“The Details Left to Chance”
I will start with the quotation from Darwin that advertises the conference. The context makes clear that Darwin was addressing the contingency (in the sense of unpredictability) of evolutionary outcomes, and he was attributing that contingency to chance variation. For a long time thereafter, evolutionary contingency was minimized, and what evolutionary contingency there seemed to be, was attributed mainly to Mendelian segregation and other forms of sampling error. What happened to chance variation as a source of evolutionary contingency? I will consider the “fate” of chance variation in this regard, philosophically and historically, from the early 20th century up to very recent developments in evolutionary biology.
“Drift Driven Diversification”
Elsewhere I have argued that drift is the natural state of evolutionary systems, that what I have called the Principle of Drift is Zero-Force Law of evolution. More recently Dan McShea and I have been working on an even more general statement of a zero force evolutionary law that would predict an ever-present tendency for complexification and diversification at all levels of biological hierarchy. This zero-force law, which we call the ZEFL, unifies a lot that is by now well known. Thus, for example, the implications of our view for molecular evolution are not surprising. The same might be said at the level of macroevolution. However, the same cannot be said for population genetics. In particular I will explore two anomalies raised by our perspective vis-a-vis population genetics. First, random genetic drift is directionless. Yet I will argue that it predictably has directional consequences. How is that? Second, drift is supposed to eliminate genetic variation in populations. Yet I will argue that drift can, and reliably will, drive diversification.
“The Role of Chance in the Evolutionary Occupancy of Phenotypic Space”
Over evolutionary history organisms have occupied only a small fraction of the multidimensional space of phenotypes. This occupancy shows a strong clustering with large regions of the space never occupied. There are a number of reasons for these empty regions including developmental and functional constraints and the tree-like constraints of ancestry. This paper will explore an important alternative, the maze-like topology of the space of alternatives accessible from previous states by sequential changes and the role that chance plays in committing the evolution to pathways that make alternatives inaccessible except by retracing long paths to critical branch points.
“The ‘Biased Chance’ of Genetic Mutations”
The concept of chance is ubiquitous in the biological literature, especially in reference to the occurrence of genetic mutations. In this paper, I will identify and define two meanings of this notion by looking at mutations from the point of view of two biological disciplines, evolutionary biology and molecular biology. Then, I will test the identified meanings in the light of some recent advances in molecular biology, i.e. the discovery of a genetic regulation of mutation frequency in response to environmental conditions, which seems to challenge the chance nature of mutations. Finally, I will argue that both meanings of the concept of chance are valid in the sense that they can account for the biased character of all genetic mutations. I will conclude that a cautious integration of molecular biology novelties in the neo-Darwinian paradigm is possible.
“Demonstrating Chance in Evolution: Lessons From an Early Drifter”
Biologists and philosophers have been extremely pessimistic about the possibility of demonstrating random drift at the phenotypic level in nature, particularly when it comes to discriminating random drift from natural selection. However, examination of a historical case – Maxime Lamotte’s study of natural populations of the land snail, Cepaea nemoralis in the 1950s – shows that while some pessimism is warranted, it has been overstated. Indeed, by describing a unique signature for drift, and by showing that this signature obtained in the populations under study, Lamotte was able to make a good case for a significant role for drift. It may be difficult to disentangle the causes of drift and selection acting in a population, but it is not (always) impossible.
“Random Genetic Drift: A Critique in Historical Perspective”
Random genetic drift has become a primary feature of modern evolutionary biology. Textbooks of evolution all include a section on random genetic drift. This is especially true of molecular evolution, where everyone knows much selectively neutral DNA changes by random genetic drift.
My first argument is that both R. A. Fisher and Sewall Wright misconceived random genetic drift from the beginning. Fisher confused meiosis and random drift along a chromosome. Wright confused inbreeding and random genetic drift. Wright’s mechanism to generate random genetic drift along the loci of chromosomes was “random sampling of gametes.” I will argue that “random sampling of gametes” produces no random drift along a chromosome, and neither do founder effects, inbreeding, random sampling of whole organisms, meiosis in producing gametes, or even small population size. I will show how the experiments of Wright and Kerr 1954, Kerr and Wright 1954, Crow and Morton 1955, Buri 1956, Dobzhansky and Pavlovsky 1957, on random genetic drift do not demonstrate random drift along a chromosome.
My second argument is that random genetic drift in molecular biology, using the same Fisher/Wright model, suffers from the assumptions that are required. One assumption is that linkage along the chromosome disappears in deep evolutionary time. If true, chromosomes should disappear in deep evolutionary time. A more severe problem is that the average recombination rate between any two DNA bases averages about 10-8. For a coding codon with a third base degenerate, this third base would have a highly selected base just before it, and just after it, the first base of the following codon, also highly selected. Thus on average the third base in question would segregate independently every 10-16 generation. The mutation rate for this base would be between 10-6 and 10-9. Thus mutation is at least 7 orders (up to 9) of magnitude greater (10,000,000 to 1,000,000,000) than any possibility of random genetic drift at this base. The same argument would apply to all three bases in a codon in pseudogenes. Movement of DNA by genetic draft (or hitchhiking) is only one of many mechanisms that change DNA around. Mutation plus this DNA movement must hold the answers to what we previously attributed to random genetic drift at the DNA level. Neutral DNA does not have to change by random genetic drift.
Final conclusion: Understanding neutral DNA for 40 years has transformed molecular evolutionary biology from a dull science to a very exciting field of study.
“Genetic Drift vs. Genetic Draft”
In a small handful of papers in theoretical population genetics, John Gillespie argues that a stochastic process he calls “genetic draft” is evolutionarily more important than genetic drift. I explore Gillespie’s argument, examining in particular his measure of ‘importance’ and whether that measure is a reasonable one for assessing the significance of draft relative to drift. More generally, I ask, “If genetic drift is not relatively more significant than other chance processes in evolution, why are philosophers so obsessed with it?”
“Markov Chain Monte Carlo: Taking Advantage of Chance”
Chance obscures what paleontologists can learn about the history of life. For instance, only a small fraction of organisms ever becomes fossilized, and only a small fraction of those is ever found. Thus a long sequence of chance events must transpire between the death of an organism and its eventual discovery as a fossil.
Chance can be an ally as well. In this talk I discuss Markov Chain Monte Carlo (MCMC), a widely applicable stochastic simulation method. Instead of attempting to minimize the role of chance, MCMC instead introduces chance into problems, event those that are deterministic (such as computing the average of a probability distribution).
As an illustration of MCMC, I describe work done with my collaborators on the end-Permian extinction, the largest mass extinction in the history of life. Using MCMC together with fossil evidence and food web models, we estimate extinction levels of plant and animal guilds in the Karoo Basin of South Africa.
“Chance as a Modal Concept”
There are at least six different senses of chance employed in evolutionary theory. One of the senses of chance is contingency, or contingent truths – meaning actual, but non-necessary states of affairs. Linked to Aristotle, this sense of chance claims that contingent events are the result of the conflagration of independent causal chains resulting in a particular event, such as a “chance” encounter with a friend at the movie theater. All species are contingent in this sense; a variety of internal and external forces, as well as random drift, coalesce to form particular species. Missing from the standard interpretation of the contingent nature of evolution, however, is the distinction that can be made between contingent (chance) forms and impossible forms. What are the limits, if any, to the variety of forms that are biologically possible? Are there any forms that haven’t been actualized, not by chance, but because they are impossible? I will suggest a strategy for distinguishing chance events from impossible events in evolutionary biology, thereby clarifying the contingent nature of evolution.