Monday, April 23, 2012

On Aquinas and the Cosmological Argument

In debating religion, I've run across a claim that has been made recently that the cosmological argument is the strongest argument out there for the existence of god.  And not just any cosmological argument, the one put forward in Thomas Aquinas' First Way, part of his Quinquae viae.  This may come as a surprise to those of you who thought that the cosmological argument had long since been soundly refuted.  But, as I've been told, none of the refutations, not a single one, have ever dealt with the real argument that Aquinas made.  Apparently, the way it's been interpreted for the past eight centuries is wrong.  Ed Feser makes this claim, but I'm afraid I don't currently have access to his book on Aquinas.  I intend to get to it in time, because I'm told it's very important.  In the meantime, this blog has been presented to me as an example of how the argument is supposed to work.  I've thought about it a good bit, and have decided to devote a little time to refuting it.

To sum up, here's the quick and simple version of the argument:

  1. Some things are being actualized
  2. Whatever is being actualized is being actualized by something else
  3. An essentially ordered chain cannot be infinitely long
  4. Therefore, there is something of pure actuality

Let's go ahead and take a look at the argument in detail.  In the first two sections of the post, we simply get a history lesson, describing the problem of change presented by Parmenides, and the division of things that exist into actual and potential by Aristotle.  It's simple, clear, and to the point, and I have no quibbles with how it begins.  I may not agree with the way the Greeks set it out, but it does its job at framing the space in which the argument is being made.

Regarding point III, however, we come to my first objection.  The distinction is made between an accidentally ordered series and an essentially ordered series.  An accidentally ordered series is one in which the causes of the events in the series need not currently exist; in an essentially ordered series, the causes must exist in the present with the effects.  But is this a meaningful distinction?  The fact that we live in a universe where there is time makes the distinction between an accidentally ordered series and an essentially ordered series questionable. Let's start by defining the present, something that we can do today in terms far more accurate than Aquinas could in the 13th century.  What we call "the present" can be divided into tinier and tinier slivers of seconds, down to (theoretically) Planck time. The Planck time is the time required for light, in a vacuum, to travel one Planck Length, which is really short: 1.616199(97)x 10−35 meters.  This makes the Planck time roughly 10−43 seconds.  And, critically, for events taking less than Planck time, no change can be detected; that's part of the definition of Planck time. For events taking more than Planck time, the information that the event has occurred can propagate no faster than the speed of light, and thus cannot actually cause any change until at least some time has passed. All causes, by our current understanding of the laws of physics, must be in the past10−43 seconds in the past may not seem like a lot, but it is the past. In "the present", i.e. any given instant in which the Planck time has not elapsed, nothing is changing. Change is always measured over time.  But the concept of an essentially ordered series depends on things changing in the present; if I push on an object, so the idea goes, the force that I am exerting on the object in the present is causing the change that moves the object.  In reality, that's not how the world works; the force I exerted an instant ago is carried across the distance between my hand and the object at no faster than the finite speed of light and affects the object, and the force I'm exerting at this instant hasn't gotten there yet, because even if I'm only the Planck length away, the information that I'm exerting force cannot have yet traveled the distance.  It appears that all series must be accidentally ordered series, and since this is admitted to not be proof of a god, the argument falls flat before we even finish the definitions.  That's a pretty good refutation, I think.  But I can do better than that.

Let's look at the three premises in the simplified argument above.

  • Premise 1: Some things are changing. This is fine, subject to the point I elucidated above.
  • Premise 2: Whatever is changing (whatever is a mix of potential & actual) is being changed by something else. This is the key point of contention, as it is in many modern refutations of the cosmological argument. It hinges, as many things do, on the fact that we've found out a lot since Aquinas' day.  This premise isn't necessarily true, but that doesn't mean it isn't true in practice; it could have turned out to be the case, even if not logically necessary. However, the observation of uncaused phenomena in modern science does mean that it isn't true. Radioactive decay is a great example, and the most familiar that I can name.  An unstable isotope of an element doesn't look unstable; it appears perfectly fine, acting just like any other atom, right up until the very instant that it decays.  There are no precursor events, no warning signs that decay is about to occur; on the level of individual atoms, whether the atom decays at any given moment is a matter of probability, and is one of the processes that appear to be truly stochastic.  We can say that, over a given period of time, about half of the atoms of a chunk of radioactive material will have decayed, but we can't say a thing about any given atom.  Nothing actually causes the decay to happen; there are necessary conditions, but no causal event.  Aquinas couldn't have known this in the 13th century, so he can't be blamed. But we should probably move on now that the universe has shown this premise to be false.  I could stop here, too, but let's keep going
  • Premise 3: This chain cannot be infinitely long, because it is an essentially ordered series. Again, the question of whether the concept of an essentially ordered series makes sense comes into play. But even if we grant it, the unstated assumption is that there must be only one terminus to all chains. This point is addressed further on, though, and I'll come back to it. But it also ignores the possibility of cycles. If we have an essentially ordered series, where A is causing B, and B is causing C, and so on, what stops Z from causing A? If we have, say, a wheel of atoms, and one is moving (from an external push in the past, outside our EOS for this wheel), then it pushes the one ahead of it, which pushes the one ahead of it, and so on, but the one behind the first is then pushing the first. The chain may not have an infinite number of members, but it isn't clear that this means it needs to terminate in a pure actuality.

So, we've got serious problems in two of the three premises, with the second being the most problematic.  The argument appears to be just as unsound as we'd expected.  But I can still do better than that.  The post then goes on to the attributes of "pure actuality". Even if we grant that we can get there from the argument, the attributes noted are either questionable or applicable to something that we would not call god. Of note is the property of unity.  As the post says, "The only way to tell the difference between two things is if one thing has an attribute that another lacks." But that doesn't mean there couldn't be an infinite number of indistinguishable ends to the chains. And it doesn't mean that a pure actuality would necessarily have all attributes; there could be things that it might not be possible for it to be, in which case it couldn't potentially be those things, and would remain pure actuality.  It also includes an odd Thomistic concept of "good"; goodness in the Thomist view is synonymous with existence.  This is a concept that really doesn't seem to make sense; the pen sitting on my desk is good, but if I need a pen and don't have it, I'm suffering from evil? This seems to be a matter of equivocation more than reason.

But I have one more point to make, one that I think is particularly devastating.  A common claim for the cosmological argument is that it provides a "first step", a way to get not to the full concept of god (although the traits noted here for pure actuality are pretty close), but to an acceptance that a god of some sort exists.  But let's look at some of these traits, and see whether they describe a god or...something else:

  • Incorporeal: we can imagine a lot of things that are not matter or energy.  Ideas, for example, or mathematical constructs, or the laws of physics...
  • Simple: Well, many things are simple, even more simple than matter and energy.  Like the equation that describes matter and energy (E = mc2) or the Feynman path integral describing the likelihood that a particle will follow a given path through space...
  • Unity: Basically, something that is the same everywhere, where you don't have multiples, but rather a single unified description...
  • Eternal, Immutable: The same not only everywhere, but everywhen, constant at all points in time from the beginning of the universe (or even before) onward...
  • Impassable: Cannot be made to change, beyond the ability of anything to alter, like some sort of fixed natural law...
  • Omnipotent: The source of all change in the universe, and the source of all change that could possibly occur, the ultimate explanation for everything that occurs...
I'll be honest with you, I don't see god in that.  "Conscious", "personal", those aren't implied.  What do I see coming from these traits that pure actuality must have?  A limited set of necessary properties of the universe (as Lawrence Krauss and some other physicists argue for) would be incorporeal, simple, eternal, unity, immutable, impassable, and (at least by the given definition of "the source of all change in the universe, and the source of any change that could ever logically occur") omnipotent. Thanks to the Pauli exclusion principle, noting that no two particles can ever be in exactly the quantum state simultaneously, and thus any change in, say, the energy state of an electron caused by rubbing your hands together might then affect every other electron in the universe (a really weird conclusion supported by some physicists like Brian Cox, but disputed by others), it could even be said that this basic set of principles is omniscient, probably in a more direct way than the roundabout "ground of all being, doesn't have knowledge but is knowledge" sense described. Since the definition of "good" being used is the weird Thomistic way of equating goodness with existence, then the Thomist would even have to conclude that the laws of physics, since they quite evidently exist, are indeed good. I wouldn't say so, but then, I think that's a wonky definition of good.

Congratulations, Thomists.  If you accept the cosmological argument, as presented in Aquinas' First Way, as a valid argument, it would be perfectly reasonable for you to conclude that the "first cause" is not god, but a limited, simple set of logically necessary laws of physics.  If you happen to be interested in what those logically necessary laws of physics might be, I recommend A Universe From Nothing by Lawrence Krauss or The Grand Design by Stephen Hawking and Leonard Mlodinow.  Both argue quite nicely that the universe did not need a god to get it started, or to keep it running currently, but rather that the laws of physics alone suffice.

Edit: Updated my brief diversion on the Pauli exclusion principle.  The science isn't quite as conclusive as I'd originally presented it.