Not thumping but pumping
One of the frustrating things about Reinventing the Sacred is that it keeps flip-flopping between condensed accounts of areas of science and mathematics I don’t understand well enough and logical leaps which don’t seem justified. So I keep flip-flopping in turn between wondering if it’s my ignorance that’s the problem and wondering if the logical leaps really don’t make sense.
Things come to a head in Chapter 4: The Nonreducibility of Biology to Physics. Here Kauffman picks up an argument which I think is originally from Carl Hempel in his essay Aspects of Scientific Explanation (1965). I don’t have this with me so I’ll stick to what Kauffman himself says:
If Darwin were asked, “What is the function of the heart?” he would answer, “To pump blood.” But hearts also make thumping sounds … yet these sounds are not the heart’s function. Roughly, Darwin’s claim is that the causal consequence for which hearts were selected was their capacity to pump blood, which conferred on the organisms with hearts a selective advantage…
…So we reach an important insight: the function of a part of an organism is typically a subset of its causal features.
We need to follow where Kauffman thinks he’s going with this:
[L]et us grant, for the sake of discussion, that [a] physicist could deduce all the features of the heart from string theory [this would be Kauffman’s paradigm for ‘reduction’ – if it actually worked]. Then what? The physicist would deduce virtually all the properties of the heart [ie including the thumping sounds it makes as well]. She would then have no way whatsoever to pick out, from the entire set of the heart’s properties, the pumping of blood as the causal feature that constitutes its function…
…(Note that functions in this biological sense do not exist in physics.)…
Put like this I am not sure if ‘functions in this biological sense’ actually exist in biology either.
Many years ago when I was an undergraduate reading Natural Sciences we had books and courses with titles like ‘Cell Structure and Function’ and ‘Invertebrate Structure and Function’. I remember being struck at the time by an ambiguity in that word ‘function’. I couldn’t work out if the ‘function’ of something was the answer to ‘What does it do?’ or the answer to ‘What is it for?’. Or when it was half of that beautiful alliteration ‘structure and function’ did the two meanings melt into a rich new composite just right for biology?
‘What does it do?’ and ‘What is it for?’ are two very different questions. The minute we discount any idea of an intentioned ‘designer’ in nature, the idea of ‘function’ as the answer to ‘What is it for?’ can only be metaphorical, whereas function as the answer to ‘What does it do?’ can be as literal as you like.
Talk of the ‘function’ of (say) an individual organ in an organism is a carryover from teleological thinking. This is fine as long as we don’t smuggle teleology in through the back door.
Imagine a depression in an otherwise flat surface of rock. In wet weather it will fill with rain. Does that mean its function is to catch rainwater and make a pond? That would be a strange use of language. But catching water is what it does.
So why do we feel (just about) justified in saying ‘the function of the heart is to pump blood’ if we can also put our hand on that same heart and say we do not in any way imply the heart was designed to pump blood? I think it is because we see the body it is a part of as if it was designed, and therefore see the part the heart plays in the whole.
Until the theory of evolution by natural selection there was no coherent solution to the puzzle of apparent design in nature, other than by positing an actual designer. But with evolution by natural selection we have, if not a complete answer, at least a conceptual framework which the ever-accumulating evidence fits into.
The heart argument tries to attack that conceptual framework. The claim is that we cannot separate the pumping of blood as the causal feature (the ‘function’) which the heart was selected for, from other features like its colour and the sounds it makes.
But this is a practical issue, not a conceptual issue. In another scientific context, if we wanted to establish if it was feature F1 or feature F2 or feature F3 of an object (or substance or event) which gave rise to result R, we would try a controlled experiment. We would take F2 and F3 out of the equation and see if R still happened. So for example if we had a radioactive sample of river-water we might try a process of elimination to see if only one of the substances dissolved in it was to blame, and which one it was.
In theory (and I mean in theory) we could do the same with the heart. We could in theory manipulate the DNA of some poor laboratory animals to give some of them green hearts, some of them silent hearts, and some of them hearts with a defective pump action. We would then track enough generations to see if any of these changes affected their reproductive survival. So in theory we could establish if it was the pumping or the thumping which gave creatures with hearts their selective advantage.
The practical (let alone ethical) implications of such a research project would make it unthinkable.
But take a different example – say metabolic pathways in the Escherichia coli bacterium. The E. coli genome would be more manipulable – as would its environment – and its fast lifecycle would mean we could assess many successive generations. We would have to pick analogues of ‘the heart pumps blood’ and ‘the heart makes thumping sounds’ – like ‘enzyme E1 synthesises protein P1’ and ‘enzyme E1 creates by-product B1’. But the heart example and the E. coli example would be identical in all important respects.
We could possibly generate:
(i) A viable strain of E. coli where enzyme E1 still synthesised protein P1 but didn’t create by-product B1;
(ii) A non-viable strain of E. coli where enzyme E1 didn’t synthesise protein P1 but still created by-product B1; and
(iii) A non-viable strain of E. coli where enzyme E1 didn’t synthesise protein P1 or create by-product B1.
Results like these would bring us close to being able to say that synthesising protein P1 is the ‘function’ of enzyme E1 because synthesising protein P1 is what enzyme E1 was selected for.
This is what ‘function’ boils down to in biological contexts like these. The ‘function’ of a component (eg the heart or enzyme E1) just is what it was selected for. The E. coli example shows we can at least sometimes separate this ‘function’ from any of other features of the component. But only in a very metaphorical way is its ‘function’ the answer to ‘What is it for?’.
More next time…
© Chris Lawrence 2011.