Are models smart or dumb?

May I point you to yet another mind-expanding blog by The OFloinn, looking at the limitations of scientific models and whence they arise. Note particularly how he categorises phenomena into organised simplicity which can be understood in detail (like Newton’s Law of Universal gravity – though that only describes what gravity does, leaving its nature as a magic force); disorganised complexity which can only be understood statistically (like the n-body problem, chaotic systems and so on) and which depend (note that word well) on the individual components being unknown and independent; and organised complexity, where there are multiple interrelated factors, which can be understood neither by simple individual laws nor simple statistical laws. These are profound distinctions.

In this blog TOF doesn’t apply what he has said specifically to evolution, though he does link to another excellent post in which he shows the inapplicability of the idea of randomness to biology, implicitly because it involves organised, not disorganised, complexity (note that has no relation at all to ID’s specified complexity, a somewhat different concept).

Yet I want to make such a link to evolution in a specific way. Whence arises the public controversy of Darwinism, if one leaves aside YEC biblical literalism? It’s not that it uncovers natural causal mechanisms in biology, or even that these point to evolution. It’s that it suggests that one blind, simple mechanism abolishes the need for all other mechanisms, including the wisdom of God. As photographer Laszlo Bencze pithily puts it:

The dullest person can understand the basic story line: “Some mistakes are good. When enough good mistakes accumulate you get a new species. If you let the mistakes run long enough, you get every complicated living thing descending from one simple living thing in the beginning. There is no need for God in this process. In fact there is no need for God at all. So the Bible, which claims that God is important, is wrong.” You can be drunk, addled, or stupid and still understand this. And the real beauty of it is that when you first glimpse this revelation with its “aha!” moment, you feel like an Einstein yourself. You feel superior, far superior, to those religious nuts who still believe in God. Without having paid any dues whatsoever, you breathe the same rarified air as the smartest people who have ever lived.

This simple analysis is perceptive, because it explains why Darwin’s theory was popular long before it was accepted by his scientific peers. It explains why creationism v evolution is presented as a “dumb v smart” dichotomy in the US and elsewhere; not because a few really clever biologists understand it, but because it’s so simple that it makes perfect sense to every dumbkof not committed to biblical literalism. Given relativity and quantum theory, that ought to make one suspicious.

Such intuitive self-evidence even underlies the theological contortions of the modern theistic evolutionists. Since the theory is essentially simple and relies on randomness, there appears neither a need, nor an opportunity, for God to be directly involved. Instead, attention is drawn to the elegance of his setting up a few simple laws (such as universal gravity, forgetting the n-body problem of course) and allowing those to work out in the whole gamut of living forms.

But TOF’s observations bring one back to earth. Darwin’s original theory was neither a simple nor a statistical law, but a model, with the whole of biology being reduced to the two variables of variation and selection. The Neodarwinian synthesis is even more obviously such a model, with its mathematical basis in population genetics relying on simplifying assumptions that were known to be unrealistic even when formulated. They have become increasingly obviously idealised – and in fact, in some cases plain wrong – as biology has progressed.

The implications of this for biology itself have been variously put by scientists, and variously received, depending on factors like ones commitment to population genetics and ones ignorance of philosophy of science. One example is this chapter, from a book too expensive to recommend, by Ward B Watt. The chapters by Clegg  (you’ll need to download and convert to pdf) and Hey speak to the same theme.

Such writers are keen to put their case against the model as a way of increasing our knowledge of evolution, not of denying it. But they do so – like the increasing array of “dissidents” such as Shapiro, Noble, Fodor and so on – and even those mainstream innovators like Gould or Ohno – by embracing the fact that the sheer complexity of biology cannot be reduced to a single simple, intuitive, model.

TOF points out that models of complex systems can still work adequately (ie their inevitable errors are less serious) if one or very few factors outweigh in importance those one simplifies away. That’s how population genetics can make real predictions in certain idealised cases. Evolution is rapidly being shown to be a situation where the multitude of variables, and their interactions, cannot be so dismissed. One could enumerate that, but every day news of non-genetic inheritance, alternative splicing, a whole heirachy of switching for genes and so on render the old model misleading. One may distribute such new factors into the piles marked “variation” or “selection” and say the theory remains the same, but that doesn’t hide the reality for long. All artifacts can be categorised as “containers” or “extensions”, but the simplicity so produced is illusory.

And so biology, freed from over-simplified models, resembles more any of the half-dozen or so human examples of organised complexity TOF gives. Human affairs, too, have been subjected to, and liberated from, such over-simplified models as the Marxist theory of history or the behaviourist theory of psychology. As Lenin and Mao’s  revolutions showed (too late for millions), for all the useful insights they give, such models are grossly inadequate, though it’s easy to get the masses to understand and act on them. But in truth there are vastly more mutually-interactive factors, not least the choices made by individual people, that make simple theories of economics, history or politics less than useful as grand narratives.

An amusing illustration of this is in a link from one of TOF’s commenters, citing a joke in which some physicists offer a cheap and simple way to predict the outcome of the Superbowl. “We have made several simplifying assumptions,” they add: “first, let each player be a frictionless rolling sphere on an infinite plane… “

What happens to Darwinism if one is bold enough to escape from the simple appeal of evolution as statistical mechanics? One discovers that: “Natural selection is real, and has some effect on the changes in biological forms over time.” To that one can then start adding all the new insights from the myriad of other processes discovered, according to ones speciality. Speaking theologically, I’d just add this: just as the existence of natural selection is, when kept within bounds, no threat to neutral theory, neo-lamarckian epigenetics, emergent complexity or even vitalism and structuralism, so it carries no more threat to God’s governance of evolution than Keynsian economics does to his governance of history.

Statistical deism, though, begins to look, like all simplified models elevated into absolutes, so nineteenth century.