Epicurus, Moran, and chance

Larry Moran, at Sandwalk, has argued that evolution is indeed a matter of chance. It is, he thinks, something that atheism requires. This is an interesting issue, one that has deep roots, both in the role of chance explanations in science and prescience, and the theistic reaction to it.

Larry says

It's perfectly okay to say, as a first approximation, that lots of evolution is random or accidental. This is a far closer approximation to the truth than saying it's the all the result of design by natural selection.

What role does randomness or accident play here? Is it an explanation to say that things occur by accident, when they are what is to be explained? There are a number of meanings to "chance" in this case. One of them is that without a perturbing cause, ensembles of events will tend to form a Poisson Gaussian distribution - the "bell curve" of beginner's statistics. If we don't have something making things shift in one way or another, then many events of the same kind will cluster most densely around a value or location on the graph. Populations do this.

Another is that there are factors at work in any explanation that are not predicted by the theory used to explain. It is not, for example, predicted by the theory (or theories - I think there are at least 7 theories involved in evolutionary biology, but that's another topic for another time) of evolution that there would be a bolide impact 65 million years ago. Once that happened, though, the theory predicts a general class of outcomes from a depauperated ecosystem.

But which lineages survived the Cretaceous bolide is also a kind of chance - called "contingency". Some lineages may have survived because they happened to have biological features that permitted them to survive the impact winter that must have followed. Some, however, will have survived because they were in the right place at the right time.

Larry wants, I think from many emails exchanged and a conversation over drinks, to see that chance, accident, is a major "force" or factor in evolution. I think he means both unpredicted factors, and contingency. Deterministic factors like selection play a secondary role in his thinking.

I am reminded here of the Original Bad Boy for theists, right back to the time of Aristotle: Epicurus. Epicurus believed, more correctly as it turned out than Aristotle's view, that what happens in biology and physics was caused by the internal dispositions of the atoms that made up living and physical things, in contrast to Aristotle's finalism and formalism. But he had a problem explaining why things weren't entirely the same everywhere and when. To "explain" this he supposed that originally (from eternity) atoms fell through the void (something else Aristotle rejected) in parallel paths, but that an accidental "swerve" brought them into collision from time to time, causing differentiation.

From this, theists, particularly Christian and Jewish theologians, attacked Epicurus as being atheist (which he may have been, but more likely thought the gods were too distant to effect change in the world), and holding that "everything was an accident". In short, he removed meaning from the world by giving chance the centre stage. Actually, he didn't, but chance did have a role in his scheme. Aristotle and the Stoics rejected this immediately, and Epicurean thinking diminished for around 2000 years.

The question here is about determinism and chance as competing explanatory strategies. If we accept a central role for chance in explanations, it is thought, then causality is at risk. Those who try for a strong selectionist or adaptationist account in evolution tend to be determinists. Those who, like Larry's minor deity Gould, believe chance as contingency has a role tend to deprecate deterministic explanations.

In one sense, statistical explanation is a mixture of chance and necessity, as Jacques Monod called it famously. It is a necessary outcome that if no confounding cause intervenes, things will fall around a mode in a Poisson Gaussian distribution. I suspect that this astounding fact, which was called the Law of Error during the 19th century, is part of what drove Darwin's thinking, although he was equally influenced by simple observation of variation in living things. But this is not the kind of chance in question - it is a causal fact of the world as we see it that things distribute that way.

Contingency, however, is not chance proper. A determinist could say quite happily that a theory doesn't need to account for the boundary and initial conditions of the phenomena or process it seeks to explain, so long as some deterministic account can be given of them. In other words, we don't need to think that bolides are created ex nihilo or that they swerve in an Epicurean sense to hit the earth in contradiction to orbital mechanics. Likewise, the changes in climate caused by the Milankovic cycle or the precession of the axis of the earth's rotation, while not part of the theory of evolution, have predictable effects once they are included in the calculation. The "accident" here is just the fact of being extra-theoretical.

So when Larry says that atheists (by which he means the non-milksop nonbelievers in God who accept evolution as rational beings, not the milksop variety who refuse to make pronouncements on God... but we digress into already trodden ground) accept chance as a factor in evolution, I think there is a bit of trading on words rather than states of affairs. Yes, molecular changes can be "random" in the sense that they are external to the theory of evolution. They are not external to the theory of chemistry. In the domain of subatomic physics, the randomness of radioactive decay or gamma radiation has to do more with statistical properties than actual chance. The rates at which such events are caused or which have an effect on biological molecules are bounded. They aren't wholly random.

All that said, is this a substantive disagreement between Larry and me? Not really, not in practice. We might think all of this, but still need to treat these events as random for the following reasons: no event occurs in anticipation of the future as it will be (not even when we plan for the future), but only in response to past events; and we do not ever act on, nor will we ever have, perfect information. All our real science has to factor in randomness because no matter how the universe may be - deterministic or stochastic - no theory, and no theoretician, can ever access all the information required. Determinism may be all very well as a philosophical dispute; it cannot be an operational part of science, not at the limit. Of course we must seek causal explanations, and they presuppose causality. There will be a point at which we have to treat noise in the data as an irreducible fact of our best theories, no matter how good those theories may get.

Theists will always kick up when chance gets mentioned, because they are the Überdeterminists. No small part of the created world can include real chance, because that is just contrary to doctrine. Atheists can allow there may be real chance, but I do not think they must. It is possible to be an atheist, milksoppish or not, and be a determinist at the post-quantum level. I think that determinism is the only metaphysics that makes sense of the world, but it is not a good standard for assessing theories or method.

More like this

A common attack upon evolutionary biology, from ranking clerics in the Catholic church to the meanest creationist blogger, is that it implies that life arose and came to result in us by accident. We are asked to believe, they say, that three billion years led to us as a series of accidents. No…
John Wilkins has replied to Larry Moran on the role of "chance" in evolution (incidentally, Moran replies to Wilkins on the same topic, but a different post by Wilkins). Here's what Larry wrote: Nobody denies the power of natural selection and nobody claims that natural selection is random or…
That paper that proposed that most cancers were due to bad luck, that is, that they were a consequence of biological factors that could not be controlled, has been surprisingly controversial. I thought it was a fairly unsurprising paper that confirmed what we already suspected, but wow, the furious…
One of the points made by Rabbi Slifkin in the article I cited recently is that if you insist on using God as an explanans in the aspects of the world we do not yet understand, that is going to mean a decreasing role for God as we learn more. This is an old point. Wesley Elsberry and I made this…

Your points are well-taken.

But some theists are not determinists, and will say that God has created the world with an allowance for chance. I believe that Thomas Aquinas is one of those accepting chance.

Seems like silly argument. Natural selection and chance (whatever that may mean) are inextricably intertwined. Without chance, there is no raw material for selection to act on, and without selection, the emergence of anything interesting (with the appearance of design) is prohibitively improbable. One of the neat things about evolution is how neatly it modularizes chance and determinism, so that even a child can see how they work together.

There is also a third requirement critical to evolution: replication, which might be considered deterministic or stochastic, depending on your POV (I see it as deterministic).

There is a hypothetical solution which is likely to be offensive to both atheists and creationists. I call it Divine Selection.

Imagine that the mechanisms of replication, repression, derepression, and so forth are as we think they are. Imagine that differential reproduction by Darwinian fitness is as we think it is. But imagine that mutation is controlled by God.

God, in this theory, aims cosmic rays and molecules of mutagens, by some means of His that admittedly involved transfinite computational capability, so as to cause exactly the point mutations that he knows will have the consequences that He plans.

By means trickier to explain, He manipulates the cellular machinery to ensure that the inversions, crossovers, and other chromosomal mutations exactly the point mutations that he knows will have the consequences that He plans.

The genetic algorithm proceeds exactrly as we surmise, but with the assumption of randomness replaced by one which is statistically random but actually directed. Randomness is too important to be left to chance.

Now, we have microevolution and macroeveolution happening exactly as the neodarwinian paradigm demands. Statistically, the population genetics equations all look right to us.

Yet God has made it be, indirectly, that His teleological agenda is fulfilled. These unicellular organisms are tweaked towards multicellularity. These dinosaurs are nudged towards being birds. Those hominids find themselves with brain development havinbg Chomskian predeliction to create and converge on the grammars of spoken language.

The strong version of this is that God has made it that this embryo shall be Bach, that one Newton, this one Darwin, that one Hitler. God works in mysterious ways. These ways do not violate physical laws, nor laws of mathematical biology. But they have intentional miraculous results.

I'm not saying that I believe this. As of now, probably nobody believes it. BUT this does meet some of the axioms of Creationism, and some of the axioms of atheism, while subverting both.

I expect withering attacks from both sides of the fence. Or, if nobody takes the bait, at least I know that I've thrown a hypothesis into the meme pool where one has not been before.

I don't think that this blogged meme came about at random. I think that I created it. Every writer thinks that he or she is doing Divine Selection of the enemble of poems, stories, science articles, equations, or whatever else is evolving in the plane of ideas. In the terminology of J. R. R. Tolkien, the writer of a Fantasy is a subcreator.

Nitpick: s/Poisson/Normal/ (or s/Poisson/Gaussian/. Or insist that N -> infinity)

More importantly (I hope), am I reading this correctly that you are assuming that causation is deterministic? You seem to be arguing (or assuming for the sake of the argument) that all randomness is epistemic.

If one allows chance as causative, then I think it becomes more difficult to separate out chance and deterministic factors: the mathematical techniques can be tricky, even if they are elegant (martingales and the like), partly because they depend on the scale you're measuring on. At the simplest, everything happens by chance, but some things are more likely than others.


By Bob O'Hara (not verified) on 26 Dec 2006 #permalink

Bob O'Hara:

Yes, for the sake of Devil's Advocate argument I am hypothesizing that causation is deterministic and that at least some (not necessarily all) randomness is epistemic.

The statistics of population genetics are neither Gaussian nor Poisson, because of nonlinear effects including error-detecting-error-correcting in DNA/RNA, multi-allele correlations, etcetera. See John Holland's lovely math in the 1976 classic "Adaptation in Natural and Artificial Systems" which I beta-tested in manuscript form while in Grad school, while I was the first person in the world (sorry, patent-holding Koza) to use the Genetic Algorithm to evolve nontrivial working computer programs (the allleles were characters in APL one-liners). In fact, what I evolved solved a previously unsolved scientific problem in the literature.

This Divine Selection (which one might also call "Nanocreationism") hypothesis is not, strictly speaking, testable, and is constructed to be untestable, as we cannot "know the mind of God" in the Spinoza/Einstein/Hawking sense. But it is a philosophical position devised to show that that are theories in between atheism and creationism.

In what the late Fritz Zwicky calls the "ideocosm" -- the space of all possible ideas -- there is some question of topology, which affects the issues of density. That is, given any two hypotheses, is there always room to squeeze on in between the two?

Diplomatic methodology seems to depend on this. As does jesuitical hairsplitting.

Of course God, to the extent that He is a Mathematician, knows all about Martingales, and han solve NP-complete questions on digraphs, and, for that matter, compute Turing undecidable functions. He is His own Oracle.

As modern Probability theory teaches (I took advanced probability with Lebesgue measure and the like) from Dr. Gary Lorden now executive Officer of Math at Caltech, and Math advisor to the hit CBS series NUMB3RS.

Viewed as a police procedural, is God guilty of determinism hiden under an alibi of chance? In the Mystery genre, show me a "coincidence" and I'll show you somebody who has something to gain by it.

See also Woody Allen's story of hiring a private Eye to determine whether or not God is Dead.

See also the various strange theories of Bayesian probability of miracles, Descarte's Gamble, and the like.

Einstein: "I do not believe that God plays dice with the universe."

Jonathan Vos Post: "I suggest that God plays with many, many loaded icoshedral dice, and we are all characters in a vast Role Playing Game."

Hard to disprove, but I hope fun to contemplate.

Jonathan - I wa actually asking John! The Poisson/normal substitution relates to the Law of Error part.


Being as I am innumerate, I always confuse Poisson and Gaussian distributions. I'll fix it in the post, thanks, Bob.

Jonathon, the problem with your divine selectionism is that it overdetermines the outcomes. This is the view of Gray that Darwin criticised. Either every "random" event is determined by God, or none are. If every event is, then saying God chose the outcome is equivalent to saying only that chance equals God's action. If the universe has a lawful randomness, then asking us to accept that in those providentialist cases God intervenes to make random outcomes turn out as he wants is to say that it would be sufficient to say they are just chance but that they aren't when it suits us theologically, which is at best an overdetermination (100% chance is sufficient, so God adds 110%?) and at worst a God of the gaps.

:Theists will always kick up when chance gets mentioned, because they are the Uberdeterminists."

News to this theist.

Either every "random" event is determined by God, or none are.

Nope, I'm not following this ... unless you are restricting that the the case of providentialism. Even then I don't see why that is any greater problem than the existence of evil (i.e. God makes choices that seem evil to us).

Otherwise, you seem to be just telling god what to do. But you're too polite for that.

Mike, suppose that there are events not under the control or will of God. Suppose that they are fundamentally and irreducibly chaotic. Would this be consistent with a monotheistic worldview? I believe this got a fair bit of discussion in the early church fathers...

John, I'm just responding to the standard view of omnipotence. If you want to allow that God only has partial control, that's very fine with me, but now we are not talking about the same deity as everyone else. They are the ones telling God what to do. I'm just taking them at their word.

The Flying Spaghetti Monster, on the other hand, has control only when He wants to...

John Wilkins:

Thank you for taking the time to offer a counter-argument. However, I share John Pieret's puzzlement on your claim: "Either every 'random' event is determined by God, or none are."

This seems to be a misapplication of the Law of the Excluded Middle. Why deny the hypothetical possibility that SOME "random" events are determined by God, but many (even most) are not?

Again taking a hypothetical example (which I consider absurdly unlikely), suppose that Earth is the only planet in the universe with intelligent life.

This is close to some medieval notions that "all the world's a stage" but there's no drama of consequence elsehere in the physical universe.

God then would have no incentive to meddle with nanomiracles through most of the cosmos, reserving the subtle tampering for Earth.

Christian science fiction writers such as C.S. Lewis and James Blish consider whether Christ died only for humans, or for all beings; and whether Christ in extraterrestrial form died on many planets.

Theological Science Fiction

Buddhism and some other religions explicitly accept that there is intelligent life elsewhere in the cosmos. It is said that when Buddha preached, from under the Bo tree, there were beings from other worlds there listening. I argued about this once with Alan Ginsberg. I asked "how did they get there?"

"Probably not," he said, "with rockets."

The question again is, if miracles are possible, and do occur, how often, and for what end?

God could, hypothetically, be nanomanaging the evolution of important speciation events, and the formation of saints and certain massively damaging people such as Hitler, Mao, Stalin, Pol Pot. Otherwise, he's hands-off.

The Manichaean heresy would suggest that God can be divinely nanocreating for Good, while Satan is nanmanaging for Evil. C. S. Lewis, in "The Screwtape Letters" bases a fine (bestselling) novel on the angelic and demonic bureaucracies devoted to getting one human to tip slightly towards Good, or slightly towards Evil.

Again, I don't believe this at all. But writers whom I enjoy reading have explored this in fiction, and philosophers whom I enjoy reading have debated this in nonfiction.

I agree with you that the Philosophy of Biology is particularly important. Your blog has intelligent debate. Hence my care in showing that there are hypotheses in between total determinism and total chance, between Creationism and scientific materialism.

A casino doesn't need to cheat every customer on every bet. Even if they can manipulate the roulette wheel, they might not do it for bets under a hundred thousand dollars. They don't change dealers at the blackjack table until someone on a "lucky streak" has quite a big pile of chips.

Again, thank you for entering the sub-arena I started, within your wider imaginative world.

I won't claim to know everyone else's god intimately but isn't there a disconnect between responsibility and agency here? An omniscient and omnipotent god is responsible for everything that happens (how could he/she/it not be?) but need not be the agent of it happening. A fine line, I know, but we are in the realm of pin-head-dancing here.

That would, if memory serves, comfortably fit within Catholic orthodoxy.

P.S. Yes, I know there is the problem of free will but, as far as I can remember, Catholic theology treats it as a commodity that God handed over to humans that carves out an exception to the big guy's responsibility.

The whole of theology can't be consistent ... anymore than any other philosophy.

There are an uncountable number of possible universes that are apparently self-consistent (e.g., as implied by quantum uncertainty and entanglement during the singularity of the Big Bang). Just as a novelist may choose one storyline from the infinitude of his imagination, so a Universal Creator could choose one specific cosmos. The history would be self-consistent, the ends would be known, but only to Him.

Well, determisnism, or preddestination might suggest a "plan", redolent of religious ideas. They might suggest that although God didnt create man a few thousand years ago, he/she/it DID implement the laws of nature, laws that were conducive to the eventual evolution of consciousness, regardless of which species it evolved in. Some people reject anything that sounds even remotely religious in tone, even if it means promoting red herrings. e.g. Fred Hoyle , stuck with Steady State Theory, and ridiculed Big Bang cosmology, probably because it sounded too religious.

Jonathan Vos Post -
I think your setup would leave some detectable traces. For example, if God is controlling mutations, it means that some genes normally thought of as "molecular clock" genes, like cytochrome-c, have an oddly regular rate of mutation (why would God do that?) Additionally, there's no reason to suppose that an organism's mutations (across their entire genome) relative to closely related organisms have anything at all to do with the actual time since divergence. Plus, it means that mutations in viruses and bacteria which confer resistence to antibiotics and AIDS anti-virals are actually controlled by God.

Here's one to think about though: instead of controlling mutations, what if God changes the actual survival rates of individuals so that He alters the path of natural selection. Because we don't know all the factors (and their quantitative weight) that influence organisms in the wild, we aren't really in a position to say, for example, that God didn't specifically favor the smartest human-chimp ancestors for survival, or influence the dumbest ones to be killed unusually frequently - thus affecting the trajectory of "natural selection" towards a humanlike species. That situation would be very difficult to detect because we could easily assume that the smartest organisms just happened to be favored by natural selection, when, in fact, there was artificial selection going on behind the scenes. (I don't believe this actually happened, just thinking about the possibilities.)