wherein he responds to the Thomist philosophers G. H. Joyce and R. P. Phillips – something for which Edwards deserves credit, given that most atheist writers not only do not address the arguments of Thomists, but seem unaware even of their existence. Unfortunately, Edwards still deeply misunderstands how Thomistic cosmological arguments are supposed to work.

Edwards does realize that Aquinas is not arguing that the universe must have had a beginning – that the first cause he is arguing for is “first” not in a temporal sense, but in an ontological sense, a sustaining cause of the world here and now and at any moment at which the world exists at all. Still, he thinks this fails to address the heart of the atheist’s critique. For why might a series of causes existing simultaneously, all here and now, not be as infinite as a temporal regress of causes might be (by Aquinas’s own admission)? And if there must be a first uncaused cause in the order of simultaneous causes, why could it not be something other than God, such as basic material particles or gravitational forces?

The very asking of these questions shows that Edwards does not understand the distinction between causal series ordered per accidens and causal series ordered per se, on which the Thomistic arguments (like other Scholastic cosmological arguments) crucially depend. Following Joyce, Edwards refers instead to the distinction between causes in fieri and causes in esse, which might be part of the problem. Not that there is anything wrong with Joyce’s terminology, but it might suggest to someone otherwise unfamiliar with the Thomistic and Scholastic arguments – as it apparently did to Edwards – that the difference between “becoming” and “being” is what is supposed to be the key to seeing why the second sort of causal series cannot in the Thomist’s view be infinite. And that is not the case.

What is key is the distinction between instrumental and principal causality (or second and first causality), a distinction which the language of per accidens versus per se (which I use in The Last Superstition and Aquinas) better conveys. An instrumental cause is one that derives whatever causal power it has from something else. To use Aquinas’s famous example, the stick that the hand uses to push the stone has no power to push the stone on its own, but derives its stone-moving power from the hand, which uses it as an “instrument.” (Of course, the stick might have some other causal powers apart from the hand; the point is that relative to the specific series hand-stick-stone it has no independent causal power.) A principal cause is one that does have its causal power inherently. The hand in our example can be thought of for purposes of illustration as such a cause, though of course ultimately it is not, since its power to move the stick depends on other factors. Indeed, there can at the end of the day be only one cause which is principal or non-instrumental in an unqualified sense, namely a cause which is purely actual and thus need not be actualized in any way whatsoever by anything else. In any event, it is because all the causes in such a series other than the first are instrumental in this way that they are said to be ordered per se or “essentially,” for their being causes at all depends essentially on the activity of that which uses them as instruments. By contrast, causes ordered per accidens or “accidentally” do not essentially depend for their efficacy on the activity of earlier causes in the series. To use Aquinas’s example, a father possesses the power to generate sons independently of the activity of his own father, so that a series of fathers and sons is in that sense ordered per accidens rather than per se (though each member of such a series is also dependent in various other respects on causal series ordered per se).

So, it is ultimately their instrumental character, and not their simultaneity, which makes every member of a per se ordered causal series other than the first depend necessarily on the first. To be sure, the paradigm cases of causal series ordered per se involve simultaneity, because the simultaneity of the causes in these examples helps us to see their instrumental character. And the Thomist does hold that the world must ultimately be sustained at every instant by a purely actual uncaused cause, not merely generated at some point in the past. For these reasons, Thomists tend to emphasize simultaneity in their explanations of causal series ordered per se, as I did in The Last Superstition.

But it is arguably possible at least in theory for there to be a per se causal series in which some of the members were not simultaneous. Suppose a “time gate” of the sort described in Robert Heinlein’s story “By His Bootstraps” were possible. Suppose further that here in 2010 you take a stick and put it halfway through the time gate, while the other half comes out in 3010 and pushes a stone. The motion of the stone and the motion of the hand are not simultaneous – they are separated by 1000 years – but we still have a causal series ordered per se insofar as the former motion depends essentially on the latter motion. I am not saying that this really is possible, mind you; it presupposes that time travel itself is at least possible in principle, which is controversial at best. But let’s grant it for the sake of argument. Insofar as the hand’s operation and existence will themselves presuppose various other factors, we have a continuation to the regress of causes ordered per se which cannot be ended until we reach a purely actual uncaused cause. The end result is the same, even if the statement of the argument needs to be made more complicated.

Now Hennessey, unlike Edwards, does see that it is the instrumental nature of the causes rather than their simultaneity that is doing the metaphysical work here. What Hennessey doesn’t see is that this completely undermines whatever force he thinks there is in Edwards’ critique. Edwards thinks that a series of causes in esse no more needs to have a first member than a series of causes in fieri does, and it is evident that he thinks this because he assumes that in the first sort of series no less than in the second, “every member is genuinely the cause of the one that follows it” (as Edwards puts it at the end of part III). Now if this assumption were correct, then it would indeed be odd for Aquinas to hold that a series of causes in fieri might be infinite while a series of causes in esse (or, better, a series of causes ordered per se) could not be. For it is precisely because they have their causal power independently of any earlier members of the series that Aquinas argues that the activity of the members of a per accidens series need not be traced to a first cause. So, if he thought that the members of a series of per se causes also had independent causal power, then his reason for tracing that sort of series to a first member would be undermined. But of course, that is not what Aquinas thinks. He thinks that they do not have such independent causal power, and so (contrary to what Edwards suggests) it is not at all odd, arbitrary, or unjustified for him to say that a series ordered per se needs to trace its activity to a first uncaused cause. Edwards misses this because he thinks that the Thomistic argument rests on an appeal to simultaneity, and Edwards doesn’t see how simultaneity requires an uncaused cause. But as I have said, the argument doesn’t rest on simultaneity as such. It rests on the instrumentality of the members of a causal series ordered per se, and instrumentality does require an uncaused cause.

[As a side note, this does not mean that there is no sense in which the members of a causal series ordered per se are genuine causes; Aquinas is not an occasionalist. But how his account avoids occasionalism is a separate issue, and does not affect the soundness of Thomistic cosmological arguments as such.]

It seems to me that the reason Hennessey misses this no less than Edwards is that he puts too narrow an interpretation on the word “first” in the expression “first cause.” He seems to think that what Aquinas was concerned to show is that if you lay out a series of causes ordered per se in a straight line, the line will necessarily have a beginning. But that is not what he was concerned to show. As Thomists sometimes point out, it wouldn’t change things in the least if we granted for the sake of argument that a series of causes ordered per se might loop around back on itself in a circle, or even that it might extend forward and backward infinitely. For the point is that as long as the members of such a circular or infinite chain of causes have no independent causal power of their own, there will have to be something outside the series which imparts to them their causal efficacy. (As the Thomist A. D. Sertillanges once put it, a paint brush can’t move itself even if it has a very long handle. And it still couldn’t move itself even if it had an infinitely long handle.) Moreover, if that which imparts causal power to the members of the circular or infinitely long series itself had no independent causal power, then it too would of necessity also require a principal cause of its own, relative to which it is an instrument. This explanatory regress cannot possibly terminate in anything other than something which has absolutely independent causal power, which can cause or “actualize” without itself having to be actualized in any way, and only what is purely actual can fit the bill.

That is the way in which it is “first” – first in the sense of being metaphysically ultimate or fundamental, and not (necessarily) in the sense of standing at the head of some (temporal or even non-temporal) queue. That is also why, contrary to what Edwards and so many other atheists suppose, it makes no sense to ask why fundamental physical particles or the like might not be the first cause. Particles and other “naturalistic” candidates for the ground floor level of reality are all compounds of act and potency, form and matter, essence and existence; accordingly, they are in need of actualization and are therefore necessarily less than the “pure act” or Subsistent Being Itself which alone could, even in principle, be that which causes without in any way being caused (or, as I would prefer to say, which actualizes potency without itself being actualized).
For those who are interested, I discuss these issues at greater length in Aquinas, which provides a more thorough, in-depth, and “academic” treatment of the Thomistic arguments for God’s existence than The Last Superstition does. (And without the jokes, polemics, or conservative political asides that some readers – including Hennessey – feel they could live without. But Hennessey does say some kind things about The Last Superstition and about this blog, for which I thank him.)


  1. Couldn't you cut the entire explanation short by saying the universe is a mixture of act and potency and requires actualization? Is talk of chains a little bit distracting since presumably every item in the chain needs its own actualization at every instant, regardless of what process its undergoing?

    1. They would say that we then would've committed the fallacy of composition. Just because we observe that parts of the universe are composities of act and potency, doesn't mean the universe as a whole is.

  2. This is much appreciated. Let me ponder.

  3. This comment has been removed by the author.

  4. "All the balls on my infinitely large pool table are in motion but nothing imparts motion to them."

    "All the rifles in my infinitely large army are firing but no triggers were pulled."

    "All the coins in my infinite fortune are in one bank but no one minted them."

    That is how infinitist objections to the CA sound to me.

    Why is motion a given? Why are causal relationships a given?

    Things having causal bearing on one another is an effect of "the Law of Causation." This Law is, then, a principle of the world which cannot be subject to causation. As such, it is an unqualified causer of all causes.

  5. Hey Codgitator. Some people would say that it there is a possible world where nothing is caused. It's not necessarily reality, but it's not logically incoherent.

  6. Three examples that might be helpful.
    1. An infinite series of moons cannot shine. It does no good to say the moon closest to us is reflecting the light of a prior moon, and that of a prior moon, and so on. An infinite series of non-luminous bodies cannot be a source of light. If the last moon is shining, there has to be a sun somewhere up the line, the first cause of the light.

    2. An infinite series of mirrors cannot reflect a non-existent bear. Suppose I'm shaving, and see a bear reflected in my mirror apparently walking through my garden. I go into my garden, and find no bear, and no bear prints, but another mirror. I trace that back, and find yet another mirror, and so on. It does no good to try and explain the bear I saw by postulating an infinite series of mirrors each reflecting the image of a bear from a prior mirror. There has to be a first mirror that first reflected an actual bear. No bear, no reflection of a bear.

    3. An infinite series of flat bed rail cars cannot accelerate itself. Suppose I come out of a the woods to find a train crossing the trail in front of me. It is accelerating, and one flat bed car after another passes by. I get tired of waiting for the train to pass, so I climb a tree to see how long the train is. and find that the train track is circular, and that there are just enough cars on the train to link up the train all the way around the circle, so that there is no first or last car on the train. I then conclude, correctly, that somewhere in that circle of cars there has to be a locomotive. I can't argue that the flat bed car in front of me is being accelerated by the one behind it, and that one by the one behind it, and so on all the way around the track. No flat bed car can accelerate itself, and so no looped or infinite series of flat bed cars can accelerate themselves. It takes a locomotive to start a train moving.

  7. awatkins69:

    I grant that kind of conceivability, but as I'm sure you agree, the CA is mounted in our world, where things happen all the time. Causal bearing seems to get smuggled into infinitist objections, when "how all that gets going" in the first place is the point. Methinks.

  8. I think for the most part the ones you call infinitists do have some sort of sense of causality as you say, but maybe to them it is a general rule as opposed to a law?

    Also, Would be Thomist gave great examples I think. Number 1 there, about the moons, is most clear I think. The causal chain does seem to logically entail that there must be some unlighted light source. So in the same fashion, God is the unmoved mover. However, I am having trouble understanding why we conclude from an unmoved mover that the being must have no potency whatsoever, is immovable, and is pure act. Maybe someone can explain this.

  9. The instant I surfed to the unbelievable The End Of Nations commentary my first thought was Edward Feser's readers totally have to witness this!

  10. awatkins69: the first mover has to be pure act, with no potential for motion or change, because unless it were a completed act, it would not exist to exert causal efficacy. Only completed acts – also known as facts – can be efficacious. The most general fact of existence as such must be eternally complete and thus logically prior to all other things whatsoever in order to be perfectly general, and so to have the capacity to condition everything that happens whatsoever.

    Also, does it really make sense to characterize a bunch of events, none of which have causal relations to any of the others, as a world? Or even as a state of affairs? How does utter uncoordination rise to the level of being? Another way to put it: is not sheer chaos the same as non-being?

  11. I'm sorry kristor, but you lost me when you said "it must be a completed act or it would not exist". Something can act while still having potentiality. I don't see this. Again, I'd really like to understand these things. I'm quite open to Aquinas' Five Ways if they will work.

    Also, I'm starting to reconsider whether our causal series here leads to God. My main objections are three:
    1) Does not happiness move the will, thus making the essence of happiness an unmoved mover? Certainly we do not call this God. Of course, from here we could turn to proof 4, but that means the First Way doesn't work alone.

    2) Don't certain natures act as unmoved movers? For instance, the law of gravity, by its very nature, is an unmoved mover. Another example is how the nature of the atom causes radioactive decay. But the nature was not moved. We do not call natures God, certainly. Of course, from here we can go to the Fifth Way, but in that case, again, the First Way fails as a proof on its own.

    3) I'm still trying to figure out how Newton's laws do not contradict this way. Motion is a change of location, thus being a type of change.

    Again, I'm sympathetic to the Five Ways, but I'm seriously reconsidering the cogency of the arguments. The De Ente argument might still work, but that's a whole different ballgame requiring a lot of metaphysical presupposition.

  12. Awatkins69:

    Sorry to take so long to respond. You write, “something can act while still having potentiality.” Indeed so; but only if it is actual can it have any property at all, including the property of having potentialities. An acorn can give rise to a tree only if it actually exists. And an acorn can actually exist only as a completed act of becoming of an acorn. Until the acorn has finished becoming what it is at time t, it cannot have the potentialities that it will have at time t. Only completed acts – things that already exist – can have causal effect.

    So the First Mover has to be complete prior to all other events. If the person who is the First Mover were to have a career composed of a series of acts, all the acts posterior to his First Act would not be First at all.

  13. """"Another example is how the nature of the atom causes radioactive decay. But the nature was not moved."""

    Yes and other examples from modern physics can be produced. Aristotle and Aquinas may suit dominoes, or billiard balls. Not radioactive isotopes

  14. Radioactive atoms are still composites of form and matter so they too would require a prior cause to be actualized.

    So anonymous above appears to be wrong.

  15. It's worse than that, even. Radioactive decay is a general case of "wavefunction collapse," or decoherence. It is caused, caused by the action of the environment on the atomic nucleus/atomic nucleus + emitted particle. Indeed, the "reduction of the wavepacket" could easily be thought of as "reduction of the wavepacket from a set of potencies to a specific act."

  16. It's worse than that for the one who claims that radioactive decay defies metaphysics: the decay of an isotope is a subtype of the more general phenomenon of "wave function collapse," which we now call "decoherence," and decoherence manifestly is caused, it is caused by the environment!

    Quantum phenomena yield surprisingly easily to interpretation by Thomistic metaphysics. "Reduction of the wavepacket" could be viewed as a shorthand for "reduction of the set of potencies associated with this quantum system to a single, specific act." The parsimony and intuitive plausibility of such an interpretation stands in stark contrast to the extreme absurdity of the "many-worlds hypothesis," and could thus be preferred on grounds of parsimony alone!

    1. Having learned more about quantum mechanics in the years since I posted this, I think I was wrong in almost every specific way, but right in general.

      The trouble is that "decoherence" and "wavefunction collapse" aren't actually synonymous.

      Decoherence is the fact that, as the wavefunction of a very complex system evolves, the interference terms get vanishingly small. Basically, this means that an electron in a two-slit experiment is small and simple enough to bump into itself after passing through both slits, but Schrodinger's Cat is too big and complicated to trip over its own corpse on the way out of the box.

      Wavefunction collapse is something stronger, something that goes beyond the unitary evolution of the wavefunction. It's not just that the cat can't interact with its corpse, it's that the cat and its corpse can't both exist at the same time in the first place. Somewhere along the line, one of the states straight-up disappeared.

      Now, the second problem is my characterization of the "many-worlds" hypothesis as unparsimonious. It's actually the simplest possible interpretation of Quantum Mechanics because it gets by with nothing but unitary evolution of the wavefunction.

      Imagine that decoherence is all that's going on in the Shrodinger's Cat experiment. Then, as light from the cat/corpse non-interfering superposition reaches your eyes, you wind up in a superposition of "seeing a live cat" and "seeing a dead cat." Since "cat+box+you" is an even more complicated system than "cat+box," the interference term is even smaller in the former case, so it "feels like" you only actually ever interact with one of the superposed states. Essentially, we get the same predictions as an objective collapse hypothesis, but don't need anything beyond unitary evolution of the wavefunction. Occam's Razor, if it can be invoked here at all (it can't, see below), manifestly cuts against the objective collapse theory, not for it.

      The guy who came up with this, Hugh Everett, called it the Relative State Formulation. What's really going on is the universe is in a superposition of states, but each state of that superposition consists of a set of states of the things in the universe that are rigidly defined relative to one another. Note the conspicuous absence of "alternate universes." There's not two cats in two different worlds, there's one cat in two mutually exclusive states in one world.

      Later proponents of the system, apparently less skilled at dealing with cognitive dissonance than Everett was, started thinking of decoherence as the "splitting of the universe" into two new ones. Thus was born the "Many Worlds Interpretation." Basically, they rebranded the system to draw attention away from the flagrant, cosmic-scale violations of the Law of Non-Contradiction.

      Needless to say, learning more about QM hasn't made me any more comfortable with the MWI. My reasoning in the old post may have been dead wrong, but my evaluation of the MWI as being "extreme absurdity" seems rather understated, if anything. The "Schrodinger's Universe" interpretation - because that's what it actually is - runs against the most fundamental principle of Reason: the Law of Non-Contradiction. There is nothing that can possibly salvage it.

  17. Is the issue of "simultaneously" really an issue of all effects existing in their essentially ordered series immediately and simultaneously without the possibility of mural causality?

  18. A quick note on wave function collapse for people to remember.
    The wave function collapse is not something that need happen linearly forward through time. It can be caused by future events.
    On what Dave said, the multiverse theory doesn't actually contradict Thomistic metaphysics, it seems to me? The multiverse theory still provides a model by which the wavelength function collapses the way it does in any instance, and explains easily why the patterns are stochastic rather than deterministic. Not to say that multiverse theory is true, simply that its still a possibility.
    It is actually more accurate to say that, with our present knowledge, it seems that decoherence isn't caused by the environment, but by separation from, or a lack of, an environment.
    Parsimony, i.e. Ockham's Razor, is for deciding what possibility to act upon from a list of possibilities, not for actually PROVING something. That being said, multiverse theory is going to need to drum up more evidence in order to be that possibility.
    That being said, I'm not particularly well schooled in this subject.

    On the topic at hand, I'm not impressed by the "movement" analogy. All it really is is an analogy. I could just as easily compare the universe to an inert object as to bullets from a rifle, and an inert object doesn't need to be fired from a gun.
    What I'm more interested in than analogy is to see if any of these premises are tautological or evident. Tautology means one could assert them with 100% confidence, evidence means one could assert them with a level of confidence that is not 0% and not 100%.
    These claims in particular seem to not be demonstrated as evident nor tautological:
    "We notice around us things that come into being and go out of being."
    I'm not sure what things are supposed to be going in and out of being? Remember that much of what we refer to as "a thing" like "a broom" or "a person" are actually groups of things, and thus their dissolution is not true destruction of being. Indeed, it seems that such patterns still exist in their relevant time frames, since there's no reason to suppose that time is an ever-changing frame rather than a set of time frames that exist simultaneously.
    "Whatever comes into being or goes out being does not have to be; its nonbeing is a real possibility."
    This would be evident if the first premise were tautological or evident, but as it stands, non-being is not shown to be a real possibility.
    "Suppose that nothing has to be; that is, that nonbeing is a real possibility for everything."
    Suppose says it all. ;^^
    "From nothing nothing comes."
    An unfalsifiable proposition.
    "The universe could not have begun."
    ...And thus, also unfalsifiable.
    "There must exist something that has to exist, that cannot not exist."
    That would be tautology? Tautologies cannot not be true. Could one not suppose that tautology is the ontological cause of reality? Tautologies are eternal and never changing, there is an infinite set of them, and one can conclude any tautology from any other tautology

    And thus the conclusions are not concluded, for the premises are not shown to be sound, be they sound, or not.
    Of course, there's the issue that the conclusion does not conclude the god of the bible in particular, but that's a separate matter.

  19. By the way, I just noticed the author of this blog post is someone praised by National Review. c: My parents read that blog, and I've had some fun times reading some of the articles there, myself. I've thought about trying to be a writer for NR myself, before, though I'm too apolitical now for that sort of thing. ;^^ I don't think anyone goes to National Review for weird biochemistry talk, which is my specialty. ;>>

  20. Hey doc Feser, when you said "As Thomists sometimes point out, it wouldn’t change things in the least if we granted for the sake of argument that a series of causes ordered per se might loop around back on itself in a circle, or even that it might extend forward and backward infinitely."

    Didn't you mean to say per accidens and not per se?

    PS: I love your blog man, keep up the informative posts!

  21. I think I follow you. The only question I have is, why should we assume the series of events in the cosmos are ordered per se, at all, rather than orderd per accidens? i.e., what is the reason for thinking all the events in the cosmos are ordered like hand-stick-object and *not* ordered like father-father-son? I see no more reason for thinking the series closer resembles the former over the latter, in which case, Aquinas' argument begs the question.

  22. Walter. Feser's video should answer your question.

  23. I'd love if someone less amateurish could step in here, but for my own growth I'll have a shot:

    @Anonymous Sept/16

    "From nothing nothing comes."
    An unfalsifiable proposition.'

    I think this is relatively obvious in regard to philosophical "nothing". It is metaphysically impossible for something to arise from nothing, by very definition.

    "The universe could not have begun."
    ...And thus, also unfalsifiable.

    Where is this stated? Aquinas didn't hold this point of view.

    "Of course, there's the issue that the conclusion does not conclude the god of the bible in particular, but that's a separate matter."

    Indeed. As Feser has stated throughout numerous blog posts, it's not the point -- if the CA is true, monotheism is necessary. Aquinas had other volumes dedicated to demonstrating the Judeo-Christian God. This is completely irrelevant to the argument.