Microcosomos

The Co-evolution of Language and the Brain
Lynn Margulis & Dorion Sagan

This is a remarkable book published in 1986. I read it when it was much newer and I find it valuable to reread much of it again.

The forward, preface and introduction each spend too much time belittling what happened after the Cambrian explosion—I don’t buy it. Their description of what happened before that is, however, unexcelled in my experience. There are in this book several notions at the heart of evolutionary theory that were unpopular conjectures in when the book was new, but due in part to the book are widely accepted today (2010).

As I read about various uses of microtubular mechanisms I wondered why a cell couldn’t apply two such solutions at once. (Walking and Chewing Gum at the Same Time) Computer designers, having developed some sort of mechanism like an adder, will exploit that invention for kindred problems by either timesharing the device, or duplicating it. More complex plants and animals exhibit exaptation where a character or organ serves two purposes at once. Why can’t the cell do the same? I think I know now and I suppose that the authors thought it was too obvious to mention. Here is my explanation: For a cell to exploit nature’s microtubular invention requires some collection of genes together with the controls to express them in the right circumstances. But what are these circumstances? They depend on which trick you want the proteins that build the microtubules to do. It is not too difficult to imagine two sets of control mechanisms, each adapted to one such trick. Such a control mechanism guides the cell thru a temporal sequence of protein concentrations, sometimes modified by exogenous signals, to accomplish a certain task. These protein concentrations correspond to a sort of point in phase space and the sequence thus defines a path thru this space, or an orbit. A particular cell can be at just one place in this space at a time for diffusion of proteins is much faster than production of proteins by the ribosomes—and thus the concentrations of a given protein is uniform thruout the cell. Orbits for two tricks can be joined in control space but with the limitation that the cell can be on at most one at a time. Which portion of the orbit may be controlled by exogenous signals. The logic for the two tricks thus coordinate with each other for there is just one population of proteins available in a single cell. You can’t have one symphony orchestra playing two symphonies at once! The organelles provide local exemptions to this restriction and can be recognized by the computer designer as additional finite state machines. This is also perhaps the initial impetus for multicellular organisms, which the I think authors suggest.

I am incompetent to disagree on most points but I find her skepticism of the variety argument for sex unpersuasive. She suggests clever alternatives to explain its existence but not enough in my opinion to explain its prevalence in plants and animals. She notes that there is no experimental evidence for the variety argument. But that criterion would exclude most of the arguments in the book. Sexual reproduction provides long term advantages which cannot be observed the short experiments that she quotes. Synergy is real—a fact that the book describes well several times over. Synergy is more than mere variation; it is the combination of solutions to diverse problems whose solutions collected into one organism provides a new sort of advantage. Sex allows synergy between mutations or even genetic drift within an extended interbreeding species.

Economists have recently begun to focus on the combinatorial nature of inventions. Kauffman has noted this by extending his notion of the ‘adjacent possible’ from the biological realms to economics.