The Baldwin Effect

Individuals can change the conditions of competition for their own offspring by the parents solving problems within their lives that their offspring will later face in their lives. Parents can affect the phenotype of their child by passing their learned behavior on. When a species finds a particular behavior compelling, it will begin to select for individuals that possess the genotypes that best enable them to engage in that learned behavior.
Over time, the changes in behavior, which affects the phenotype, will begin to direct the development of the genotype. In this manner, creatures that are capable reinforced learning will not only have greater freedom to affect their phenotype but will see their genotypes change more rapidly than creatures that are not capable of reinforced learning.
“Baldwin was an enthusiastic Darwinian, but he was oppressed by the prospect that Darwin’s theory would leave the Mind with an insufficiently important and originating role So he set out to demonstrate that animals, by dint of their own clever activities in the world, might hasten or guide the further evolution of their species. Here is what he asked himself: how could it be that individual animals, by solving problems in their own lifetimes, could change the conditions of competition for their own offspring, making these problems easier to solve in the future?”(77, Darwin’s Dangerous Idea)
“Consider a population of a species in which there is considerable variation at birth in the way their brains are wired up. Just one of the ways, we suppose, endows it possessor with a Good Trick — a behavioral talent that protects it or enhances its chances dramatically…
Those few individuals in the population that are lucky enough to have the Good Trick genotype will typically have difficulty passing it on to their offspring, since under most circumstances their chances of finding a mate who also has a Good Trick genotype are remote and miss as good as a mile.
     But now we introduce just one ‘minor’ change: suppose that although the individual organisms start out with different wirings (whichever wiring was ordered by their particular genotype or genetic recipe)…they have some capacity to adjust or revise their wiring, depending on what they encounter during their lifetimes. (In the language of evolutionary theory, there is some ‘plasticity’ in their phenotypes. The phenotype is the eventual body design created by the genotype in interaction with environment. Identical twins raised in different environments would share a genotype but might be dramatically different in phenotype.) Suppose, then, that these organisms can end up, after exploration, with a design different from the one they were born with. We may suppose their explorations are random, but they have an innate capacity to recognize (and stay with) a Good Trick when stumble upon it. Then those individuals who begin life with a genotype that is closer to the Good Trick genotype — fewer redesign steps away from it — are more likely to come across it, and stick with it, than those that are born with a faraway design.
     This head start in the race to redesign themselves can give them the edge in the Malthusian crunch — if the Good Trick is so good that those who never learn it, or who learn it ’too late,’ are at a severe disadvantage. In populations with this sort of phenotype plasticity, a near-miss is better than a mile….
     In the long run, natural selection — redesign at the genotype level — will tend to follow the lead of and confirm the directions taken by individual organisms’ successful explorations — resign at the individual or phenotype level.
     The way I have just described the Baldwin Effect certainly keeps Mind to a minimum, if not altogether out of the picture; all it requires is some brute, mechanical capacity to stop a random walk when a Good Thing comes along, a minimal capacity to ‘recognize’ a tiny bit of progress, to ‘learn’ something by blind trial and error. In fact, I have put it in behavioristic terms. What Baldwin discovered was that creatures capable of ‘reinforcement learning’ not only do better individually than creatures that are entirely ‘hard-wired’; their species will evolve faster because of its greater capacity to discover design improvements in the neighborhood.” (77-79, Darwin’s Dangerous Idea)
Darwin’s Dangerous Idea by Daniel Dennett

Leave a Reply

Your email address will not be published. Required fields are marked *