Physical Principles and Biology

13 Dec 2015 13:44

The construction of the universe is certainly very much easier to explain than is that of a plant ...
---Lichtenberg, Aphorisms J 4
I don't exactly mean biophysics in the usual sense (e.g. things like looking at the physical properties of proteins, or the other parts of what used to be molecular biology, before the DNA-sequencers pre-empted that name), interesting though that is. It's more like wondering how much of biology --- especially the Big Things, like evolution --- can be more or less directly explained by physics, or how tightly physics constrains biology. Physics constrains computation, (e.g., Landauer's principle: erasing a bit produces \( kT \ln{2} \) joules of waste heat) --- might it also constrain evolution in an analogous way? How useful are the mathematical tools physicists have come to know and love in understanding biology? And so on.

One long-running speculation along this line is that evolution has something or other to do with thermodynamics. It would be nice to think so, but I've never encountered any argument to this effect which is even remotely convincing; the most prominent one these days is that advanced by Brooks and Wiley in their book Evolution as Entropy: they claim that speciation and natural selection are instances of the increase of entropy. Unfortunately, they know squat-all about thermodynamics and statistical mechanics, and some of their examples lead me to think they don't really understand probability either. --- That said, I'd be willing to bet (in a very modest way) that some version of the thermodynamic formalism would actually be useful in describing evolution.

A nice symmetry to biologists who don't understand physics is physicists who don't understand biology: these also usually claim a connection between physics and evolution, only in the area of self-organized criticality and the supposed drive to the "edge of chaos." This involves less the active errors of people like Brooks and Wiley as sheer impatience with biological reality: as I heard one of its advocates put it memorably, "Details don't matter!" But of course they do. (My paper will Bill Tozier, below, is devoted to this critique.)

Pure mechanics seems to be much more successful at saying interesting things about biology; but much of this, like the square-cube law, is very old (that in fact goes back to Galileo). The work by James Brown (no relation), Geoff West et al., explaining "quarter-power" scaling laws in physiology, also looks reasonably solid (even though it's about circulatory systems).