Sunday, 30 November 2014

My altruism boycott

Although the term "altruism" is a popular one among evolutionary biologists, I mostly boycott it. In this article I give my reasons:

  • The dictionary defines altruism "as concern for the welfare of others".
  • There's a biological conception of altruism - which means taking a hit (in personal fitness) for another party.
  • Then there's William Hamilton's (1964) conception of altruism - which argues that it means taking a hit (in inclusive fitness) for another party.
It is widely argued that kin selection and group selection are explanations for altruism. However, they can only explain the first two types of altruism.

I prefer the third definition of "altruism". However practically nobody uses it these days. Indeed there's a lot of confusion and muddle surrounding exactly what the term "altruism" refers to. This confusion is well documented by West, El Mouden and Gardner in Sixteen common misconceptions about the evolution of cooperation in humans (See section 6.1.1).

Alas, the definition they prefer is one of the ones I do not like. They say:

An individual's personal fitness is defined as the number of offspring that she produces that survive to adulthood (Dawkins, 1982; Grafen, 2007b; Hamilton, 1964; Maynard Smith, 1983; also termed neighbour-modulated fitness). From an evolutionary point of view, a behaviour (or action) is social if it has fitness consequences for both the individual that performs that behaviour (the actor) and another individual (the recipient). Hamilton (1964) classified social behaviours according to whether the consequences they entail for the actor and recipient are beneficial (increase personal fitness) or costly (decrease personal fitness) (Table 2). A behaviour which is beneficial to the actor and costly to the recipient (+/-) is selfish, a behaviour which is beneficial to both the actor and the recipient (+/+) is mutually beneficial, a behaviour which is costly to the actor and beneficial to the recipient (-/+) is altruistic, and a behaviour which is costly to both the actor and the recipient (-/-) is spiteful (Hamilton, 1964; Hamilton, 1970; West et al., 2007b

It is rather ironic that they cite Hamilton, 1964 here. Hamilton was pretty clear in saying that it was inclusive fitness he was talking about - not personal fitness. He says so plainly - at the bottom of page 14 of the paper.

I'm not drawing attention to this blunder in order to bash West, El Mouden and Gardner. My point is just that - if a paper by these folks trying to clear up confusion in the area is in this kind of conflict with its primary sources, then the situation is confusing indeed.

If I did use the term "altruism" I would have to constantly define it. My usual strategy would be to put up a page rather like this one, saying what I mean by the term - and then hyperlink to it whenever I needed to use the term. However, in this case, my preferred definition is a minority taste. W. D. Hamilton was apparently on my side of the issue - but that was long, long ago. I can't bring myself to abandon inclusive fitness and adopt personal fitness - that just seems scientifically stupid to me. However, rather than trying to fight this battle, I have mostly decided to just abandon the term "altruism", as being too polluted. Instead, I mostly use the term "cooperation". It doesn't mean the same thing - but it is much, much less confusing.

The idea that meme-based altruism is just another kind of mistake

I read another paper Andre and Morin: Questioning the cultural evolution of altruism This paper also discusses cultural relatedness - a key concept in cultural kin selection.

The authors say:

biological altruism cannot be conceived as arising from a process of gene–culture coevolution

Culture does not change the selective pressure on genes
On the face of it, these statements are completely mistaken. What is going on with these authors? They say:

cultural evolution solves the problem of altruism only to convert it into another problem: the fact that individuals learn from others in a way that eventually leads them to acquire maladaptive behaviours.

We already understand this second "puzzle", though. Cultural transmission is adaptive overall - or at least it has been historically. It's what has enabled humans to conquer the planet. It does sometimes lead to humans acquiring maladaptive behaviours - but that's the price you pay for a package that has been positive overall, so far.

The paper is mainly concerned with trying to to make sense of the literature on cultural group selection. The paper argues that much of this literature is rather muddled. That is something that I can agree with.

The paper insists that altruism resulting from cultural transmission must be maladaptive from the perspective of the host.

However, just because something is bad for an agent, that doesn't necessarily mean that it will be bad for populations containing those agents. The classic example of this is smallpox. Smallpox isn't good for you when you get it. However populations carrying smallpox conquered the world - systematically obliterating populations without smallpox. This was no accident - the European colonists won, in part, because they were smallpox carriers.

Culturally-mediated cooperation that harms donors could be like that - causing individual harm and group benefit simultaneously.

The authors close with an argument that altruism caused by cognitive mistakes (e.g. heuristics) is a lot like altruism caused by cultural transmission - which the authors treat as another way of making cognitive mistakes. There's certainly some similarity there.

However, I would point to a significant difference: with most cognitive mistakes, there's no opponent. No other agent is actively trying to manipulate you into making those mistakes. With cultural transmission, though, agents face an evolutionary process with a conflicting optimization target and a rapid pace. Meme evolution is actively working against their own interests - memes are working to get them to make mistakes. This results in significant differences between the dynamics of the processes involved. This is part of the reason why meme-based hypotheses are more fertile as sources of cooperative behaviour than hypotheses involving cognitive resource limitation or over-generalization.

Programmable animals

Humans are programmable animals. The programming language involved varies between humans - some humans are programmable in English, some in Japanese, and others in Swahili.

Other animals are programmable too. You can train dogs, dolphins, birds and other apes to perform tasks. However, few other animals have such advanced languages, programming them is trickier and they have a less flexible behavioural repertoire.

Drawing a parallel with Turing-completeness seems appropriate. Before the 20th century, machines could be built that performed more than one task, or had functions that could be changed by the user. However then the universal computer was invented - and suddenly machines could do anything - at least anything that was permitted by the available memory, sensors and actuators, if you gave them a suitable program.

Humans are the animal version of a universal computer. You can make them do almost anything. Never before has an animal's behavioural repertoire been so flexible or user modifiable. Cultural transmission is a lot like sharing subroutines.

This is not a new observation. There's neuro-linguistic programming - which is all about influencing behaviour using language. There's a book from 1998 by Jack Balkin titled "Cultural Software" that makes the link between culture and software explicit. Daniel Dennett has drawn a similar analogy: while laptops are programmable by virtue of their Java virtual machine implementation, humans have "necktops" which are programmable via their implementation of the "English Virtual Machine". Just as computers may be given instructions in Java code, so humans can be given instructions in the form of recipes, requests and purchase orders. David Deutsch gives human language as an example when describing his concept of a "jump to universality".

What of any use follows from viewing humans as animals programmable via a universal language?

That humans have already made the jump to universality means that this isn't something that has been left for machine intelligence. Of course, machines can be faster or have more memory than humans - but they can't be more computationally universal than humans already are. Humans are already Turing complete.

Perhaps there is some mileage to be had out of viewing education as similar to software upgrades. Uninstalling existing software isn't always easy - sometimes uninstallation takes important components with it by accident.

Another area of similarity is viruses and spyware. When installing new software, one is often concerned about whether its source is trustworthy. Some software resists being uninstalled again later and does not really have its users interests at heart.

There are a few areas of mismatch: Installing culture often requires repetition - this is not normally needed for software. Software often comes with licenses, but other culture does so much more rarely. Culture can involve "non-disclosure agreements" - but these are only rarely used to protect software.

However, overall, I think that the link between culture and software is useful and under-appreciated.


Monday, 24 November 2014

Critiquing a critique of Universal Darwinism

The essay Evolutionary social science and universal Darwinism argues that cultural evolution and "biological evolution" [sic] differ in various ways - and so a close parallel is inappropriate. Several of its arguments are ones that I haven't seen before. Here I will spend a moment going through them. To start with, here's a summary from the paper:

I want to propose here that there are at least four intertwined “details” about the evolution of human culture that differentiate that process from biological evolution,or at least the standard model thereof, in important ways. They are, first, the often major role of human purpose, intelligence, and intellectual interaction, both in the generation of variety, and in the selection process. Second, selection criteria and mechanisms seldom involve directly issues of human survival or reproduction. The well being of certain kinds of organizations may be at stake, but often not. Third, the entity that that is evolving — aspect of human culture — is a phenomenon that cannot simply be characterized as the aggregation of the population of traits possessed by individuals, but has a collective property. These aspects of the evolution of human culture all are involved in a fourth important difference; the way human individuals and groups are involved with culture and its evolution is different in many ways from the manner in which genes and living entities are related in the evolution of species.

In turn:

Human intelligence and purpose often provide selection on human DNA as well as on cultural variation. It is called "sexual selection" and it is standard orthodoxy in evolutionary biology.

If you argue that cultural evolution differs from biological evolution because it "seldom involves directly issues of human survival or reproduction", you should support the same claim for ant evolution, worm evolution, fish evolution - which also "seldom involves directly issues of human survival or reproduction". Yet the idea is that Darwinism applies to ant evolution, worm evolution, fish evolution and human evolution - despite their differences. Evolution need not directly involve issues of human survival or reproduction. This is widely understood.

Next, it is not true that in standard evolutionary theory, the entity that that is evolving must be a phenomenon that can be characterized as the aggregation of the population of traits possessed by individuals. An ant nest is not simply the aggregation of the population of traits possessed by individuals. A multi-cellular organism is not simply an aggregation of the population of traits possessed by individuals. Such a perspective ignores emergence, which is a standard idea in complex systems theory.

The author's presentation of their last point is a bit vague. The author expands on it later, writing:

Even in its more flexible version, it seems to me that a number of the proponents of Universal Darwinism are arguing not simply that cultural change proceeds through a process that involves variation and selection, which was Darwin’s broad proposal regarding the evolution of biological species, but also that it is useful to see cultural evolution as involving gene-like things, and phenotype-like things, with their relationships similar to those in biology. It should be obvious that I am in full accord regarding the former proposition. However, the latter part strikes me as trying to see the details of cultural evolution as like the details of biological evolution. This may not be helpful to attempts to see what really is going on, or at least not helpful regarding all areas of culture.
That's right - cultural evolution has memes - and there are meme products as well as gene products. You can quantify the proportion of the variation in an object that is due to genes and memes. For example, most of the variation in kilts is due to memes, while most of the variation in haemoglobin is due to genes.

In cultural evolution there are more "naked memes" than there are "naked genes" in the organic realm. However, this is likely to be because the organic realm is more mature. There were probably more "naked genes" in the organic realm back closer to the origin of life. It is hard to see how this counts as a real difference between the two realms.

That just leaves mutations that are directed by intelligence. That's a reasonable example of how cultural evolution actually differs from the evolution of nucleic acid and protein systems. Of course, nucleic acids are catching up these days - with genetic engineering - so the distinction is rapidly vanishing.

Of course, one can argue that the brain is not the only organ that outputs adaptive complexity within the lifetime of the host organism. The immune system does this as well - finding adaptive solutions to environmental problems without the host dying. The gut also outputs a stream of adaptive complexity in response to the diet of the host - in the form of bacteria. Seen in these contexts, the brain is not doing such a special job by solving problems.

The alleged intelligent mutations are actually themselves the product of an extended evolutionary process within the brain - involving multiple rounds of copying with selection and mindless variation. Entities copied within brains include axon spikes and ideas. Once you understand that the brain evolves, mutations that are directed by intelligence do not seem so miraculous and special: there is Darwinism all the way down.

Sunday, 23 November 2014

Richard Dawkins warns about Universal Darwinism

In a section of a 2008 article in The Guardian, titled Why Darwin Matters Richard Dawkins wrote:

Although Darwin's theory can be applied to much beyond the evolution of organic life, I want to counsel against a different sense of Universal Darwinism. This is the uncritical dragging of some garbled version of natural selection into every available field of human discourse, whether it is appropriate or not.

Maybe the "fittest" firms survive in the marketplace of commerce, or the fittest theories survive in the scientific marketplace, but we should at very least be cautious before we get carried away. And of course there was Social Darwinism, culminating in the obscenity of Hitlerism.

Less obnoxious but still intellectually unhelpful is the loose and uncritical way in which amateur biologists apply selection at inappropriate levels in the hierarchy of life. "Survival of the fittest species, extinction of poorly adapted species" sounds superficially like natural selection, but the apparent resemblance is positively misleading. As Darwin himself was at pains to point out, natural selection is all about differential survival within species, not between them.

So: "Universal Darwinism" is "obnoxious" - according to Richard Dawkins! Frankly, an excess of caution is what got humanity into the current scientific mess surrounding Darwinism. It is why 155 years after Darwin published, the social sciences are still mostly stuck in a pre-Darwinian time warp. The precautionary principle is a disastrous approach to science and policy alike - as adeptly explained by Max More in The Perils of Precaution.

That Darwinism leads to "Hitlerism" is what the social scientists seem to think. Where's the science behind that? Understanding of Darwinism has not - to my knowledge - been shown to be correlated with any kind of poor behaviour. The nearest thing I can think of is the work which shows that economics students are more selfish than average - and that seems to be a bit different to me. The idea that understanding Darwinism somehow leads to bad social outcomes is currently a pseudo-scientific claim, not supported by the available evidence.

I'm not terribly concerned about "getting carried away" by Darwinian excesses. It seems like a non-hazard to me. What I am much more concerned about is the ignorance, stupidity and conservatism that are preventing the understanding of Darwinism in modern times. I am not talking about the clueless theists, but rather the anthropologists, historians, psychologists, developmentalists and physicists who should be taking advantage of Darwin's insights, but currently aren't. Some of these people are influential folk who educate the next generation and influence policy. We need to make sure that they have a basic working understanding of their own fields - and that will inevitably include the basics of Darwinism.

As for high level selection not really being "natural selection", Dawkins is wrong here - according to my understanding of these terms. Are we to believe that differential survival of species is not "natural"? Or that it is not "selection"? Of course differential survival of species qualifies as being "natural selection".

The idea that species selection is not a kind of "natural selection" is probably based on using "natural selection" as a technical term with a counter-intuitive meaning. I often dislike terminology that uses ordinary English words and gives them a counter-intuitive technical meaning. "Natural" and "selection" are ordinary words that work together pretty well in the context of Darwin's theory. There is absolutely no good reason to assign them a counter-intuitive technical meaning.

Of course, it would be possible to debate to what extent adaptations existed to benefit species - but that seems like a very different issue to whether the term "natural selection" applies to species.

Rob Boyd: The puzzles of cooperation

Here Rob argues that the idea cooperation between humans which is not due to kinship must be due to reciprocity is wrong. That's correct. There are also manipulation and virtue signalling to consider - to mention just two other mechanisms.

Saturday, 22 November 2014

Large swarms of stupid minions

One of the reasons I got involved with cultural evolution and memetics was the belief that humanity badly needed a solid science of cultural evolution to help successfully navigate the transition to a world economically dominated by machine intelligence.

The coming memetic takeover will be a kind of genetic takeover - and there are good reasons to believe that these are likely to be disruptive evolutionary events - accompanied by mass extinctions and the loss of significant quantities of adaptive information. A rapid transition seems as though it would be undesirable - with an increased chance of things getting lost or damaged.

One of the conclusions my my studies so far so far has been that the geologically-recent explosion of cultural evolution that we are now witnessing was triggered more by social networking skills than by factors associated with brain size or intelligence.

Human behavioural imitation apparently required the complex ability to mentally put yourself in another person's shoes while watching them perform tasks. For our ancestors, this appears to have required relatively advanced cognition - as ably explained by Susan Blackmore in The Meme Machine. So, there is a limited sense in which the conventional wisdom that "intelligence did it" is correct.

However, in memetics, large brains are seen more as a consequence of cultural evolution than a cause of it. Big brains have evolved to be meme nests. Large brains are the nervous system equivalents of ant domatia. They are homes for memes. Psychological support for cultural transmission - rather than intelligence - was really the key here.

This strongly suggests that sub-human level machines could effectively reproduce the human explosion in cultural evolution. Machines can copy each other easily. We can engineer them to be social. I think this means that we can forget about attempts to directly reproduce "human level" machine intelligence, and work instead on swarms of relatively stupid minions. Then the power of collective intelligence and the wisdom of crowds can be used to get them to perform useful work for us. It will be a new kind of society of mind.

Machine progress has occurred largely by them being strong in domains where we are weak. If the aim is to reproduce human cultural evolution in a machine-based substrate (in order to better make progress) then a frontal attack on directly reproducing human cognition in machines doesn't seem to make very much sense.

To a large extent, I think that using swarms of minions is largely what humans have been doing anyway. We do already have huge numbers of not-too-smart computer systems - and we have been a putting considerable amount of effort into networking them together.

Today, much of the main action is in the process of moving out of brains and into data centers. I expect data centers to become the main social centers for machines. The coming explosion of machine intelligence looks set to take place in the computing cloud.

Interestingly, data centers are usually out of town - where land is cheap. The centers of machine civilization and human civilization thus look set to be geographically separated - although intimately connected by high-speed networks. This will create an interesting dynamic.

Friday, 21 November 2014

Meta memetics #fail

Religious folk fairly often disagree with memetics. Some of them write whole books about the topic - and so we have bizarre works like:

One of the tactics is to argue that, if religion is a plague of viral memes, then so is science - and so is memetics - hah! take that, science!

It looks as though the religious apologists just got some support for their "theory" from academia. Two new books (published this year) explicitly treat memetics as a religion. Here they are:

The blurb from the first book reads, in part:

This book presents an objective method for understanding and comparing belief systems, irrespective of their subject matter and of whether or not the investigator happens to agree with them. The method, descriptive logic, is illustrated through analyses of various phenomena, including Zoroastrianism, Dawkinsism, Fabianism, 9/11 Truth, 'alternative' Egyptology, Gnosticism, flying saucer sightings, and the hymns of Charles Wesley.
"Dawkinsism", eh! I can just imagine Saint Richard rolling his eyes towards the "heavens".

Thursday, 20 November 2014

First mover advantage

The concept of first mover advantage is perhaps more familiar in cultural than organic evolution. Many people are familiar with the idea that early players have advantages in business contexts.

So: my task here is rather the reverse of my usual one. Frequently, I claim that some aspect of organic evolution applies to cultural evolution more than might naively be thought. Today, I'll argue that a well-known aspect of cultural evolution also applies to the organic realm. The theme is the same though: evolutionary dynamics in the two domains are more similar than they at-first appear.

Early players in a space harvest easily-accessible resources - which are then not available to later arrivals. They use these resources to reproduce, grow and adapt themselves to the niche. This raises the barrier to entry. Newcomers are in unfamiliar territory and they face competition which has already adapted to the niche and is fighting on familiar ground.

Since the issue is fairly clear and simple I don't plan to spend too long on it. Many animals are territorial - and the territory owner often has a significant advantage. Like those inside medieval castles, territory holders can build traps for intruders, build barriers to impede prospective invaders and surround their home turf with an inhospitable moat.

The hermit crab that gets into a shell first has a clear advantage in any dispute over it. The mason bee that finds a hole first will dam it up, and take it out of service. Late-arriving bees simply don't have so many accommodation options.

First mover advantage is a thing for parasites and mutualists too. Famously, if you get infected by , you won't be colonised by smallpox. This was the basis of the first smallpox vaccine. There's an advantage to penetrating the immune system before it has built any defenses. Similarly, an ant colony in a thorned Acacia is pretty hard for later arrivals to displace.

Though first mover advantage exists in the organic realm, it isn't always enough to result in success. The marsupials in Australia had the advantage of being first in their territory, but they are busy following in the footsteps of the marsupials in South America. They are busy being displaced by placental mammals. New Zealand's birds were also unsuccessful early arrivals. Sometimes the first colonization wave gets overrun by later waves.

Since the focus of this blog is memetics, I should mention how first mover advantage applies to memes. Like parasites memes benefit from getting into young and inexperienced hosts. They face a weaker memetic immune system and suffer from reduced levels of competition from other existing memes. As a classic example of first mover advantage among memes, the QUERTY keyboard layout has become a locked in standard. It's main advantage was early adoption.

Thursday, 13 November 2014

The long reach of the meme

Richard Dawkins' 1982 book "The Extended Phenotype" is subtitled "The Long Reach of the Gene". Most of its example are very gene-centric. It discusses beaver dams and spider webs to illustrate the concept of genes having extended effects. However, what about memes and cultural evolution? Memes have a long reach - just as genes do - as was pointed out by Steven Jan in the year 2000.

When a pagan is stoned by Christians, the memes in the bible are exerting effects that go far beyond the bounds of the book and the minds of the Christians that form their main inheritance pathway. Memes that have spent most of their reproductive life in a camera manufacturer in Japan may deter criminal activity by humans half way around the world. Memes that only reproduce inside NASA have created meme products - such as the Voyager spacecraft - that have traveled beyond the edge of the solar system.

Overall, the long reach of the meme leaves the long reach of the gene panting far behind.

Tim Tyler: Manipulation


Hi. I'm Tim Tyler and this is a video about manipulation. More specifically it's about the role of manipulation in producing cooperative behaviour.

Manipulation involves a biological entity skilfully influencing another biological entity - for its own ends.

Manipulation can be deceptive or honest. A manipulator may be forceful or subtle. They can target their victim's body, their perceptions or their environment. Manipulators may use rewards or punishments to help to elicit the behaviour they desire. Or they might use weapons, drugs or misdirection.

Manipulation one of the types of biological interaction which is capable of producing cooperative behaviour. A classic example of manipulation producing cooperation involves cuckoo foster parents. Through their own feeding behaviour, the hosts take a reproductive hit on behalf of a non relative with no hope of it being repaid. They do this because they are being manipulated. The cuckoo chick fools them into mistakenly believing that it is one of their own offspring.

Although, in this example, kin recognition is involved, manipulation is a different idea from kin selection. It need not involve relatedness. The example of a cuckoo chick shows that it can take place between individuals of different species. The creatures involved need not have much in common. Manipulation is also a different idea from reciprocity: with manipulation, the victim need not benefit - and quite often they don't benefit.

The definition of manipulation I gave mentioned that it took place between "biological entities". That's a intended as a broad category that can include anything from individual genes to entire governmental departments.

Manipulation is common. Parents frequently manipulate their offspring - for example by punishing them. In turn, offspring manipulate their parents - for example by crying when they want attention. Sellers try and manipulate buyers into thinking their products have high value. Buyers try to manipulate sellers into thinking they are short of funds but might still go for the right deal. Manipulation is also a common mechanism which produces cooperative behaviour. It helps to keep workers in eusocial colonies in line. It helps avoid genes on chromosomes defecting against each other by bypassing meiosis. However, despite manipulation being widespread, it is much less well known as a source of cooperation than kin selection or reciprocity.

Manipulation is common in symbiotic relationships. Some parasites manipulate their hosts into contact with other hosts - since parasites require contact between hosts to facilitate their own reproduction. To give three examples: Toxoplasmosis makes rodents more likely to interact with cats; the rabies parasite promotes contact with other prospective hosts and malaria-carrying mosquitoes are more likely to bite humans.

Humans also manipulate other humans for their own benefit - and for the benefit of friends and relatives. Some biologists have called some sorts of manipulative behaviour among humans "Machiavellian" - after the writer Niccolò Machiavelli, who appeared to endorse political strategies involving cunning and duplicity.

Unfortunately, manipulation is poorly-understood as a mechanism capable of producing cooperation. Martin Nowak fails to mention manipulation on his list of mechanisms favouring cooperation in his book SuperCooperators. Karl Sigmund doesn't mention manipulation in The Calculus of Selfishness either. Gintis and Bowles don't treat the topic in A Cooperative Species. Manipulation was slow to be understood historically as a mechanism responsible for producing cooperative behaviour and still today remains an under-appreciated force.

In cultural evolution, memes induce pro-social behaviours in humans by manipulating them. It seems likely that they do this partly because human friendship promotes the contact between their hosts that they need to spread. Memes use promises, threats, sex, desire, misinformation - and numerous other tricks to manipulate humans into being nice to other humans. Memes may be engineered to do this (e.g. by prospective human recipients) - or they may evolve via natural selection to behave in this way.

Manipulation is implicated in the evolution of eusociality. A queen will often manipulate their offspring to make them better serve her. This manipulation typically results in colony-level cooperation. Manipulation also is the basis of the symbiont hypothesis of eusociality. Originally developed to explain cooperation between termites, the symbiont hypothesis holds that host eusociality arose, in part, because it facilitated the transfer of symbiotic microbes down the generations. Each symbiotic microbe must regularly find new hosts. They do this by finding their way into young termites - where they rapidly multiply and adapt - successfully repelling subsequent invaders. This requires contact between hosts, which the symbionts facilitate by manipulating their hosts.

Memes are well known for doing something similar. They colonise the minds of children, and once established there are hard to displace. From the perspective of memes, children's minds represent especially valuable real estate to control. An early beach head in a host can avoid direct combat with a fully-developed immune system; there are fewer existing inhabitants to compete with for resources; a young mind offers the most time for adaptation to the host's environment - and there's a lifetime's opportunities ahead to spread to others. Memes get into children's heads partly by manipulating the behaviour of adult instructors.


Sunday, 9 November 2014

Cultural kin selection may have driven imitation capability

What drove the human ability to transmit culture with high enough fidelity to support the current cultural explosion?

There are a couple of common explanations for this:

  • One explanation invokes DNA evolution. This explanation says that acquiring intact memes was beneficial to their hosts - and so acquiring them without error was favoured.

  • The other explanation involves cultural evolution. This suggests that culture evolved in order to improve its copying fidelity. Gesticulation led to grunts, which led to speech, which led to writing, which led to printing, which led to the internet - with the copying fidelity increasing at every step. Here the benefits of high-fidelity copying accrued primarily to the memes involved - not to genes.

These explanations are not mutually exclusive - and fairly clearly each played a role at different points in time.

Recently, I've seen another kind of explanation which involves cultural kin selection:

In DNA-based kin selection, your genetic relatedness to another human is an accident of birth - something that you can't easily change. The best you can do is to try and manipulate perceived relatedness cues. In cultural kin selection, the situation is a bit different. The proportion of memes you share with another human is not fixed. You can fairly easily increase your memetic relatedness to another human - by the process of acquiring memes from them - or perhaps their teachers or associates.

It has long been known by psychologists that humans manipulate other humans by imitating them. Interview technique books are full of advice about mirroring your interviewer's posture and copying them in other ways - in order to appear more similar to them. The idea is that this process may have actively pressured humans into improving their imitations skills - in order to appear more similar to other humans, so as to better manipulate them. The improvement could have involved DNA-based genetic evolution, cultural evolution - or a combination of the two.

This is an intriguing story - partly because we can see the process involved acting today. However, I think the idea needs more comprehensive study and quantification. Chimpanzee studies are one area which might illuminate the issue. Chimpanzees have the ability to transmit information down the generations culturally. However do they also imitate each other - in order to appear more like kin to each other? It is an interesting question.

I came across the idea in this article while researching my article on homophily. The paper is in the references for this article. I'm not yet sure where the idea originated.



Kin selection results in greater cooperation between relatives. A side effect of this is greater levels of interaction between similar organisms. A name has been given to this phenomenon: homophily. As the etymology suggests this term means: liking those similar to you. This idea is captured in the proverbial saying: "birds of a feather flock together".

Here's what Wikipedia says on the topic:

Homophily (i.e., "love of the same") is the tendency of individuals to associate and bond with similar others. The presence of homophily has been discovered in a vast array of network studies. More than 100 studies that have observed homophily in some form or another and they establish that similarity breeds connection. These include age, gender, class, and organizational role.

Individuals in homophilic relationships share common characteristics (beliefs, values, education, etc.) that make communication and relationship formation easier. Homophily often leads to homogamy—marriage between people with similar characteristics.

The term was coined in the 1950s. More recently, a significant literature on the topic has developed.

Kin selection seems to have been largely ignored or rejected as an explanation of homophily - apparently because it takes place between similar organisms - who need not necessarily be kin. However a broader interpretation of kin selection that includes memes as well as genes positions kin selection very centrally as a theoretical explanation of homophily. Almost all similarity in nature is based on copying - whether through blood kinship, mimicry, teaching, behavioural imitation or learning in shared environments. The fundamental logic of kin selection is based on copying heritable information - and so it applies to all of these phenomena.

There have been some studies of the evolution of homophily - including one published in Nature. However, this was Funded by the Templeton Foundation, and authored by kin selection hater Martin Nowak. Needless to say, it makes no mention of kin selection. It's proposal is that homophily offers direct fitness advantages. It gives an example: "homophily may yield fitness advantages because individuals using the same mode of communication may be able to act together more effectively". Sure, but a shared communication system is going to be down to shared genes - or shared memes. Kin selection - or cultural kin selection is thus applicable. Normally a cooperative system featuring shared genes or memes would be followed by mention of the work of W. D. Hamilton. Not here, though: this is ideology, not science.

The study of homophily has resulted in a significant literature, most of which bears pretty directly on the topics of kin selection and cultural kin selection. There's quite a lot of quantitative data available on the topic. Many of the studies involve humans as subjects - helping directly with the study of cultural kin selection. So far, the topic has lacked a central organizing principle. Kin selection neatly explains homophily. The study of homophily to date has produced an abundance of highly-relevant data. Now that there's a good theory to ground and guide our observations of homophily, progress in the area should come more rapidly.


Saturday, 8 November 2014

Culture vs innovation and norms

I think that memetics should cover all of cultural evolution. However various more narrow kinds of study have also been proposed:

  • Diffusion of innovations;
  • Evolution of norms;
  • Evolutionary epistemology;
While specialization has its merits, it seems to be that there's too much Balkanization here. I've long been aware of the work on diffusion of innovations - and how it myopically parallels the cultural evolution literature. However, there appears to be a similar phenomenon going on within sociology - where a whole raft of researchers obsessively focus on the evolution of social norms. I'm sure that the evolution of norms is a very interesting topic. However, most of its regularities and rules apply to other forms of culture as well - and it often seems as though that isn't being appreciated.

Looking at the literature on norms and innovations it is hard to avoid the conclusion that many of those involved often don't see the bigger picture of cultural evolution.

The overall picture is reminiscent to biology before Darwin. The topic was fragmented. Darwin brought a theoretical framework that unified it all.

Even among those who understand cultural evolution, Blackmore (1999) and Boyd and Richerson (1985) once proposed that only imitation and teaching were significant - and that these should be the main objects of study for cultural evolution. Yet culture that is transmitted in other ways mostly obeys the same rules. I think these folk have now mostly retracted their earlier opinions on this topic.

Monday, 3 November 2014

On Coyne on Kelly

I read Jerry Coyne's review of What Technology Wants - and now duty calls.

Coyne argues that technological determinism is "a theory of such mind-blowing generality that it can’t be disproved". I think that this is contrary to conventional understanding of the issue - we can see to what extent technological determinism is true by looking at convergent evolution - much as Kelly does in What Technology Wants. Coyne goes on to write:

Kelly is strangely keen to tie his theory of technological development to biological evolution. I am not sure why; perhaps he thinks his progressive view of technology is more credible if it’s seen as an extension of the established scientific vision of evolution. But his take on biological evolution is one that, while beloved of creationists, is completely rejected by scientists: he sees it as teleological, driven by external forces to achieve certain goals. Sadly, evolution doesn’t work this way.

Coyne's answer to technological determinism is Stephen Jay Gould's "contingency". He rattles off Gould's argument that evolution is directionless as though it is an established fact. Let's get one thing straight here. Gould was politically motivated to weigh in against evolutionary progress - because progressive evolution punches big holes in the precious idea that everyone is born equal. If there is evolutionary progress, that would probably mean that there are higher and lower races and cultures - and then, oh, the horror. Gould's scientific egalitarianism was political, not scientific - in much the same way that his farcical book on intelligence testing was. Gould is a terrible authority to cite on this topic - because of this obvious politically-motivated bias.

The truth surely lies somewhere between the extreme positions. Contingency plays a role, but there is clearly evolutionary progress - due to technological determinism and its pre-human equivalent. Not just the passive kind of progress that Gould fantasized about - but one due to progressive optimization. Ecosystems are getting progressively better at seeking out energy gradients and dissipating them. Evolutionary progress is one of the commonalities between biological and cultural evolution. Just as cellulose and chloryphyll allowed some organisms to better degrade sunlight, these days, progress in solar panel technology is repeating this feat. This isn't some kind of idle coincidence - evolution is a gigantic optimisation progress. It accumulates natural and human-made technology / innovation - which gives it a ratchet effect: early there are few inventions; later there are many. Under the sufficiently low-frequency of cosmic bombardment that we happen to observe, that process results in evolutionary progress.

Maybe when our descendants are all near-perfect angels there will be a shake-out of accumulated evolutionary contingency. However, for the moment, there's some contingency too. Bats and birds both have wings, but there are some differences - for example birds have feathers. Dolphin brains and human brains exploded in a similar fashion, but they aren't exactly the same.

In his review, Coyne argues that Kelly is mistaken. However it is Coyne who doesn't understand here. He doesn't understand the links between organic and cultural evolution. He doesn't understand evolutionary progress. He isn't in a good position to review Kelly's book because he doesn't understand the topics it deals with.

Sunday, 2 November 2014

Tim Tyler: Memetic programming

Transcript: Hi. I'm Tim Tyler and this is a video about memetic programming.

The term "memetic programming" is often used to refer to mental software.

In computer science there is a hardware software divide - which is based on how easy systems are to upgrade: hardware is difficult to upgrade while upgrading software is easy. Much the same classification scheme maps pretty neatly onto animal nervous systems. In particular, brains are hard to upgrade, while learned information is easy to upgrade. Socially learned phenomena are a lot like a type of cultural software. Culture is easily hot-swapped and modified much like software is. As with computer software, there are memetic programs written in memetic codes - and there are memetic programmers and deprogrammers.

While this usage of the term "memetic programming" is interesting, it isn't the topic of this video. Instead, this video is about the subset of memetic algorithms that are based on using a computer program of some kind as their genome.

To recap, genetic algorithms were used to solve optimisation problems, starting in the 1960s. In the 1990s, genetic programming was popularised by John Koza. Koza used a computer program as a genome and developed programs with a tree-like structure, in an attempt to make the crossover operation more useful. Around the same time, memetic algorithms were conceived. Where genetic algorithms attempted to mimic organic evolution, memetic algorithms drew inspiration from cultural evolution. They typically featured individual or social learning in combination with more conventional aspects of genetic algorithms.

Computer programs are an interesting and flexible target for evolutionary approaches to optimization problems. They typically help to provide a neat split between the genome, a developmental program and the resulting phenotype. The use of tree-shaped chromosomes has advantages in that recombination potentially becomes more wide-ranging. A tree-shaped genome makes it a bit harder to exchange functionally-corresponding modules between two trees - but this is a manageable problem.

One issue associated with most kinds of computer program is that they are often relatively poorly suited for parallel programming. You can fork execution threads - but this is a clumsy approach to parallel execution. Automatic parallelization is another possibility - but this also has serious limitations. Similar universal media which are more suitable for parallelization include neural networks, networks of logic gates and circuit board floorplanning. However attempts to make mutation and recombination-friendly versions of these have not been very successful so far. Consequently, these parallel representations tend to be treated as phenotypes - produced from computer programs by a developmental process - which brings us back to having a computer program as a genome again.

There are quite a few papers on memetic algorithms, genetic algorithms and genetic programming. However, memetic programming hasn't received much attention from academics yet. There are only a few papers on the topic so far. It ought to be a fertile mix of genetic programming and memetic algorithms.

Many picture a future in which teams of machine programmers collectively work on the source code of the next generation of machines. This is the essence of memetic programming. If these sorts of dynamics are going to be important in the future, then we should start studying them more seriously now.


The Princeton Guide to Evolution's section on cultural evolution is an ignorant mess

The chapter from the 2013 Princeton Guide to Evolution is online here. It's by Elizabeth Hannon and Cambridge philosopher - and meme hater - Tim Lewens.

I don't recommend that people get their information on cultural evolution or memetics from these folk - or this book. They don't know what they are talking about. Their chapter is a hatchet job on memetics - and a highly hostile account on cultural evolution. If many people are confused about cultural evolution, this sort of material is a good example of why.

The whole thing is so transparently a straw man attack that I can't imagine that many will be taken in by the contents. Maybe Elizabeth and Tim really do think that meme advocates are stupid. However, by doing so, I think they just make themselves look stupid. Step one for a critic is to try and find a sympathetic understanding of the material they are criticizing. The value of their efforts often depends heavily on how well they perform this step. Elizabeth and Tim have failed at this stage: they don't understand the material they are criticizing. The results are hopeless. It seems like a matter of embarrassment for the editors how this kind of material made it through the review process.

155 years of resistance to Darwinism within anthropology

Despite cultural evolution being close to the theoretical centre of gravity of their discipline, anthropologists have been some of the most reluctant scientists to embrace evolutionary theory.

Historically, Francis Boas and Claude Lévi-Strauss were among those responsible for the decline of Darwinism within anthropology. Retrospectively, their motivations seem political - often openly so. Boas saw it as his mission to expunge the concept of race from the scientific literature, and replace racial differences with cultural ones.

The document: The Boasian School of anthropology and the Decline of Darwinism in the Social Sciences tells the story of how this happened. It is embarrassing reading - portraying a triumph of ideology over science.

I notice that many of the modern critics of memetics are anthropologists. Their activities generally fit into the historical pattern of Darwin denialism in the social sciences. Even among those who have some sympathies for Darwin, we have documents minimizing the influence of Darwinism - such as How Darwinian is cultural evolution? - Claidière, Scott-Phillips and, Sperber. Their answer seems to be "not very". Boyd and Richerson have a different tack. They accept Darwinism, but take pains to emphasize how cultural evolution is different. My differences remain exaggerated article goes over this strategy. The basics of Darwinism are accepted, but the detailed implications - such as "Hamilton’s inclusive fitness rule" - are rejected.

Why do some anthropologists prefer "cultural variant" to "meme"? My theory is that "meme" is too openly biological - too reminiscent of "gene". Many anthropologists would react to this with a the same forceful immune reaction that they use to reject other biologically-influenced theories. "Cultural variant" is not such "in your face" biology - and so has a better chance of being tolerated by other anthropologists.

One the one hand, it is good to see a few anthropologists finally getting to grips with the application of evolutionary theory to culture. However progress within anthropology is frustratingly slow. The Darwinian revolution is going much more slowly in the cultural realm than it did in the organic realm.

I should add that it isn't just anthropologists who are at fault here. Evolutionists must accept some of the blame. For some reason, many evolutionists prefer to focus their educational efforts on other targets. Instead of educating their fellow scientists, many evolutionists seem to focus their outreach on the unlikely target of creationists. I am sceptical about whether this "gutter outreach" approach is an effective use of resources. I think a "leading from the front" approach would be better. Arguing with creationists makes evolutionists look stupid - in my opinion. It is too much like shooting a fish in a barrel. Evolutionists should spend more time taking on folk closer to their own intellectual size. Resistance to Darwinism within the social sciences is bad news for everyone - but few evolutionists seem interested in fixing the problem.

Part of the problem is specialization. Cultural evolution lies between two stools. Specialization is natural and good - but you have to make sure that you put more workers on the boundaries to make sure that valuable things don't fall down the cracks between the traditional disciplines.

Evolutionists need to clearly explain where their discipline applies to humans. Of course, that should be the job of anthropologists - but they have spent most of the last 155 years fumbling that ball, and still don't seem to have a good grip on it. Thus my equal time for cultural evolution proposal.

Saturday, 1 November 2014

Peter Richerson's olive branch

Recently Peter Richerson wrote:

I don’t have a big problem with the concept of memes so long as the meme-gene analogy is not excessively rigid. Susan assures is that Rob’s, Joe’s and my old fears in this regard are unfounded.

There is good and bad here. It is good if it means that one source of anti-meme FUD in academia will quit with the misguided criticisms based on an unsympathetic interpretation of memetics. I have become pretty fed up over the years with people citing Peter Richerson as a reason for not using the "meme" term. Peter may be an expert on cultural evolution, but his criticisms of memes were generally feeble and inaccurate. As I see it, this situation is his fault - since he is the aggressor. If he quits with the invalid criticisms of the efforts of his fellow researchers, hopefully we can all get along a bit better.

However, I see two negative points. Firstly, memetics is not based on an analogy. Instead we have the Darwinian algorithm playing itself out in multiple media.

As Susan Blackmore put it:

memetics is not based on analogy but on the principle of universal Darwinism: the idea that memes undergo the same evolutionary algorithm as genes.

Just so. Secondly, according to my favored classification scheme, memes are not just like genes, they are genes. Genes are the units of heredity in evolution. Genetics is the science that studies heredity. Memes are the units of heredity in cultural evolution - and so they are a subset of genes - just as cultural evolution is a subset of evolution.

Of course, this is an argument about terminology. However, according to a literal reading of his comment, Peter would have a "big problem" with the position that memes are not just like genes - they are genes.

I've though long and hard about this over the years, and the terminology I favor is the best that I can think of - given the history of the field. I don't see much of a "big problem" here. Indeed, if there is a "big problem" in the area, it is with those who try and use the terms "gene" and "genetics" for something else. The science of heredity and the unit of heredity are ideas that deserve to be taken seriously. Parochial treatments of these topics are not acceptable as science.