Organic Evolution (aka the "Baldwin Effect"): Intentionality in Evolution?
19 November 2009
Much of the debate that followed publication of the Origin focused on the question of evolution itself, rather than on Darwin’s proposed mechanism. After the fact of evolution had become established, however, Darwin’s proposed mechanism came under close scrutiny. Beginning about 1880, many scientists—including some of Darwin’s most ardent defenders—expressed discomfort with the element of randomness in the theory of natural selection, suspecting that random variation, even if selected, could not alone produce the seemingly endless examples of living biological adaption.
This situation led to the advocacy of various strands of anti-Darwinism, ultimately culminating in a crisis for Darwinism around 1900, a period Julian Huxley would later refer to as the “eclipse of Darwinism”. One of the strands of anti-Darwinism that garnered much interest was the notion of Organic Evolution, as conceived simultaneously by psychologists James Mark Baldwin and Lloyd Morgan, as well as paleontologist Henry Fairfield Osborne, around 1896. The notion was essentially a means by which a learned behavior could spread through a population and eventually become genetically encoded as instinct. What made this appealing to folks uncomfortable with Darwinism is that it granted individual organisms some agency in their own biological evolution—a “kinder, gentler” Darwinism, if you will.
During the discussion we will run a simulation of the Baldwin Effect. Hopefully this will make clear the difference between the two parts of the process: the rise of plasticity (or the ability to learn) (part 1) and the “genetic assimilation” (part 2). I hope that we will then be in a position to ask why, if intentionality (or the ability to simultaneously consider and choose among alternative futures) has evolved (part 1), why have we not lost it (part 2)?
Meeting summary (Paul)
Greg's introduction focused on historical (and continuing) resistance to the notion that adaptive organization in biological systems can be accounted for simply by random variation and natural selection, with no "striving for perfection" or impact of individual learning on the genes transmitted from one generation to the next. The Baldwin Effect suggests a way to re-introduce the latter (and perhaps the former?) into thinking about biological evolution via two steps: "phenotypic plasticity" and "genetic assimilation."
To further explore this possibility, Greg introduced a model of the Baldwin Effect published by Hinton and Nowlan in 1987 (Complex Systems 1: 495-502) suggesting that "the ability to learn has a cost," which in turn raises some quite general questions about the long-term stability of phenotypic plasticity and hence of the likely significance of the Baldwin Effect as well. There was extensive discussion of the legitimacy and usefulness of this model in turn raising a whole series of more general issues:
- Does a model need to be in line with biological findings?
- How generally significant are the observations on this model, for biological evolution? for other things?
- Is there more going on in the model, in the Baldwin Effect, in general than just selection for adaptive characteristics, of which various forms of phenotypic plasticity might be one?
- Does selection operate on phenotypic plasticity, on variability in general? If not, why do these things persist?
Continuing conversation, in on-line forum below