… it becomes easy to see one of the flaws in memetic thinking. Changes in “culture” differ from changes in biology in that they are not random; they are directed toward a specific challenge or concern.
underverse has made a couple of conflations here that are rife in the literature, and the reason is that it is really hard to disentangle the issues. So allow me to try.
First of all, it doesn’t follow that because one is a cultural evolutionist, one is committed to meme theory, most especially that one is not committed to the Blackmore view that human selves do not exist, and that memes are somehow propagated independently of intentions. That is one view, but not the only one. It may be that we intentionally propagate those cultural items that we do because they are somehow more adaptive to our cognitive dispositions; in other words, they make us want to propagate them.
Intentions in biology are not empty, but neither are they something that evolution depends on (which was Butler’s Word Game Error). Lions have intentions to eat gazelles. Gazelles have intentions to not be eaten. What ends up working depends on factors that are independent of the directionality of the intentions – that is, in how well gazelles evade lions and how well lions hunt. And then there’s the famous quote about whales:
“The species of whale known as the black right whale has four kilos of brains and 1,000 kilos of testicles. If it thinks at all, we know what it is thinking about.” [Jon Lien, “Whale Professor” at St. John’s University, Newfoundland, speaking to the Norwegian Telegram Agency (spring 1995)]
So intentions are a red herring here. Cultural evolution occurs, as does biological evolution, upon all factors, including intentions. If wishes were horses, then beggars would ride. Intentions are not enough.
If we strip the term “meme” of its Dawkinsian connotations, and as a matter of cultural evolution that is exactly what has happened, it stands as a useful stand-in for “cultural items that are passed on with variable fidelity, including ideas, rituals, institutions, traditions, practices, etc.”. In other words it becomes a name for the things in culture that evolve. This is not malign.
Second, it is a fallacy to think that evolution of the darwinian kind (I lowercase the term as an adjective for evolution by natural or other selection) must involve equiprobability of all possible variants. Mutations need not be equally likely. You may have biases in what mutants can form, from, say, constraints in development or in replication of genes, and the end result – selection based on what of the actual variants are the fitter in that circumstance – is still darwinian. So another red herring is that if variation is biased this makes it non-darwinian. So long as there is some variation with respect to fitness, there will be selection (ceteris paribus*).
So biasing the sorts of potential “solutions”** does not undercut the darwinianness of cultural evolution. Intentionally biased variation that was unreasonably successful without selection would do, but this we do not see, contrary to what engineers may think. Engineers may think they come up with solutions by introspection, but they still have to test them out in the real world. And the solutions they come up with by introspection are the heirs to centuries of tradition, trial and error, and instruction. There are no magical inferences here.
Third, the speed of cultural evolution is pretty much the same as the speed of biological evolution. The problem is that the “rate” is not absolute. Speed in evolution is always relative to generations, not to years. I feel that cultural evolution will tend to be roughly the same relative rate as biological, if only because error rates tend to be at or about the same general level in transmission processes.
Finally there is this comment:
There is a reason why cultural evolution happens so much faster than biological evolution, and that is because it is a function of intelligence. It need not rely on random occurrences to proceed.
Intelligence is not magic. The reason why cultural evolution occurs more quickly is that the generation time between instances of the memes is orders of magnitude greater than that of human biology. We have a new generation of human beings every 33 years or so; we have a new generation of memes in a matter of hours or days, in the case of interpersonal discussion and experiment. No special process or properties are required to explain why intelligence works. Since Hume we have known two things: we discover by induction, which is trial and error generalised; and induction works by what has worked in the past. If things change, induction can fail. This is not any different to a darwinian process. Intelligence does nothing miraculous here but apply rules and knowledge that have been applied successfully in the past, along with variations that arise for all kinds of ways.
One aspect of intentional behaviour underverse hasn’t addressed that makes intentionality itself a darwinian process is that it is probably the end result of a quorum of partial neural activities – Minsky’s notion of a “society of mind” (see also Plotkin). Agents do not have immediate epistemic access to the solution – they run through a series of solutions in their heads, based on the accrual of either past experience or of dispositions that were successful in the past. The key lies in finding the right level of darwinian description.
underverse then, in a second post, attempts to make the Baldwin Effect a non-darwinian solution to this conundrum of his own making. He misreads the effect as being a reconciliation of conflicting ideas, but I fail to see why. Certainly Mayr had little time for it, but generally, both Baldwin and the other two codiscoverers of it, Henry Fairfield Osborne and C. Lloyd Morgan didn’t exactly set up the effect in a clear manner, so that is understandable. Later, Simpson wrote that it was possible but unlikely, giving a clearer view. Eventually later work made it clear under what conditions it would occur, and that they were rare.
However, underverse conflates the Baldwin Effect with Waddington’s “canalization”. This is a serious mistake. Canalization is the view that certain developmental trajectories are buffered against change in genes, so that they come out “right” even when there are perturbations. The Baldwin Effect is the view that behaviour can deform the fitness landscape such that variants (which may be themselves biased) that arise later will be fit. That the Baldwin Effect can occur is, I think, unremarkable and certainly not nondarwinian. The question is how often behaviour persists relative to number of generations, to make genetic evolution likely to occur.
Canalisation (the right spelling!) is less unlikely – developmental trajectories are likely to become buffered against change if they occur long enough in a variety of environments, and this is very likely to have occurred. Each of us is the end result of milliards of developmental events, and the range of environmental perturbations are certainly going to have been great. Developmental systems that can withstand such changes will evolve almost certainly under such conditions. To conflate the tow is a sign of misunderstanding.
I fail to see why he writes this:
It is tempting to invoke memes where genic explanations falter, but such an invocation amounts to a deus ex machina of the kind Dennett dubs a “skyhook.”
It would be a skyhook if intelligence were able to overcome the limitations of evolution, and Hume’s Problem, as underverse claims. But a simple cultural evolutionary view is not a skyhook by definition, but a classic Dennettian crane. We learn by trial and error and transmission without prior knowledge of what will work. We do this by building up our knowledge piecemeal. This is a crane. The idea that intelligence is somehow able to clairvoyantly identify what needs to be known, presumably by Putnam’s “noetic rays“, is as sky hookish as it gets.
So even before we get what is shaping up to be a misreading of Lamarck in the next post, I think that there is no case to answer. Cultural evolution is a process that truly explains what is going on in epistemology (but not, I hasten to add, what should happen – evolution is not a teacher about the future).
To conclude, let me say that I greatly appreciate the opportunity underverse has provided to discuss these mistakes, and I am not having a go at him particularly. Even Gould made some of these errors of reasoning; and people continue to do so. It is important to identify what the fitness bearers are in any evolutionary process, and the cultural agents often aren’t the biological agents. I am many cultural agents, for example, but I am at best one biological one. Failing to distinguish the two is where much of the trouble arises.
* To avoid a potential misunderstanding here, note that this is only true if the fitness differential between variants is great enough to overcome sampling error, which in genetic terms is either drift or neutrality. If you have a small population or the fitness differences are small, then you get no selection.
** Evolution is not a problem solver. It does not find solutions. It is a process whereby variants that survive to reproduce more effectively than other variants in the environment of a population will tend to develop into the majority or totality of the makeup of the population. Using intentional terms to describe evolution sounds like a nice idea until it systematically misleads you in your thinking about evolution. In the case of social evolution, however, we do try to find solutions to problems some of the time. So when that happens, evolution of cultural variants, a process exactly analogous to natural selection, will take our proffered solutions and if they are fitter they will tend to be passed on.