Darwinism Evolving: Systems Dynamics and the Genealogy of Natural Selection.
Darwinism Evolving is a history of the life, times, and fate of one of science's most important ideas. Depew and Weber trace evolution from its pre-Darwinian roots through its development by Darwin and his immediate followers. They then recount its initial rivalry, then synthesis, with genetics, and then the extension of that synthesis to include systematics, biogeography and paleontology through the 1940s and 1950s. The final section takes us to the contemporary debates and future prospects of the idea. The chain that binds all this together is natural selection. For Depew and Weber, Darwinism stands or falls with the fate of natural selection as an explanatory idea. The history of Darwinism as Depew and Weber retell it is thus largely a history of the changing understanding of this idea. The book plays a double role in both depicting the changing understanding of natural selection, and in offering suggestions about how it should be understood. Reasonably enough, their role as participants in the battles rather than reporters of them is much more marked in the final chapters of the book. Here the issue is Darwinism's possible future rather than actual past and present, but even the final chapters are only episodically partisan.
This narrative is impressively contextualized, for a central theme is the continuing importance of physical and especially dynamic theory in the ongoing development of Darwinian ideas. Depew and Weber read this influence not as a reductionist denial of the distinctive phenomena of evolutionary theory, but as a productive source of models and metaphors leading to an explanatorily richer understanding of natural selection. The methodological context is important for Depew and Weber too. They argue that prevailing conceptions of the proper way to do science have been important in the evolution of Darwinism. For example, in their view, the Synthesis's ascendancy owed something to a link with logical empiricist conceptions of science, and the decay of that conception contributed to the softening of the Synthesis.
Let me say at the beginning that this is a very fine book. Depew and Weber have mastered a huge and diverse literature: they have constructed their book from the history of science, evolutionary biology, philosophy of science, and philosophy of biology. The work that emerges is clear and readable. Moreover, despite its size, the reader does not lose the wood for the trees. That is not to say that I agree with all they say. I have reservations both about their historical interpretation and their report on the current state of play.
I have already noted that an ongoing theme of Depew and Weber is the influence of physical ideas on evolutionary theory. They see that influence exemplified in the influence of Newtonian ideas on Darwin himself, in the influence of statistical physics on the between - the - wars synthesis of Darwinism with genetics, and in the alleged "complexity revolution" on tomorrow's Darwinism. I remain somewhat sceptical.
Let's begin with Darwin. They argue that Darwin's picture of evolutionary theory is an instance of a Newtonian dynamics, funnelled through political economy. I think Depew and Weber do show Darwin's debt to the pictures of the economy and the geological world as machines that ran themselves. But these metaphors seem to me to be broadly mechanistic rather than specifically Newtonian. I have no doubt that Darwin was a materialist and a mechanist, and if that were all Depew and Weber meant by Darwin's debt to Newton no doubt they would be fight. Darwin cannot but have been influenced by such an influential model of successful science, both in the way he thought about science himself, and in the way he presented his theory to others. Moreover, Hodge and others have made it clear that Darwin was informed by, and responsive to, contemporary philosophy of science. That philosophy was most certainly deeply influenced by the Newtonian model. Depew and Weber themselves make a convincing case that Herschel's 'law of higgledy-piggledy' dismissal of The Origin was driven by a Newtonian philosophy of science. Variation from heterogeneous causes, and selection for equally heterogeneous reasons, did not count as a unified and lawlike causal mechanism for a Newtonian true believer.
This is all good news for an uncontroversial variant of their thesis. But they often mean something stronger than this in thinking of Darwin as a Newtonian. Depew and Weber characterize Newtonian dynamic systems as those that are closed, deterministic, reversible, and atomist. Moreover, they think of Newtonian explanations as typically externalist: the system's specific trajectory is a consequence of external forces acting to overcome its intrinsic, inertia-like characteristics (pp. 91-2). They think this externalist explanatory scheme fits Darwin's Darwinism. Natural selection is the external force analogous to gravity. This externalism is precisely what distinguishes Darwin's evolutionary ideas from the intrinsically driven, developmentalist conceptions of evolution which both preceded and followed him.
I agree that Darwin's evolutionary explanations are externalist, but the rest of the Newtonian portrait does not seem to me to fit very well. Though Darwin's Darwinism was deterministic, for him biological evolution was neither in a closed system nor reversible. Nor is it clear that Darwin's Darwinism is atomist. Depew and Weber take Darwin to be an atomist because they think of him as an individualist. But while Darwin is famously hard to pick on the reality of species, Depew and Weber's own treatment of his theory is at odds with this individualist characterization. For Darwinian explanations rest on properties of populations which do not reduce smoothly to properties of their component individual organisms. First, Darwin appeals to the role of sex in explaining the integrity of species. As Depew and Weber point out, for a thorough going anti-essentialist like Darwin, the problem becomes one of explaining the apparent reality of species. For Darwin, sexually reproducing populations (not a property of individual organisms) play an essential role in that explanation. Second, the central role of continuous variation in a population does not fit the atomist characterization. Depew and Weber chart Darwin's abandonment of the role of sports in evolutionary change in response to Fleeming Jenkins' 'reversion to the mean' objection. Evolutionary change fuelled by sports does genuinely seem an atomist theory of variation. But it's much less clear that an evolutionary theory dependent on fine-grained individual variation within a population is individualist. The individualism of Darwin's theory could be restored, of course, by taking the atoms to be species or populations. But that would be inconsistent with Depew and Weber's take on the debt to Newton through political economy (pp. 127-8), and with Darwin's somewhat ambiguous attitude to the reality of species.
Thus, a specifically Newtonian image of the workings of the world does not seem to me to play a central and direct role in Darwin's theory. I have similar doubts about the role Depew and Weber accord to physical ideas in the development of the modern synthesis. Just as Depew and Weber see Darwin's Darwinism as indebted to a Newtonian conception of the world, so they see Synthesis Darwinism as developing from, and indebted to, Boltzmann and the probabilistic revolution in physics.
I am left with the same unease. On the one hand, there is once more a sense in which Depew and Weber must be right. The development of the Synthesis was contemporaneous with great developments in ways of thinking about probability, and the slow abandonment of probability measures as a reflection of epistemic limitations in favour of the idea that probabilities are propensities. The conception of probabilities as real, explanatory, and in principle ineliminable was surely of great importance in the birth of the Synthesis. It is also true that some of the Synthesis figures self-consciously saw themselves as extending the methods of physics to evolutionary biology - this is especially true of Fisher but perhaps also of Haldane. Mayr, however, is a clear counter-example. One continuing theme of his great output is the inappropriateness of modelling the biological sciences on physics.
I am not clear how close a tie Depew and Weber see between probabilistic evolutionary theory and probabilistic physics. They give no explicit characterization of a Boltzmannian conception of a theory. They do develop a parallel between Fisher on fitness maximization under natural selection on a population of genes exhibiting a fitness gradient, and the maximization of entropy in a population of particles that begin with an energy gradient. But there does not seem to be more than a formal analogy here. As I understand it, the maximization of entropy is a feature internal to physical systems, whereas natural selection cannot be thought of as internal to a population of genes. Fisher could (and no doubt did) abstract away from the causes of fitness differences. His interest was in the consequences rather than the causes of fitness differences. But no defender of natural selection denies the role of factors external to gene populations in driving those populations to fitness maxima. So in contrast with the maximization of entropy, fitness maximization has an externalist explanation. Moreover, as the story moves beyond Fisher and Wright to Dobzhansky and Mayr, Depew and Weber show that the idea of fitness landscapes becomes increasingly concrete. They are increasingly identified with real landscapes divided in various ways by physical and ecological isolating mechanisms. As this understanding of fitness evolves, the role of models from statistical physics becomes more and more remote. Moreover, Depew and Weber themselves demonstrate (in Ch. 12) that the key theoretical achievement of the Synthesis is its realist, historical understanding of species as evolutionarily isolated populations of organisms. This conception of species and speciation seems to owe nothing to statistical models in physics, though no doubt the 'population thinking' central to this understanding owes a lot to more sophisticated ways of thinking about chance and probability.
Let's move to current controversies. The last twenty years or so have seen an adaptive radiation of (alleged) revolutionary alternatives to, or major revisions of, Darwinism. The final third of Darwinism Evolving reviews these challenges. Assessing these would-be revolutionaries is tough for three reasons. First, it depends on a proper understanding of the new proposals. It is notorious, for example, that 'gene selection theory' has been presented both as a major revision of Darwinian orthodoxy and as an equivalent formulation of it. Second, the proper bet often depends on a difficult empirical assessment: do most species really show a 'punctuated' life history? Do developmental facts really rule out many apparently possible phenotypes? Was diversity really maximal just after the Cambrian explosion? And so on. Third: grant that we have a genuinely new idea with some empirical support: how greatly would it stretch the fabric of Darwinism to include it?
The various challenges - from paleontology, genetics, and development - are by and large logically independent of one another. None the less, there is a rough tendency for response to one to predict the response to others. So at the sceptical end of the continuum, we have those like Dawkins (The Blind Watchmaker ) and Dennett (Darwin's Dangerous Idea ). Their predominant response to these ideas is to think that they are probably wrong, and even if they are not, they are minor wrinkles in the Darwinian fabric. At the credulous end, with the likes of Goodwin (How the Leopard Changed Its Spots ) or Salthe (Development and Evolution ), we get endless burials of Darwinism and many 'new paradigms'. Depew and Weber are certainly not at or even close to the credulous end: their account of these debates is intelligent and fair-minded. But they are certainly closer to that end than me, so let me note a few of my reservations about their take on these issues.
Let me begin with some thoughts on the molecular revolution, and the related issues of neutralism and gene selection. One striking recent discovery is 'neutral evolution': the persistent, perhaps clocklike, replacement of genes by variants which seem to be their phenotypic equivalents. I agree with Dawkins that neutralism is mostly irrelevant to the Synthesis. One could accept the neutralist conception root and branch without blemish to the Synthesis account of speciation, adaptation, or diversity. Neutralism from this perspective adds a new and unexpected phenomenon: the existence of pervasive gene and protein change independent of natural selection. Depew and Weber boost the importance of neutralism in two ways. First, they point out that it shifts our default expectation - out expectations in the absence of selection - from no change to change at full clock speed. Selection changes from a mechanism that causes change where there would be none to a mechanism which damps down change by editing out a few of the changes that are happening all the time. Second, Depew and Weber conclude that since selection damps down the pace of change, the 'creative force' in evolution shifts from selection - now seen just as a culling agent, an editor - to drift (p. 363). The creative force is the mechanism that drives the clock. Here I get off their boat. I do not see why they attribute explanatory primacy to the mechanism producing variation, when both it and natural selection are essential. Natural selection does not just damp down the rate of an endogeneous change. It biases changes in certain directions, and hence the running of the clock together with repeated cumulative selection causes a shift in evolutionary trajectory. Natural selection causes the actualization of phenotypes with complex adaptive structure - phenotypes which would never be found without it.
Depew and Weber also link the molecular revolution to gene selectionism. They see gene selection as a revival of Fisher's idea in the context of the molecular revolution that we can treat each gene as a having a fitness that behaves additively (p. 377). But Dawkins' central argument is that though a gene gets its average fitness in complex, interactive and context dependent ways, none the less it has phenotypic powers in virtue of which it can be selected. Moreover, Dawkins quite deliberately defines the gene qua replicator in functional rather than molecular terms. If anything the molecular revolution is bad news for gene selection. For it makes it unclear whether the functional unit - complex and perhaps even physically dispersed on a chromosome - is a unit of replication.
A continuing theme of Depew and Weber is the role of physical models in evolutionary biology. So they are enthusiasts for the work of those, most especially Kauffman, who think that the survival of Darwinism depends on a major injection of new dynamic ideas from the 'complexity revolution'. Depew and Weber's Chapter 16 serves as a good introduction to Kauffman's Origins of Order . His boosters often interpret his work as showing that order and structure 'spontaneously' arise in complex systems of certain kinds. The result is 'order for free' and hence natural selection is redundant, or confined to a minor role in explaining the elaboration of complex structure that arises autonomously. Whatever the merits of Kauffman's work. Depew and Weber make it clear that neither they nor Kauffman share this reading. It was once common to suggest that any domain characterized by variation, differential success, and heritability would see evolution under natural selection. Not so. Evolution under natural selection depends on cumulative selection which in turn depends on very specific features of the biological world, not just very general properties of it. In one respect, Kauffman's work can be read as a continuation of that line of thought. Selection is ineffectual in excessively chaotic worlds, and is excluded in excessively Laplacian ones in which the trajectory of events derives directly from underlying physical principles. While developing this line of thought, Depew and Weber bring out an interesting ambiguity in Kauffman. One shift in recent Darwinian thought is an increasing emphasis on the particularity of evolutionary history and hence evolutionary explanation; so much so that G.C. Williams elevates 'historicity' to one of the three fundamental tenets of contemporary Darwinism. Depew and Weber see Kauffman as swimming against this tide (p. 455), and endorsing the sufficiency of universal laws of nature in the explanation of much of importance in evolutionary history. I line up with G.C. Williams in this debate. This may explain our differing bets on the importance of the complexity revolution for tomorrow's Darwinism.
Despite these reservations about detail, Darwinism Evolving is a very good book. It is probably too long, and perhaps too historically organized, to serve as the main text for a philosophy of biology course. But its clarity, and its excellent reading guides, would make it an important supplement for any such course. It would be an excellent tool for a graduate group trying to bring themselves up to speed in the area, or for anyone in a related area wanting an overview of the whole field, but one with enough depth and enough sources to ground a more detailed exploration of some of its domains. Its breadth is so great that even those of us professionally involved in these issues will find much that is new and worthwhile.
|Printer friendly Cite/link Email Feedback|
|Publication:||The British Journal for the Philosophy of Science|
|Article Type:||Book Review|
|Date:||Dec 1, 1996|
|Previous Article:||Force and Geometry in Newton's Principia.|
|Next Article:||The Scientific Realism of Rom Harre.|