Friday, January 18, 2013

Is the neo-Darwinian synthesis intuitive?

In a sober [1], concise, careful and clear review of Thomas Nagel's Mind and Cosmos, H. Allen Orr quotes a lengthy passage from the book which serves to ground Nagel's subsequent arguments against neo-Darwinism in favor of a "teleological" model: 

I would like to defend the untutored reaction of incredulity to the reductionist neo-Darwinian account of the origin and evolution of life. It is prima facie highly implausible that life as we know it is the result of a sequence of physical accidents together with the mechanism of natural selection. We are expected to abandon this na├»ve response, not in favor of a fully worked out physical/chemical explanation but in favor of an alternative that is really a schema for explanation, supported by some examples. What is lacking, to my knowledge, is a credible argument that the story has a nonnegligible probability of being true. There are two questions. First, given what is known about the chemical basis of biology and genetics, what is the likelihood that self-reproducing life forms should have come into existence spontaneously on the early earth, solely through the operation of the laws of physics and chemistry? The second question is about the sources of variation in the evolutionary process that was set in motion once life began: In the available geological time since the first life forms appeared on earth, what is the likelihood that, as a result of physical accident, a sequence of viable genetic mutations should have occurred that was sufficient to permit natural selection to produce the organisms that actually exist?  [my emphasis]. 
I think there is something to this.  Not because I think Nagel's right; Orr's review disposes off his objections quite convincingly.  But there is something non-intuitive about the neo-Darwinist synthesis.  A simple mixture of random mutations and natural selection that leads to mammals?  Creatures with complicated mechanisms like eyes, an immune system, a circulatory system (and elaborate processes of clotting and repair!) and on and on -- how on earth could all of that arise randomly?

 I felt like this for a while.  Obviously not badly enough that I became a creationist or anything.  But it was puzzling.  What solved it for me was reading the first few chapters of Daniel Dennett's book Darwin's Dangerous Idea: Evolution and the Meanings of Life.  Dennett thinks that natural selection is actually (or like, sometimes the distinction is unclear) an algorithm. "Darwin had discovered the power of the algorithm"  (p. 50).  Evolution was not designed to produce us per se, he says, but there's no reason to believe that it is an algorithmic process that has in fact ended up producing us. 

To illustrate his point, Dennett gives us this nifty little diagram.  Think of what natural selection does, he says, as an example of a tennis tournament draw.  Whatever happens, a tournament has to have a winner!  And natural selection picks winners; there is nothing inevitable about these winners -- they are contingent -- but the process of picking winners is inexorable.  And this happens over long, long, periods of time -- millions and millions of years. 

So first, you have an algorithm that picks a winner.  And then you have an algorithm that picks winners over large time-scales.  That nailed it for me.  As a programmer, you are constantly faced with time and space constraints when you program. Think about writing a program to play chess where the program essentially tries to look as far ahead as it can.  How far down the tree of moves should the computer look?  Ideally -- all the way down!  But wait - then it'll take forever for the program to make the next move, so we need to compromise.  Or maybe we can store everything in the memory all at once so that it won't take forever?  No luck again because memory is limited. 

When you work with programs, the power of algorithms is evident.  And the constraints on algorithms are all too visible.  Evolution then is like an algorithm with no constraints; it gets infinite time and infinite space to do its work.  And for something like that, anything is possible -- mammals, conscious mammals, insects, plants, whatever.  It doesn't seem non-intuitive at all.

Now obviously, there's a lot of flimflammery to Dennett's thesis.  Is evolution actually an algorithm?  Or is it like an algorithm?  And Dennett clearly has a lot more up his sleeve: his point is not to make evolution intuitive (that was a byproduct for me; not everyone works with algorithms), but rather to show that evolution is indeed something like "universal acid" -- a concept that can explain the deepest philosophical mysteries: the mind-body problem, consciousness and so on.  Critics, quite rightly, beg to differ.

No matter.  My point was to show that there are ways in which natural selection's workings can be made to seem entirely intuitively.  The trick is to find the metaphor that works for you. 


  1. I use the word sober for a special reason.  You would think Orr would have more in common with Daniel Dennett who at least is not rejecting the neo-Darwinian synthesis like Nagel is (by starting with doubts from intuition).  (In fact, one might argue that Dennett assigns it far too much significance.)    But while Orr's review of Nagel is scrupulously respectful, his review of  Darwin's Dangerous Idea, the book that's the subject of this blog-post, is quite scathing.  His correspondence with Dennett is even more so.  (Dennett doesn't get much love at the New York Review of Books.)     

No comments: