Wednesday 29 January 2014

Memes are more gene-like on the internet

The internet has turned into quite a crucible for all things meme-related. In the wake of the 2011 internet meme explosion, there have been many papers on memes - which probably wouldn't have been published otherwise. When I got into the field in 2008, it was pretty inactive - and the 2011 meme explosion has been a very welcome upturn.

The whims of youth/geek/nerd culture is probably primarily responsible for the explosion. However, one thing that has probably helped with meme adoption among academics is the fact that memes found on the internet are more like the genes of molecular biology. One of the historical criticisms of memes is that they are not digital and discrete - and so are not much like the genes of molecular biology. Of course, the term "gene" has historically had a quite different meaning in evolutionary theory - but not everyone understands that.

Anyway, memes on the internet are digital and discrete - at least for some of their lifecycle. The ease of classifying the resulting discrete variants makes it more obvious that gene-like dynamics apply to their evolution.

Daniel Dennett described the internet as the drosophila of memetics in 2009. The internet is the main object of study for memetics - but it's also the crucible in which many memes form. Not just the the "drosophila of memetics", but also its "warm little pond".

Many evolutionary processes tend to start off with low fidelity. After a while, they become advanced enough to invent digital copying, there's a digital revolution, and things mostly stay digital from then on. We've seen a digital revolution going on over the last few decades. It seems likely that memes will retain their more gene-like dynamics far into the future - effectively eradicating one of the sources of confusion about memetics in the process.

Sunday 26 January 2014

The host-centric approach to cultural evolution

I am aware of few defenses within academia of the host-centric approach to cultural evolution. Those involved mostly don't seem to be properly aware that there are other approaches - or think that the other approaches are wrong for unrelated reasons.

One of the few attempted justifications I have encountered comes from Peter Turchin. He once wrote:

I still don’t like these memes and cultural survival machines that are disembodied from real people and human groups. They tend to be treated as a kind of undifferentiated agar growth medium on which the real dynamics take place. Within the framework of multilevel cultural selection we do not need to abstract away from humans and human groups.
It is true that the Price equation allows you to partition cultural variation however you like. However that doesn't mean that the different partitions are equally useful - or equally helpful.

Imagine that someone had said:

I still don't like the idea of smallpox viruses. They tend to be treated as a kind of undifferentiated agar growth medium on which the real dynamics take place. Within the framework of multilevel cultural selection we do not need to abstract away from humans and human groups.
It is technically true that using the Price equation you could model selection and variation caused by smallpox viruses at the level of humans and human groups - and build models that produce useful predictions. The 'pox' trait could be considered to be a property of hosts that spreads horizontally, vertically and obliquely between them. You could track the 'pox' trait - and there would be no need to mention hypothetical smallpox viruses. However, it should be obvious that this process misses a lot - and is an incomplete and impoverished way of modelling the situation.

This is the problem with a huge raft of modern cultural evolution literature within academia - in a nutshell.

One obvious case where humans and human groups is not enough is when modeling memes reproducing and competing within individual minds. In intracranial memetics - you have to "abstract away from humans and human groups" - and go down to the level of the meme.

The host-centrism of so much modern cultural evolution literature can't be part of the coming consensus on cultural evolution. It acts like blinkers on the field, causing massive structural damage to it.

Memetics had this right from the beginning. Many modern students of cultural evolution are perpetuating a problematical approach which was rendered obsolete decades ago.

The most comprehensive critique of the idea that I'm aware of is the book Contagious Ideas: On Evolution, Culture, Archaeology and Cultural Virus Theory by Ben Cullen.

Recently, Steven Shennan has been saying much the same thing as me, writing:

The variation, selection and retention processes that underlie cultural evolution were laid out in detail more than 25 years ago (Cavalli-Sforza and Feldman 1981, Boyd and Richerson, 1985) and have been extensively elaborated on since (e.g. papers in Boyd and Richerson, 2005). However, this has mostly been done from an agent-centred perspective and not from that of the cultural lineages themselves - the "memes eye view" - and the two are not the same.
As far as I can tell, this problem became established in the literature with Boyd and Richerson (1985) - who wrote:

This does not mean that cultures have mysterious lives of their own that cause them to evolve independently of the individuals of which they are composed. As in the case of genetic evolution, individuals are the primary locus of the evolutionary forces that cause cultural evolution and in modelling cultural evolution we will focus on observable events in the lives of individuals.
They meant human individuals. However, in cultural evolution there's a symbiosis involved - with human hosts on one side and cultural symbionts - things like bibles and dogmas - on the other. One should not just focus on the hosts! That misses out half of the creatures involved in the process! Students of "genetic evolution" [sic] typically do not make this mistake - since they have a rich theory of symbiology to draw on.

A possible defense of host-centrism in cultural evolution is that it's a modelling simplification. Simple models can certainly be useful. However, one should not then mistake the simplified model for reality - and something like that appears to be happening in this case. When 99% of the papers discuss the model and very few address the gap between the model and reality, we have a clear case of reification on our hands.

However, it isn't really true that we are dealing with a modeling simplification here. We already have models of symbiology. Adding a whole bunch of new models of cultural evolution that partly duplicate these isn't really a simple approach to the whole topic. If you look at symbiology in the organic realm, few people interested in the spread of symbionts bother with host-only models - they just seem ridiculous. I would classify this as a case of bad models - not just simplified models.

Next year, this awful muddle will have its thirtieth anniversary. It seems to me that it needs to die quickly. Hopefully, explaining the problem clearly will hasten its demise.

References

Selfishness != parasitism

There's a longstanding muddle associated with the "selfish gene" terminology of Richard Dawkins. The idea is that genetic selfishness tends to lead to anti-social individual behaviour. This misunderstanding has been transferred by some to memetics. Peter Richerson provides our example of this today:

One of the problems with the meme concept as it evolved is that users of the term focused far too heavily on the selfish potential of memes. But I think it is near to undeniable that cultural variants are sometimes selected to become selfish patthogens along the lines that Dawkins suggested.

Here, Richerson seems to be equating - or associating - selfishness with parasitism. To a memeticist, this is a false equation. Most memes behave selfishly. By that, what is meant that they behave as though they are self-interested agents that act so as to maximize their representation in future generations. Most genes also behave selfishly in this way. That is the meaning of the title of "The Selfish Gene".

Memeticists correctly focused on the selfish potential of memes. Memes are typically selfish - and treating them as such allows you to understand and explain their behaviour. However, selfish genes can reside in deleterious symbionts (like fleas), or mutualist symbionts (like gut bacteria). Selfish memes are the same: they can reside in helpful ideas (such as knots) or unhelpful ones (such as obestity-promoting memes).

Selfishness != parasitism. This is just a misunderstanding of the "selfishness" terminology - when applied to memes and genes.

I've picked on this Richerson quote before - but that was for different reasons.

Saturday 25 January 2014

The germ theory of culture

Numerous students of cultural evolution have got a lot of mileage over the years out of viewing animal brains as the central locus of cultural evolution. Brains are envisaged as repositories of cultural information that selectively adopt it and selectively transmit the cultural information to other creatures.

It's always been a central thesis of memetics that this is only one way of looking at the situation. As well as focusing on the cultural hosts, one can alternatively model the cultural symbionts themselves. The cultural symbionts have their own life cycles and lineages - which are independent of those of their hosts. From this perspective the hosts are one part of the environment. The environment includes multiple other copying devices (such as printing presses) and multiple other sources of selection (such as fires). Animal brains are just one part of this bigger and better picture.

It is worth noting that biology has gone through this conceptual revolution once before. It was once not widely recognized that many diseases were caused by microscopic pathogens - such as bacteria and viruses. The germ theory of disease was developed in the 19th century. Before that time, many diseases were modeled in a variety of ways using many inaccurate theories. Miasma theory held that some contagious diseases were the result of airborne pollutants. Sometimes pathologies were considered to be the results of hauntings, curses, karma, witchcraft or the work of the gods. Some of the theories came close to the right answer. The ideas of demonology and possession attributed some maladies to the actions of external agents that lived inside the host's body. These theories were partly validated by the germ theory of disease - though the causitive agents turned out to be rather different.

I think we are now facing a situation which is similar to that faced by 19th century physicians - with the modern study of culture. The cultural equivalent of the germ theory of disease is the the germ theory of culture. This was proposed by Cloak (1975) and Dawkins (1976). As with the germ theory of disease, the idea faced resistance, and took a considerable quantity of time to be widely adopted - 50 years approximately. Memetics further supports the basic model present in demonology and possession - that pathology is caused by invasion of hostile agents from outside the body - which can potentially be evicted again.

Today people routinely talk about culture "going viral". Cultural epidemiology has been enshrined in the dictionary - in the form of "memes" and "contagions" - which dictionaries explicitly acknowledge take cultural forms. However, the whole the idea still faces resistance within academia. Many still just don't understand that "viruses of the mind" are real things that can cause or contribute to pathologies. The concept is widely denigrated as being an imperfect "analogy". Others describe fear of manipulation by microscopic entities as "Darwinian Paranoia". Others opine that cultures are complex wholes that can't be meaningfully split up into germ-like pieces - and so on.

It is rather tragic to watch this situation unfolding - in the light of all those who died during the development of the germ theory of disease. You might think we had learned that small, information-carrying entities that can be transmitted between hosts can cause real problems. A wider recognition of the germ theory of culture seems likely to lead to more and better diagnosis and treatments of transmissible pathologies - just as happened with the the germ theory of disease.

Friday 24 January 2014

Random mutations are shifting goalposts

Many evolutionists of a neo-Darwinian bent have long claimed that mutations are random.

They don't consider this as a modelling assumption, but rather something that is actually true.

Of course, in cultural evolution, mutations are not random. Instead they are what Donald Campbell (confusingly) referred to as "blind" - which he defined to mean: "based on existing knowledge". This just means that they are nor prescient or miraculous - but rather are naturalistic.

Many have claimed that this makes cultural evolution different from organic evolution - where mutations really are random. However, mutations aren't really "random" in the organic realm either.

Unfortunately, the idea of "random mutations" has turned into a morphing concept. It represents goalposts that shift - whenever the idea that "mutations occur at random" is falsified.

Consider how Jerry Coyne defined what biologists mean by mutations being "random" in his book "Why Evolution Is True". Jerry said:

The term “random” here has a specific meaning that is often misunderstood, even by biologists. What this means is that mutations occur regardless of whether they would be useful to the individual.
This is what I learned the term "random mutations" meant (in evolutionary biology) when I was growing up.

However, now Jerry is saying something very different. He now says this:

What we mean by random mutations is simply this:

“The chance of a single mutation being “adaptive” for the organism (i.e., promoting the replication of the gene in which it occurs) does not depend on the environment in which it finds itself.”

See how the goalposts have shifted here? In the first quote, there's no mention of dependence on the environment. In the second quote we have this entirely new concept.

Of course the reason for this is that the first statement is now widely recognized to be false. Mutations are just not random in that sense. The most harmful mutations are the ones most likely to be the target of error correction mechanisms and it's widely recognized that (harmless) mutations in junk DNA are more common than (deleterious) mutations in coding regions.

Jerry's second statement sounds harder to disprove - but it's still wrong. Basically Jerry's assertion boils down to saying that the fitness of mutations is, on average, environment-independent. However, we know that is not true: mutations are more likely to be adaptive in some environments rather than others. Here's my comment, explaining why the new idea is still wrong:

The chance of a single mutation being adaptive for an organism does typically depend on the environment in which the organism finds itself. Some organisms are on fitness peaks - where the surrounding fitness landscape all lies below them and there are no nearby mutations that give a fitness boost. Some organisms are climbing "fitness ridges" - where there's some scope for going uphill and others are on "fitness slopes". In these latter cases, some nearby mutations do result in adaptive, uphill progress. The the chance of finding an adaptive mutation fairly trivially does depend on where you are on the fitness landscape. So: mutations are not typically 'random' - in the proposed technical sense given in this post.
This is an example of a case where debating the dogmatic neo-Darwinians is so frustrating. You show that they're wrong - but then the goalposts get moved and everyone pretends that that was where they were all along.

In the case of the first definition, the term "random" was excusable. However, in the second case, using the concept of "randomness" has become pretty ridiculous. Jerry's post contains a whole section apologizing for the terminology. However, it isn't just the terminology that is bad here. The concept is wrong as well. Random mutations are a (very useful) modeling assumption. Mutations are not actually random - in any useful sense of the word.

Kevin Kelly is right about this - though for the wrong reasons.

Mutations in evolutionary theory must be constrained. Unconstrained mutations are consistent with any observations - and so explain nothing. One day, biologists will give credit to Donald Campbell for correctly characterizing the constraint on mutations in evolutionary theory. The history books will say that his discovery was not widely recognized among biologists for many decades - due to neo-Darwinian dogma.

Sunday 19 January 2014

The evolution of memes on Facebook

Facebook data science seem to have got into memes recently.

Here's their 2014 article, The evolution of memes on Facebook, by Lada Adamic, Thomas Lento, Eytan Adar and Pauline Ng. There's a PDF.

The authors explicitly go after some of the basic questions in memetics, writing:

So can memes really be modeled as genes? After all, Richard Dawkins originally coined the word "meme” to draw the analogy to genes when describing how ideas or messages replicate and evolve. How would one test the hypothesis that memes undergo a process akin to biological evolution? First, tracing biological evolution is notoriously difficult because one must discern the lineage of specific genetic sequences through generations, without having the genetic sequence of many intermediate instances. But when studying Facebook memes, we have a very unique opportunity to actually trace when copies and mutations occurred, and these are the two basic ingredients in the evolutionary process.

They conclude:

We’ve observed a number of remarkable parallels between how information evolves in a social network and how genes evolve. Drawing these parallels simply hasn’t been possible before for lack of large-scale data containing the evolution histories of many memes. Here we examined near-complete traces of hundreds of memes, collectively comprising over 460 million individual instances. Although the study is limited to just the Facebook context, and just on format of meme (textual status updates), we believe it provides useful insight into the behavior of ideas transmitted via social ties in general.

Here's one of the images from the article. The caption says: "A tree showing the lineage of each variant of the 'no one should' meme".


Parsimony in the social sciences

Parsimony - in the form of Occam's razor or Solomonoff induction - is an important foundation of science.

However, in the social sciences, simplicity is often seen as a bad thing. Often models are criticized as being overly simplistic.

I haven't written about this before - since Boyd and Richerson's Simple Models of Complex Phenomena says practically everything I would say - and much more in far more detail than I would bother with.

However, Jonathan Haight's critique of "the pursuit of parsimony" (see also 6:10 here) makes the subject topical - so now seems like a good moment for my two cents.

Complex models have a number of drawbacks. They are difficult to construct, difficult to use, difficult to maintain, difficult to publish, and difficult to teach. In short, complexity costs.

The argument for complex models is that they are more accurate. This is often not true, but even when it is, complex models are often not worth the associated costs.

Simple models are ubiquitous in science. People still use Newtonian mechanics - even though it is less accurate than general relativity. In population genetics, one often assumes that the population size is infinite, or that gene fitnesses combine linearly to produce organism fitnesses, or that mutations occur at random. If other scientists took the social science approach, no work would get done.

Why are social scientists so obsessed with making their models perfect? I think it's because they are such sharp critics of each others' work. Anything that falls short of total realism gets shot down in flames as a naive over-simplification. Inaccurate social science models aren't just scientific mistakes, they can led to capitalism, euthanasia or Social Darwinism.

Sometimes the charge of oversimplification is genuinely useful. Perfectionism can sometimes be a virtue - but often it is a destructive and paralysing obsession. This has certainly been a common result in the social sciences. These have yet to embrace Darwinism in many cases. Much social science has been stuck in a pre-Darwinian backwater for 150 years. If you reject simplified models, you don't have basic models that can be improved on. Whatever methodology social scientists have been pursuing, it hasn't been working very well.

Anyway, no one should ever write:

Let's retire the pursuit of parsimony from the social sciences. Parsimony is beautiful when we find it, but the pursuit of parsimony is sometimes an obstacle to the pursuit of truth.

Michael McCullough bashes cultural evolution

Michael McCullough's 2014 Edge question piece was on Human Evolutionary Exceptionalism. He followed it up with a blog post on why human groups and human language don't represent "Major Evolutionary Transitions".

Here, I'll contrast my views with Michael's. Michael writes:

Humans' niche construction activities have undoubtedly exposed new covariances between genetic variation and fitness during human evolution, but those activities have neither created that variation nor filtered it, so they don't constitute an evolutionary process.

This is the fallacy that evolutionary processes must be based on DNA genes. Michael doesn't mention DNA - but it is clear that is essentially what he means. He's saying that cultural reproduction of books isn't really "genetic" - and so doesn't really evolve. This is a huge mistake - culture evolves. It has genes of its own, they are called memes. If there's a problem with "niche construction", it is its name - which confusingly classifies a lot of destructive activities as "construction".

On major transitions, Michael writes:

Their second category mistake was to hold up human language as the outcome of major evolutionary transition. To be sure, human language, as the only communication system with unlimited expressive potential that natural selection ever devised, is biologically exceptional. However, the information that language conveys is contained in our minds, not in our chromosomes. We don’t yet know precisely where or when human language evolved, but we can be reasonably confident about how it evolved: via the gene-by-gene design process called natural selection. No major evolutionary transition was involved.
Hang-on! Language evolved by a process of cultural evolution, in which words and syllables adapted themselves to the human mind. Yes, there might have been a little genetic change as a result - as genes adapted to their new culturally-constructed environment - resulting in "babbling babies" and explaining why chimpanzees learn language poorly. However, language is primarily a cultural phenomenon that evolved via cultural evolution. We can see languages evolving in real time, and we can see that their adaptations to their human hosts have precious little to go with changes in DNA genes - since languages evolve far too quickly.

It may still be a little bit early to classify this phenomenon as a "Major Transition" - since it is still going on. Memes have already had an enormous impact on the biosphere, but any "transition" is pretty clearly still "in progress" - and won't be complete until human brains have all been scanned into the matrix.

In his blog post, Michael seems to say that even this wouldn't count - since evolution only works on DNA genes, and humans uploading themselves in the matrix wouldn't make much difference to this - since bacteria would be unaffected (maybe). Maybe. He says:

Evolutionary transitions are about information stored in DNA, not about information in people’s minds.

However if 99% of the biomass on the planet becomes based on memes - rather than DNA genes - that would be a pretty major change - even if the remaining DNA transmission wasn't very much changed in the process. Michael's emphasis on DNA is a mistake - since what we are looking at is a memetic takeover in progress. See my Genes are not sections of nucleic-acid post for more on this whole topic.

Lastly we have this:

In the past, all of these cooperation-related phenomena spent time on evolutionary scientists' lists of "unsolved puzzles about human cooperation." The good news is that scientists have already succeeded in nudging many of them toward the "solved puzzles" list. The bad news is that some scholars have gone in the opposite direction: They have moved these problems onto the list of "mysteries"—problems so perplexing that we should abandon hope of ever solving them within the standard inclusive-fitness-maximizing view of natural selection. Their mystification has led them, at turns, to invoke evolutionary explanations that are inappropriate for species in which all individuals reproduce, to propose new evolutionary processes that are not evolutionary processes at all (but rather, proximate behavioral patterns that require evolutionary explanations), and to presume without justification that certain quirks of modern social life were selection pressures of our deep evolutionary past. Explaining the exceptional features of human cooperation is challenging enough without muddling the problem space even further with conceptual false starts, questionable historical premises, and labyrinthine evolutionary scenarios.

This is a full-throated critique of cultural evolution. Cultural evolution is a fact. Yes, it does involve some rather "labyrinthine evolutionary scenarios", but that's just how our species evolved. You can't reduce the complexity of the process by sticking your head in the sand and pretending that cultural evolution doesn't exist.

Remember, folks, that not much in human origins makes sense, except in the light of culutral evolution.

On the positive side, cultural evolution works a lot like the way in which organic evolution does. So, the symbiosis between memes and our brains is not so much different from the symbiosis between termites and their gut bacteria. We can understand human ultrasociality as the product of memes manipulating humans into frequent close contact to facilitate their reproduction, in much the same way that termites are manipulated by their gut bacteria into frequent close contact to help the bacteria reproduce. Similarly, humans use their memes to adapt to their spatio-temporally diverse environment, just as termites use bacterial evolution to adapt in real time to different food sources.

Nor are humans the only creatures with cumulative cultural evolution. Chimpanzees have it too, and so do whales. We are an exception - in that we have cumulative cultural evolution and an opposable thumb. Every species is different in some way. However, our difference does seem to be one that makes a difference.

Michael evidently didn't feel the paradigm shift yet. He directs the Evolution and Human Behavior Laboratory at the University of Miami. You can be that that is one place where the study of cultural evolution is being stifled.

Saturday 18 January 2014

The concept of 'culture' nominated for retirement

The 2014 annual question at The Edge is: What scientific idea is ready for retirement?

Looking at the responses shows that many of the nominations are misguided. One wonders if they are XKCD386-like attempts to troll the internet in an attempt to attract attention. For example, people take on entropy, altruism, computer science, artificial intelligence, causality - and many other silly candidates.

The concept of 'culture' got nominated for retirement three times, as follows:

As subscribers can probably guess, I am not impressed. Two of the authors raise similar objections. Pascal Boyer writes:

But culture is splendidly diverse only because it is not a domain at all, just like there is a marvelous variety in the domain of white objects or in the domain of people younger than Socrates.
John Tooby writes:

Attempting to construct a science built around culture (or learning) as a unitary concept is as misguided as attempting to develop a robust science of white things (egg shells, clouds, O-type stars, Pat Boone, human scleras, bones, first generation MacBooks, dandelion sap, lilies…).
I think these objections are ridiculous - and unscientific. Living things are diverse too - ranging from elephants to bacteria. However that doesn't mean that we can't have a science of biology - we do, in fact have such a science. It's the same with evolution and genetics - the diverse subject matter involved actually has relatively uniform and comprehensible roots and can be covered by general principles.

The same goes for culture, cultural evolution and memetics. It's embarrassing for me to see that some people think we shouldn't even try for a science of culture. Culture is big and important. It should be - and indeed is - an obvious target for scientific investigation.

Tooby compares the study of cultural interactions with the hypothetical study of "building interactions", saying:

Consider buildings and the things that allow them to influence each other: roads, power lines, water lines, sewage lines, mail, roads, phone landlines, sound, wireless phone service, cable, insect vectors, cats, rodents, termites, dog to dog barking, fire spread, odors, line of sight communication with neighbors, cars and delivery trucks, trash service, door to door salesmen, heating oil delivery, and so on. A science whose core concept was building-to-building influence ("building-culture") would be largely gibberish, just as our "science" of culture as person to person influence has turned out to be.

This analogy has some issues. We do in fact, have a science of biological interactions, called symbiology. One reason we have a science of organism interactions and not a science of building interactions is because organisms are central in biology - while buildings are not. Another reason is that organisms interact more than buildings do. We have a well-known science of brain behaviour (psychology) and not such a well-known science of buildings (though of course some scientists do study buildings). In short, this analogy gets most of the intuitive force that it posesses from the significance and status of buildings in science - and their static, separated nature - and not from the intrinsic silliness of sciences that study interactions.

In fact, a science of culture would cover buildings and many of their interactions - along with many, many other things. It is deep and general - in contrast with the proposed science of "building interactions", which is narrow and specific.

It is a little strange how so many evolutionary psychologists systematically don't understand cultural evolution. You would think that - with their interests in psychology and the mind - people like Pinker and Tooby would be well placed to understand a domain adjacent to their own. Instead, they seem to be among those most confused about it.

As I understand it, the history here is that evolutionary psychologists - like memeticists - have put up with a lot of misguided nonsense about culture from anthropologists over they years. The anthropologists involved usually think that "culture did it" - and can point to the cultural variation that makes them believe this. This irritates the evolutionary psychologists - who seek the shared genetic basis of traits. From this perspective, cultural variation is irrelevant noise that you have to filter out. Enthusiasts for the significance of culture thus seem like the "other team" to the evolutionary psychologists. Since they have spent so long battling bad cultural science from anthropology, they generalize to: all cultural science is bad.

This is unfortunate. Evolutionary psychology has historically studied human universals coded in DNA genes. Memetics has studied human similarities (and differences) transmitted via human culture. Both topics are important ones. These perspectives are both firmly grounded in evolutionary theory - and they should not be in conflict.

Other "Edge annual question" entries relevant to the topic of this blog include:

I reply to Michael McCullough's article in more detail here.


Update 2015-01-06: Alberto Acerbi also responds - Culture: A scientific idea "ready for retirement"?

Thursday 16 January 2014

Meme and gene fitnesses: not aligned!

Jason Collins has another controversial blog post out. This one is about how genetic selection acts to align cultural and genetic fitnesses. He writes:

Now, imagine a mutant in that population that causes people to pay attention to a cultural trait that is more highly correlated with genetic fitness. As these mutants have higher genetic fitness, they increase in proportion of the population, and cultural and genetic fitness are now more correlated. Cultural fitness now promotes genetic fitness. In the long-run, the two will be perfectly correlated (the exception being where cultural traits are neutral to genetic fitness).

Of course, it is well known that deleterious memes are widespread. The obesity epidemic is one example. The smoking epidemic is another example. Cultural and genetic fitnesses are not, in fact very well aligned. So, there is something wrong with this argument.

To see what it is, imagine we are talking about DNA-based symbionts - such as bacteria - rather than cultural symbionts. It is true that genetic mutations in the host that improve relations with 'good' bacteria, and hamper relations with 'bad' bacteria will often tend to spread. However, the bacteria themselves are rapidly reproducing and very numerous. In practice bacterial pathogens have a field day, causing all manner of infections to their human hosts, by simply out-evolving their defenses. Plus the immune system has to compete for resources with all the other organs in the body - including important ones such as eyes and gonads. It's just not going to be perfect.

In Jason's defense, he does go on to say:

The catch in that last sentence is the “long-run”. As cultural evolution can be so much faster than genetic evolution, systems can be far from genetic equilibrium until the genetic response evolves. Fertility in developed countries would be an example of this. There may also be some constraints that prevent perfect alignment, such as the presence of appropriate learning mechanisms.

For some reason, many evolutionists seem to ignore the deleterious aspects of culture - simply treating it as adaptive for its human hosts. Culture, after all, is what resulted in the human domination of the globe.

Yet, look at Japan. It is one of the most culturally-advanced nations on the planet - and the natives have sub-replacement fertility of around 1.3 kids per woman. The extent to which meme interests oppose gene interest in the modern world is enormous. The signs suggest that the conflict of interests between them will increase further in the future, as memes grow in power and numbers, and approach their crescendo. Memes have already overtaken genes on the internet. This is still just the beginning.

Jason closes his article with some highly dubious claims about how cultural evolution differs from the familiar kind:

El Mouden and friends note that there is a host of complications not present in the genetic case. Cultural relatedness can vary wildly across cultural traits, whereas the nature of genetic transmission means that relatedness is similar across most of the genome. Recognising the pattern of inheritance is also a challenge, as ancestor numbers can vary in number and be of vastly different biological ages. In that context, there is no such thing as a standard length of generation.

In fact, although cultural relatedness can vary wildly across shared cultural traits, this is much the same as how genetic relatedness can vary wildly across shared symbionts.

Similarly, DNA-based symbionts can have very different biological ages and average generation lengths.

Cultural and organic evolution only look different in these respects if you choose not to see how they are related.

Tuesday 14 January 2014

Jason Collins on cultural kin selection

Thanks to Jason Collins for recently writing a recent article about cultural kin selection.

Jason is essentially reviewing the paper Cultural transmission and the evolution of human behaviour: a general approach based on the Price equation.

This paper has some problems. Basically, like so much of the cultural evolution literature, it is symbiology-blind. Instead of cultural symbionts, their human hosts are the focus of this paper. The paper writes:

Person A is a cultural ancestor of person B if the value of z person B has was influenced by the value of z person A had

This is not how memeticists look at things. In memetics, the ancestors of cultural traits are other cultural traits. Humans and cultural traits are in different lineages - and so they are not related. A human can't be a "cultural ancestor" of another human - that muddles together the organic and cultural realms.

Cultural traits are the most obvious evolving entities to track in cultural evolution. One can usefully consider the fitnesses and relatedness of monetary units, knots, recipes, songs, phrases - and so forth.

However, the Price equation approach lets the modeller choose how to divide things up. They are free to do this however they like - regardless of whether they are carving nature at the joints or not. Cultural traits - and not their human hosts - are the most obvious thing to track if modelling cultural change using the Price equation.

The approach discussed in the blog post and the associated paper is rather like considering the two humans who share fleas to be related to one another. While it is possible to consider that shared symbionts increase the relatedness of their hosts, it is typically much more useful to consider the fitnesses of the fleas themselves - and their relatedness to each other.

It is true that shared symbionts can cause cooperative behaviours between their hosts. AIDS patients may preferentially associate with other AIDS patients, for instance. In cultural evolution there are more useful examples - shared catholicism memes cause catholics to cooperate with other catholics, for instance. I am not saying that this whole approach is completely useless.

However, it is important to understand that considering memes shared between humans is not the only level of analysis - or even a particularly helpful one. Memes that are shared between my printer and my computer allow them to cooperate, for example. It is important to understand that the billions of sterile cloned dollar bills in circulation are all closely related to one another - and act as sterile workers for the treasury that made them. A focus on the human hosts of cultural traits is a limiting one that misses out the fitnesses and relatednesses associated with cultural traits and cultural artifacts.

Basically, cultural kin selection is much more useful and general than these articles suggest.

Monday 13 January 2014

Culture's leash revisited

I expect that we'll have a widely-accepted consensus theory of cultural evolution at some stage.

That means that some of the current positions on the topic are likely to fall by the wayside.

One historical point of disagreement has been to what extent genes hold culture "on a leash". Memeticists have often insisted that any alleged leash could be broken - resulting in a memetic takeover. However, Wilson (2012) maintained his position on the topic, writing:

I am further inclined to discount the widespread belief that robotic intelligence will overtake and potentially replace human intelligence.

This was also one of Susan Blackmore's objections to Boyd and Richerson's position. She wrote:

Although Richerson and Boyd describe us and our culture as like obligate mutualists, they still maintain that “Culture is on a leash, all right” even if the dog on the end is big and clever.

Peter Richerson revisited the topic in a recent essay. He wrote:

Charles Lumsden and Edward O. Wilson (1981) famously argued that the gene-culture coevolutionary process meant that culture was on a genetic leash. But if cultural processes are actually driving genetic evolution, it is by no means clear that genes control the coevolutionary process in the one-sided way they proposed.

This seems like a bit of an improvement. However, I don't see why we can't all be on the same page on this one. The possibility of a memetic takeover is widely recognized in the circles I move in. Denial that the "leash" could be broken would surely be irresponsible foolishness. A DNA-filled universe is the "Star Trek" fallacy - few futurists think it is at all plausible. It is the job of all scientists to make predictions. If the scientists involved fail to recognize this possibility, they are guiding the rest of us with a lamentable blind spot.

Sunday 12 January 2014

Culture is alive!

Not all memeticists think that it follows from memetics that culture is alive. For example, Daniel Dennett claims that culture shares the living status if viruses, which - he says - are not alive.

However, the Maynard Smith and Szathmary definition of life and the NASA definition of life clearly classify culture as being "alive".

Maynard Smith and Szathmary said life was:

any population of entities possessing those properties that are needed if the population is to evolve by natural selection.

The NASA Astrobiology Institute Definition of Life says that life is:

A self-sustaining chemical system capable of Darwinian evolution.

These definitions of life seem buggy to me, but they're pretty close.

It seems true that we should really have a science of evolving systems - rather than a science of life, since "life" isn't really a scientific concept. However, "biology" seems here to stay - and scientists should probably try to make the best of the situation.

However practically any sensible attempt to give a scientific blush to the idea of life, winds up including culture. Not all culture perhaps. Some cultural artifacts have the same status as hair and nails - the product of a living system, rather than independently alive. Other cultural artifacts are more like dormant seeds. Seeds don't have any of the metabolic characteristics of life. They just sit there, doing nothing. However, under the right conditions, they can sometimes come to life. Dusty books in library basements are rather like dormant seeds.

I think that classifying cultural systems as being alive is a feature - not a bug - of these definitions of life. If someone says that this makes "life" less like human intuitions of what is alive, scientists can just shrug and say that scientific concepts are often counter-intuitive.

On the other hand, something like the NASA definition of life defines drainage basins as being "alive". That is a bit of an edge case, but perhaps that is more like a bug than a feature. It certainly seems as though it is a bug from NASA's perspective.

Saturday 11 January 2014

Population dynamics

"Universal Darwinism" is a convenient term to cover the modern expansion of the scope of evolutionary theory. However, it doesn't seem to be to everyone's taste.

Some have objected to the word "universal" - arguing instead for "generalized". These objections are not technically correct, I believe - though "generalized" is also a reasonable term.

However, the term "Darwinism" also has some problems. In science, it's normally the idea, not the man, that is celebrated. "Darwinism" sounds too much like a religion consisting of followers of Darwin. Some object that evolutionary theory has moved on since Darwin's era, and the term "Darwinism" fails to recognize this properly. Neo-Darwinism has been tried - but alas, used dogmatically - and it isn't clear whether the term is still usable.

This post is about an alternative set of terminology for the topic. The field could be referred to as "population dynamics" - which would be part of dynamical systems theory.

The main problem I see with "population dynamics" is the term "population". For many, this term has a fairly specific technical meaning - which is not really appropriate for the topic. However, I think that this technical meaning is narrow and bad - and that we should just recycle this term to refer to any-old collection of things.

The term "population dynamics" has also been used in a limited way - to refer mainly to births and deaths. I don't see this usage as much of a problem - science can't allow "population dynamics" to mean that.

Memetics as a paradigm shift

One of the problems people have in understanding memetics is that it represents a paradigm shift.

Richard Brodie, in Virus of the Mind (1996), put it this way:

Viruses of the mind, and the whole science of memetics, represent a major paradigm shift in the science of the mind.

In Thought Contagion, AAron Lynch (1998) wrote:

Memetics represents just such a paradigm shift. In a nutshell, it takes the much explored question of how people acquire ideas, and turns it on its head - the new approach asks how ideas acquire people. More exactly, memetics treats both questions as valid, but finds the new form more useful for some topics, the old form better for others.

In my book, Memetics (2011) I wrote:

[Memetics] is a major paradigm shift - probably the biggest disruption in evolutionary theory since 1859. However it isn't just a paradigm shift. Not only is our understanding of evolutionary theory changing, but the way in which evolution actually happens is changing too.

This explains I believe, why some people have such a hard time understanding memetics - when the topic isn't really particularly complicated or difficult. They are still stuck in the old paradigm.

Monday 6 January 2014

A "Questionable Analogy" paper

My last post referred to a paper which referenced:

I have some comments on that paper as well. I think it is a bad paper. The authors claim to have found:

key differences between the two domains that compromise, we think, the attempt to understand cultural evolution on par with genetic evolution.
The paper concludes:

This suggests that in each domain, specific cognitive mechanisms lead to the emergence of domain-specific cultural dynamics. There is therefore no particular reason to build models of cultural evolution based on an analogy with population genetics (Daly 1982).
It is true that there are domain-specific cultural dynamics. However, there are also domain-specific dynamics in the organic realm. That doesn't prevent models from population genetics being useful there. It seems to me that the merit of models based on techniques from population genetics in cultural evolution has been well established empirically over the last four decades. The authors don't really present a case which counters this large body of work.

Instead they focus on the issue of "Transmission Modes" - which they define as follows:

A central feature of population genetics is the reliance on the concept of transmission modes (TMs). A TM is a way in which genetic material is transmitted between individuals.

The authors claim in the abstract:

Modes of cultural transmission are, by analogy with modes of genetic transmission, ways in which cultural information is transmitted between individuals. Despite its importance across the behavioral sciences and for theories of cultural evolution, no attempts have been made, to our knowledge, to critically analyze this analogy. We here aim at such detailed comparison and show that the fundamental role of modes of transmission in biology results mainly from two properties of genetic transmission: (1) what is transmitted generally does not influence the way in which it is transmitted; (2) there is a limited number of simple and stable modes.

However to my mind they fail to establish either of these points.

There are an enormous number of methods of transmission in the organic realm. Genes may be transmitted between individuals by sneezes, by sexual intercourse, by contaminated water, by blood transfusions, by biting insects - and so on and so forth. If you are classifying transmission methods into discrete "modes" and arguing that these differ in number in cultural and organic evolution, the you should present a classification scheme, and argue for its utility in both domains. However, the authors present no methodology for doing that. Also, few other authors do that: "transmission method" is a much more common phrase in genetics than "transmission mode" is - as can easily be seen by performing some searches.

Saying that a transmission mode is "a way in which genetic material is transmitted between individuals seems vague. Are mosquito bites and tick bites different "ways" of transmitting genetic material between individuals? Or is it one "way" - on the grounds that both are biting insects? Are coughing and sneezing different transmission modes? - or one mode? - on the grounds that both involve airborne particles. The authors don't say - and don't provide any way of answering such questions.

Since they don't defend a classification scheme for transmission methods, it seems to me that they have failed to substantiate their case that the organic realm features a limited number of simple and stable transmission modes. Looking at the range and diversity of transmission methods in the organic realm, this claim seems implausible to me.

Also in the organic realm, what is transmitted very often does influence the way in which it is transmitted. When a cold virus is transmitted, it makes the host sneeze. When a warts virus is transmitted it makes the host itch. When a stomach bug is transmitted, the host gets diarrhoea - and so on and so forth. Where there are multiple tranmission methods, mutations may alter the balance between them - showing that genetic control over the transmission method exists.

For me, these simple observations cause the whole thesis of the paper to collapse.

Interestingly, the paper describes itself in grandiose terms, saying:

This article therefore specifies the fundamental properties upon which the analogy between cultural and genetic transmission modes rest, suggests different interpretations of previous data, raises challenging modeling opportunities and develops a new hypothesis regarding the origin of the difference between biological and cultural transmission.

I think this captures some of the excitement of those who seek to characterize the differences between the cultural and organic realms. These explorers are seeking out new laws and new principles. However, step one for scientists is to familiarize yourself with the existing literature. Often, once you have done that, not everything looks quite so new and shiny. To me, many of the supposed innovations associated with cultural evolution fit into the category of phenomena that have already been characterized by evolutionists.

Sunday 5 January 2014

Alberto Acerbi on the differences between organic and cultural evolution

Memetics takes the relationship between the organic and cultural realms more seriously than any other theory of cultural evolution I am aware of. However, the close parallels seen by memeticists are not always shared by cultural evolution enthusiasts in academia. Alberto Acerbi helps us to highlight this difference in perspective today. To him, the close parallel seems wrong - and he explains why.

He recently posted about the differences between organic and cultural evolution on his blog. Here's one of his comments:

Moreover, a part from having effects in cultural dynamics, regulatory traits represent a difference between cultural and biological evolution. This is an “hot” topic in modern cultural evolutionary theory so I do not want to go in depth here (let me just say at least that I think it is interesting to study also the differences between the two evolutionary processes). “Rules” of genetic transmission tend to not be under genetic control, and models of cultural evolution inspired by evolutionary biology tend to consider the rules of cultural transmission (for example the propensity to learn from others) in the same way not under “cultural control” (they are usually considered under genetic control).

...and here's my reply - in which I explain why I regard the parallel as being closer:

Surely, everyone believes that the rules of genetic transmission are under genetic control! Probably, you are talking about how easy the rules are to change by making genetic modifications.

Note that, if you are comparing genetic with cultural evolution, you should really be looking at whether the capacity of organisms to transmit DNA-based symbiotes (e.g. gut bacteria, and parasites) is easy to change by making genetic modifications. In which case, you will see that the way in which such genes are transmitted is relatively malleable. It is easy enough for genetic changes to make an organism more or less likely to sneeze, for instance.

If you adopt this perspective on evolution in the organic realm, you will see that the case of DNA-based symbionts turns out to be closely comparable to the case of cultural symbionts - with respect to how easy it is for DNA to modulate the transmission pathways involved.

If anyone has ever reasoned from organic evolution to cultural evolution - arguing that since genetic transmission is not easy for genes to regulate, it will not be easy for culture to regulate cultural transmission - then probably their understanding of the dynamics of genetic transmission is faulty. There's no need to question their understanding of the relationship between organic evolution to cultural evolution on these grounds - since this is an example of where these processes are very similar. There are, in fact, plenty of known cases where genes can easily regulate the transmission pathways of other genes residing in symbionts.

In the associated paper, the authors say:

Although more flexible modalities of gene transfer exist [34, 35], genes typically propagate to offspring from just two (sexual reproduction of chromosomal DNA) or one parent (asexual reproduction or sexual reproduction of mithocondrial DNA). Cultural information instead can be transmitted in many different ways and, potentially, from any individual to any other individual, which creates the opportunity to regulate the flow of information in a more fine–grained and context–dependent way.
Cultural information typically propagates from one (or sometimes more) parents to offspring as well. However, the parents aren't human, they are cultural. For example, the parent of a bible is typically another bible. To conceptually mix together human hosts with cultural descendants would be to muddle together cultural and organic evolution. If you do that, you typically do get into a muddle about these issues - but the solution is to not do that. Humans aren't the "parents" of cultural information. They are the parents of their own children. Cultural information is unrelated to them. It's in a completely different lineage, which almost never recombines with human genes.

In both the organic and cultural realms, symbionts may potentially be transferred from any individual to any other individual. Organic symbionts - such as flu viruses may be passed between any two individuals. Similarly, in cultural evolution, bibles may be passed between any two individuals. The situation is a pretty close parallel in this regard. That's because both cultural and organic evolution are implementations of universal Darwinian rules.

Of course, there are differences between cultural and organic evolution - but they are more subtle than this. A failure to recognize the similarities results in existing work not being reused - and reinvention-of-the-wheel syndrome. Also, exaggeration of the differences hampers the development of a unified theory that covers both the cultural and organic realms. We have known about the importance of symbiosis for a long time. Culture is well modeled within a biological framework using the idea of cultural symbionts - and the existing theory of symbiosis - as was recognized long ago by the cultural evolution pioneers Ted Cloak and Richard Dawkins.

It is frustrating for me to see modern students of cultural evolution struggling to understand the topic without using the concept of symbiosis. In 2011, Alex Mesoudi managed to write a whole book on the topic without event mentioning concepts from symbiosis at all. To me, it all seems a bit like time-traveling back to the 1950s - before the idea of symbiosis was widely understood.

References

Friday 3 January 2014

Memes in the context of other brain viruses

Memes - especially deleterious memes - have been characterized by many as "viruses of the mind". However, they aren't the only brain viruses out there. They have to compete for that title with DNA-based exogenous viruses. These include Human herpesvirus-6 (HHV-6), cytomegalovirus, rabies, JC virus, HIV, Borna disease virus, LCMV, and probably many others.

Schizophrenia has previously been speculated to be virus-mediated. The Borna disease virus has been linked to it. Endogenous retroviruses were also linked to it in 2010.

Recently, 'Jumping Genes' have been linked to schizophrenia - with significantly elevated levels of transposons being found in brain tissues of those afflicted with schizophrenia - and some other mental disorders.

Probably the most famous behaviour-control brain parasite in humans is Toxoplasmosis. It manipulates human behaviour working under the assumption the that human is a mouse. The Rabies virus is another brain parasite that infects humans and manipulates their behaviour, mostly under the assumption that they are lower animals.

However, there are probably many other behaviour-modifying parasites out there. A better understanding of these might help to improve public understanding of memetics. If people learn that their behaviour is being manipulated by DNA viruses, they may become more open to the proposal that their behaviour is also being manipulated by software viruses.

References

Wednesday 1 January 2014

Cultural evolution in annual reviews

I've previously cited textbooks as evidence for the poor penetration of understanding of cultural evolution. However, textbooks are typically conservative and out-of-date - perhaps this isn't the best way to sample recent developments.

Annual reviews may offer another perspective on the issue. For instance, here is New Scientist's 2013 annual review: of "evolution and life sciences":

It seems as though this scores zero for understanding of cultural evolution.

The reviews from previous years tell a similar story:

The negative effect of carving the humanities out of the life sciences - and then not properly studying them scientifically - continues to be devastating.

As we enter 2014, the field of cultural evolution could evidently still do with some more promotional activity - to improve public knowledge and understanding of this important topic.

The best internet memes of 2013

To get meta, here is a curated list of other folks' lists of the best memes of 2013. Enjoy:

Here are my corresponding lists from 2012 and 2011.