Pross, Addy. 2012. What is Life? How chemistry becomes biology. Oxford University Press. 200 pp.
Chapter 2 The Quest for a Theory of Life
Pross discusses previous attempts to develop what he calls a theory of life, beginning with Aristotle. The only aspect of Aristotle’s views that he describes, though, is telos. He also characterizes Copernicus, Bacon, Descartes, Galileo and Newton as banishing telos from the universe, instead of only from their philosophical explanations of motion. [It is worth noting that he retrospectively applies the name “science” to what they and others were doing.] Pross quotes Jacques Monod as saying that a purposeless cosmos is the most important discovery of the past 200,000 years. Besides being completely unverifiable and hence clearly unscientific, the supposed discovery doesn’t even seem that obviously useful. I guess you could say it frees us to do destructive experiments on animals, but our current regulations suggest that we don’t think that. Pross says it propels us into a new conceptual reality. What does he mean by that? Pross also adds that Schrodinger, in his What Is Life, said that the explanation of living things would involve as yet unknown laws of physics.
Pross thinks, along with Monod, that teleonomy requires an explanation. Isn’t teleonomy only supposed to be apparent purposiveness? So what is the problem? If we assume organisms lack real purpose and simply obey the laws of chemistry and physics, then there is nothing to explain except our perception of purpose. That may be a problem, the problem of consciousness. Is he going to solve that with his chemistry?
In his section on definitions of life, he carefully distinguishes individual living things, which cannot evolve, from populations, which can evolve, but he then talks about a population of mules, possibly not seeing that there can be no such thing.
He does seem to be on track in suggesting that most attempts to define life fail. The examples given either make mistakes like saying life is self-sustaining without qualification, instead of pointing to reliance on energy inputs, for instance, or only list some characteristics of life as known to us, or seem just ridiculous, like Freeman Dyson’s information definition.
Chapter 3 Understanding “Understanding”
Pross links understanding to induction, citing Bacon. He says all scientific explanations are inductive, being based solely on pattern recognition. True, patterns in some sense must match, but induction is a reasoning process, so it should describe not the explanation but the way it was derived. In that case, it seems clear that deduction plays as great a role as induction in our understanding. In talking about mathematics’ role in explanations, he goes from pattern recognition to pattern formulation, without noting that he’s moving between induction and deduction.
In discussing the problem of where the underlying patterns come from, that is, what is the reality behind them, he denies we can know that scientifically, and he quotes Wittgenstein to that effect. This would seem to put him into the linguistic positivists’ camp, but I doubt he’s that clear about questions like realism vs. anti-realism, although so far, his statements seem consistent with anti-realism. He does however seem to qualify himself at one point by saying that patterns are to some degree subjective. He also distinguishes quantitative, qualitative and statistical patterns. Then we get a dose of pragmatism to the effect that adequate understanding is whatever works. Then, in another twist, he says that the patterns we recognize are only reflections of the underlying reality of nature. Once again, it is not clear whether he’s an anti-realist, as he seemed to say earlier, or some sort of Kantian realist. Could he even be a Platonist? Images of reality?
The reductionism vs holism section doesn’t add anything. The problem is that he’s leaving out any discussion of the environment of life. If you frame the problem as what environment and what inputs do I have to supply to create a self-replicating molecular system that can undergo natural selection, you have a pretty good reductionist program for developing an understanding of life. If by life, you mean the biosphere, then you still have a long way to go, and it becomes necessary to use more complex terminology than what you would use to describe life in a simple experimental system.
Chapter 4 Stability and Instability
Pross agrees with my idea of auto catalysis: if something is auto catalytic the rate of formation increases as there is more of it around: dn/dt = rn provided you maintain steady inputs of reactants, while in a normal chemical reaction with a catalyst dn/dt = r, where n is the concentration of product and r is the rate of conversion of reactants to products. He expresses the idea in terms of the time required to produce a given amount of product, if you have a given amount of catalyst. For the Spiegelman RNA autocatalysis, you should get a logistic growth pattern, because the rate will be constrained by both the RNA and the protein enzyme acting catalytically. This seems like it ought to apply to PCR, for example.
Another thing about the RNA replication reaction is that it is template replication, so it actually yields copies with a highly specific structure – meaning that analogies to information become possible. Is that what all the talk about “information” in biology is, a physical analogy? How would the idea of a physical analogy apply to a computer or a brain? It seems as if information theory is a mathematical formulation applicable to understanding a variety of things, some of which (cells, telephone signals, computers) we think of as physical and others (language) that seem not to be. I would say that what goes on with cells is physical and the information is only metaphorical. A computer seems more problematic, especially since what it does can be represented as a Turing machine, and even though it isn’t a machine but a mathematical hypothesis its relation to meaningful information seems very immediate. Since information theory involves representations in mathematical symbols of concepts that are not physical, why invoke physical analogies? In all the physical systems covered by information theory, is there a point at which a mind is needed to interpret the meaning of the information? That seems to have been the original motivation in fields like cryptography, communications, etc. but in cybernetic systems there may be times when the information is used only by the machine. Still, someone has to eventually determine whether the machine is doing what it is supposed to, at least until we find ourselves in the Matrix, etc. Stephen Hawking apparently worries that this is where Artificial Intelligence is leading us. A biosphere is like that. It doesn’t need to be meaningful to us to be a biosphere.
What about crystal growth? Clonal growth?
What sense does it make to talk about kinetic dynamic stability or about the “efficiency” of maintaining a large population (p. 74) by rapid replication? I would think that in a way, autocatalysis is very unstable, because it tends to exhaust resources so quickly. He talks about Cyanobacteria being around for billions of years. Is persistence of a clade with little obvious development or change the meaning of stability? Success, might be a better term. To me, the Heraclitean flux is the only really persistent feature of the biosphere. Moreover, it looks as if the pace of change is accelerating: metazoans only in the last billion years, a full terrestrial biosphere only in the last 300 million years, hot blooded life only in the last hundred million, and cultural evolution only in the last six million? Is this all the result of auto catalysis? Is dn/dt = rn, where n is “information?”
It seems as if “stability” is not a very good word to encompass the persistence of biological entities through time, given the tremendous range of life histories found among living things. The mathematical complexities are very great (cf. Cole, L.C. The population consequences of life history phenomena. Quarterly Review of Biology Vol. 29, No. 2, Jun. 1954, pp. 103-137) and there are many dimensions to the whole problem of what is it that persists: genes, phenotype, species, clades? What about the stability of Redfield ratios? If true, it is an indication of an extremely widespread pattern. He claims the more stable replaces the less stable. Doesn’t that imply that species should last longer and longer in the fossils record? What is the actual pattern? TO BE CONTINUED