Roger White has a very interesting paper on origins of life research in Nous from a couple years ago, which I only just got around to reading. I think he makes some very good points about the general reasons why apparent coincidences often shouldn’t be taken as evidence against a chance hypothesis. He makes these arguments in order to suggest that origins of life research rests on a mistake – in particular, he suggests that we have no positive reason to believe that the initial appearance of life in the universe was anything other than chance. However, I think his argument for this claim fails.
His argument is basically as follows. Some state of affairs S disconfirms the hypothesis that the system was evolving due to chance only if the probability of S on the chance hypothesis is less than the probability of S given the negation of the chance hypothesis. (This is just a theorem of probability, together with the claim that disconfirmation is a decrease in probability.) Further, S can be less probability on the chance hypothesis than the negation of the chance hypothesis only if S is more probably on at least one of the specific alternatives to chance than it is on chance.
Then he considers what the specific alternatives to chance are that are available, and divides them into “intentional bias” and “non-intentional bias” hypotheses. He rejects intentional bias as a hypothesis that would increase the probability of S (in this case, the origin of life), because of what he calls the “preference problem”. Clearly, if some agent were able to bias things in the relevant way, and preferred that the universe contained life, then there would be a higher probability of life emerging. But White points out that we have no particular insight into what the preferences of a being capable of biasing things the right way would be, and thus can’t assume that the existence of such a being would give the right kind of bias.
He then goes to make a similar argument that we have no reason to think that there was any non-intentional bias in favor of the existence of life.
Now let’s suppose that the process by which these complex molecules arose was not just a matter of chance, but rather was (non-intentionally) biased towards certain molecular configurations. Are self-replicating, life-producing molecules more likely to appear on this assumption? I am unable to see any reason to think so. We can think up any number of ways that the process could be biased. We can speculate about a range of possible laws and physical conditions such that simple atoms and molecules tend to cluster in certain ways rather than others. Some of these may favor life’s emergence; others will disfavor it. As in the cosmological case, what makes certain molecular configurations stand out from the multitude of possibilities seems to be that they are capable of developing into something which strikes us as rather marvelous, namely a world of living creatures. But there is no conceivable reason that blind forces of nature or physical attributes should be biased toward the marvelous.
However, I think this evaluation isn’t quite right. It’s not just the fact that life is marvelous that makes it stand out – life actually has a bunch of features that I think are interesting from the point of view of fundamental physics, and not just from the parochial point of view of life itself. In particular, much of what is interesting about life is the way that living things make use of every source of energy available to them (on Earth alone we see life making use of solar energy; the chemical energy of sulfur and hydrogen in deep-sea vents, and iron and other chemicals in all sorts of extreme environments (see extremophiles), and oxygen in the atmosphere; and of course the chemical energy stored in other life forms, whether already dead or still alive). Ecosystems develop in a way so that some organism uses any new energy source almost as soon as it becomes available (either having been created as a byproduct of the operation of some other organism, the way that animals emerged to use the oxygen created by plants, and rats and flies move in to use the piles of waste created by humans). Historically, life hasn’t been especially effective at extracting energy from the wind, tides, and radioactivity, though the first two have been widely used for purposes of transportation and reproduction, and all three are being used now by humans.
It’s clear that many features of the flow of energy are of fundamental physical importance, given the status of thermodynamics in modern explanations of the universe. The fact that life plays an important role in apparently resisting the second law of thermodynamics, but only by greatly exploiting every source of energy available to it, seems to me to suggest something natural and fundamental. At least, it seems to be more than just the parochial bias of some living thing finding life “wonderful”.
And I think this idea helps make sense of some of the examples White considers.
For instance, on p. 470, he uses Fred Hoyle’s metaphor that “the random assembly of the very simplest living system would be like a tornado blowing through a junkyard and assembling a Boeing 747”. White says, “obviously part of what makes something complex in this sense is that it has a heterogeneous structure, being made up of very many parts of various shapes and sizes. But any pile of 747 parts meets this condition.” He then considers the suggestion that the plane meets a very specific functional specification, but points out that any other collection meets another very specific functional specification, which we can call the specification of what it means to be a “schmane”. “Any considerations which make planes stand out as special as compared to schmanes, are intentionally related—whatever intuitions we have about the case have to do with what we think an agent is likely to do.”
But I don’t think this is correct – one important thing that distinguishes planes from schmanes is that they distribute energy in complex, highly efficient, and long-lasting ways. A schmane where all the pieces of the plane are arranged as dominoes would demonstrate a relatively long-lasting distribution of energy if triggered in the way to knock them down. Similarly, a schmane where certain interactions trigger sparks that detonate the fuel in the tank give highly efficient uses of energy. But very few arrangements utilize the very large amounts of chemical energy in the fuel in ways that last longer than a quick burn. An actual plane is one of the very few arrangements that does very interesting things with the distribution of energy.
Compare the plane to the tornado itself that is involved in the metaphor. There are all sorts of distributions of heat and moisture in the air, but very few of them are as long-lasting and highly efficient in use of energy as a tornado. And in fact, phenomena like this do arise all the time even in purely non-living nature. In reading something about the discovery of rivers of liquid methane on the solid H2O surface of some moon, I realized that the very existence of rivers seems to be something fundamental, but that I wouldn’t have initially predicted just given the thought that liquid can evaporate and recondense. But once there’s a solid landscape, it turns out that some regions of it will tend to have large catchment areas and will thus end up as rivers. Similarly, we get things like weather patterns, plate tectonics, and geysers, not to mention the very small and very large phenomena of chemical atoms and stars. (Think of how many more ways there are to arrange lots of hydrogen and oxygen molecules than as liquid water, and how many arrangements of hydrogen atoms fail to produce nuclear fusion as in stars.)
White also considers a final example, of someone throwing Legos at random off the top of a skyscraper, and the pieces happening to bounce around and assemble themselves into a working model of a cell nucleus and ribosome. Everyone has the immediate intuition on imagining seeing such a phenomenon that they would think that there had to be some kind of intentional bias. This leads us to discard the chance hypothesis, and assume there must be some non-chance explanation. The idea that science requires a search for non-intentional explanations over intentional ones at all costs therefore leads us to assume there must be some non-intentional bias here. But as White (rightly) points out, this is just a non sequitur – if the example motivates us to believe in an intentional bias, this gives us no reason to believe in a non-intentional bias. So if we reject the intentional bias hypothesis, then we have no reason not to stick with the chance hypothesis.
But even this example I think is somewhat misleading. The fact that we see all sorts of phenomena in the universe that are amazing in ways similar to life in their highly structured dispersal of energy (like the tornadoes, rivers, and geysers mentioned above) suggests a slightly different thought experiment. Imagine not just that one person had thrown Legos off the top of a building and they spontaneously assembled into a model of a cell nucleus and ribosome, but that many people had thrown Legos off the tops of buildings, and some of them had formed working models of tornadoes, geysers, rivers, and other phenomena with similar patterns of energy dispersal. Even though we are considering the emergence of the first living organism, that single entity is hardly unique in the universe in many of the interesting features of its behavior. Given the existence of all these things, we might be led to postulate something like the ideas of Ilya Prigogine, or perhaps something a bit less far-fetched, as an explanation for the existence of all these interesting structures, rather than just relying on chance to have produced each separately and spontaneously.
Even if I’m totally wrong in thinking of these facts about the use and distribution of energy as reasons to think of life as special, there are other proposals for why the universe might be expected to be biased towards life.
And in fact, at least some cosmologists respond to the fine-tuning phenomenon (which White mentions on pp. 465-7) in this way. White says about this phenomenon:
While there is at least room to argue that a rational agent is likely to influence the physical parameters in order to allow for the evolution of life, to suppose that impersonal physical laws are likely to constrain the constants in this way can only be based on a confused anthropomorphism. What makes the particular narrow range in which the actual values of the physical constants fall stand out among the full range of apparent possibilities, is that just that they permit the existence of something that strikes us as special or interesting or valuable, namely life.
However, I believe Lee Smolin has suggested an appropriate impersonal mechanism that would lead to this sort of selection. He suggests that black holes are a way for universes to reproduce and produce off-shoots that have very similar (but slightly different) fundamental constants to them. As a result, natural selection will tend to produce universes that give rise to lots of black holes, which (as it happens) tends to involve settings of the fundamental constants that are very much like the ones needed for life. It’s not purely anthropomorphic reasoning that sets this collection of constants apart from the others – without these tunings, we don’t get atoms, or everything collapses into a single neutron star, or fusion fails to produce enough energy to last long. It’s not just that life ceases to exist – so does most of what is objectively interesting about the natural world.
Thus, although I think Roger White’s individual arguments about what it should take to disconfirm the chance hypothesis are right, I think he’s wrong about the claim that we should have no reason to favor a non-intentional bias hypothesis over chance. There are several features of life that seem to be of objective interest, and not just anthropomorphic interest, that could be used in exploring the fundamental physical laws to see how they might be biased in favor of the existence of life.