Thursday, 30 October 2014

Recent work on informational genetics

I've long been a proponent of informational genetics. This is the idea that genetics is based on information theory - as opposed to the (ridiculous) idea that genetics can be based on molecular biology.

The basic idea was pioneered by G. C. Williams (1966,1982), Richard Lewontin (1970), Richard Dawkins (1976, 1982) and others. To recap, information theory provides a theoretical foundation for a science of heredity - a generalised genetics that unites the organic and cultural realms and acts as a foundation for universal Darwinism.

I'm pleased to see that the topic has seen recent activity. Gérard Battail wrote two recent books on the topic:

Also, Steve Frank published a long and interesting paper on the topic in 2012, which is freely available online:

These recent contributions are encouraging. Critics have previously complained that proponents of applying information theory to biology lacked a systematic and rigorous approach to the topic (see references below for examples). This criticism was never very viable, but nonetheless, these recent contributions to informational genetics are welcome and encouraging.


Saturday, 25 October 2014

Natural Selection among the molecules

T. H. Huxley applied natural selection to molecules, in The Genealogy of Animals (1869):

It is a probable hypothesis that what the world is to organisms in general, each organism is to the molecules of which it is composed. Multitudes of these having diverse tendencies, are competing with one another for opportunity to exist and multiply; and the organism, as a whole, is as much the product of the molecules which are victorious as the Fauna, or Flora, of a country is the product of the victorious organic beings in it.

Darwin wrote back on October 14 1869:

I am very glad that you have been bold enough to give your idea about Natural Selection among the molecules, though I cannot quite follow you

This early exchange makes T. H. Huxley an early pioneer of universal Darwinism.

The details were not all fully understood at the time - but we now know that molecules are the product of an extended evolutionary history - which includes splitting and merging events as well as iterated selection. Splitting was always a fundamental feature of Darwinism - but it really wasn't until the symbiology revolution of the 1960s-1980s that the significance of mergers to evolution was understood. Iterated selection is a central characteristic of Darwinism. Molecules exhibit this extensively - with every molecule the survivior of a long series of potentially lethal selection events. The stability of observed molecules can be usefully understood as a form of Darwinian adaptation. Unstable molecules are possible an theory - and are sometimes observed fleetingly - but the law of survival of the stable means that most of them are short-lived.

Tuesday, 21 October 2014

Memetic sex

Memes like genes clearly recombine. It is a simple matter to point to what were once independent memes fused together. Portmanteaus are an obvious example. For example:

  • "Sheeple" - a portmanteau of sheep and people;
  • "Spork" - a portmanteau of spoon and fork;

It is fairly common to use the term "sex" as a synonym for "recombination". For example, this is what Matt Ridley seems to do - when he says that ideas have sex.

However, not everyone shares this broad conception of sex. For one thing, you can recombine your own genes with each other - for example as an error-correction strategy, but - unless the process involves outcrossing - few would classify this as being sex. Another issue is pathogens. A retrovirus that infects your cells is certainly recombining its genes with yours - yet many people would be reluctant to call refer to this form of recombination as being "sex".

This leads to the question of how best to distinguish between sex and disease. One apparent difference is that sex involves members of the same species - whereas disease involves members of different species. However the most common definition of what a species is invokes the idea of sexual reproduction. To simultaneously define "species" in terms of "sex" and "sex" in terms of "species" would be circular.

Other differences might be invoked: parasites have shorter generation times, smaller bodies and smaller genomes than their hosts. However, most of these differences apply to genuine males somewhere or another in biology. They seem unsuitable to act as a basis for classification.

My view is that the best way to distinguish sex from disease is to see whether the recipient benefits from its injection of genes. If it willingly accepts the genetic donation, that's sex. If it rejects them - or tries to - that's disease. This raises some corner cases - such as cases of rape. These cases suggest a slightly different criterion - it's sex if there are adaptations favouring the incorporation of the injected genes - and it's disease if adaptations resist them.

I wrote up this idea in 2007 - in an essay titled: "Sex is not a disease" - and I think the idea has stood up well to the test of time.

Having said what we mean by sex, it is now possible to address the issue of sex in cultural evolution. The portmanteaus we originally gave as examples of recombination don't look much like sex or disease - they are one-off events. However, there are some cases of cultural evolution that seem more sexual:

  • Fashion products - Fashion generates enormous cultural diversity. If you just think about shoes, the number of varieties on sale at any time is enormous - and cultural sexual recombiantion seems like a tempting explanation. Here, recombination is ritualised into annual fashion shows - where the designers go to copy each other's products.
  • Printer-cartridge diversity - in my 2011 memetics book I gave printer-cartridge diversity as an example of parasite driven recombination. The parasites in this case are the cartridge-cloning companies. Parasites drive a lot of sexual recombination in the organic realm - and it looks as though similar diversity is produced in cultural evolution in response to parasitism.
  • Experimentation on consumers - some internet companies (such as Google) run regular experiments on their users. They change the colours scheme, the positioning of adverts - and so on. At any time people may be participating in many different experiments - and multiple experiments are run simultaneously - to gather data about their interactions. The dynamics here involve a kind of ritualized recombination.
  • Genetic algorithms and memetic algorithms - anyone who denies the existence of cultural sex surely faces a strong challenge in explaining how these optimisation techniques work without invoking the concept of cultural sex.
Even with the relatively strict definition of "sex" given here, it seems as though cultural evolution exhibits sexual recombination. I am not really clear on what Richard Dawkins meant in 2005 - when he said (of memes):

One respect in which they are not like genes is that there is nothing obviously corresponding to chromosomes or loci or alleles or sexual recombination.
Memes do have sexual recombination. I don't think there is room for much argument about that. If your conception of sex doesn't apply to cultural evolution, you probably don't have a very useful conception of it.

Sex has been described as being "nature's masterpiece". It is true that there are quite a lot of monocultures in cultural evolution. Many products for example, occur in many identical copies. Microsoft's Windows is a monoculture. However like many monocultures it has been plagued by pathogens.

However, nonetheless, there is sex in cultural evolution. Even a fairly cursory look at the fashion world shows the impact of regular sexual recombination on product diversity.

In the organic realm, sex often leads to gamete size dimorphism. It is possible to see this in companies - with some large, fat organizations provision resources, while other small, meme-carrying mobile agents of various types flit between them. There's also an equivalent of female choice. Interviews, testing by R+D departments, and carefully scrutinized fundraising pitches all qualify here.

Sex has proved popular in the organic realm. No doubt it will become similarly popular in cultural evolution. After all, regular recombination with outcrossing is the best way that is known to explore large, complex search spaces.

Sunday, 19 October 2014

Equal time for cultural evolution

Cultural evolution is a neglected runt-child of the scientific establishment. This is - as far as I can tell - the result of ignorance and stupidity. The main culprits are anthropologists. They are the people who study this material in their day jobs. However evolutionists must also accept some of the blame.

For humans, it is often a lot more important and significant to understand cultural evolution than it is to understand organic evolution. Yet most educational efforts devoted to evolution concentrate on organic evolution. Most scientific papers concentrate on organic evolution. The whole topic is insanely biased away from cultural transmission. What people learn in school is blinkered Darwinism. Some of these people grow up into educators and pass their blinkered Darwinism on. The wheel of ignorance rolls down the generations.

To redress this imbalance, I have a modest proposal. I think those teaching evolution - and those studying evolution - should try and split their time equally between the organic and cultural realms. Because we are human beings - and because cultural evolution is so significant for humans, it would actually make sense to have a lot more than 50% of the examples in the cultural realm. Also, there is currently an insane bias away from cultural evolution. To redress this bias, we really need to devote as much energy to cultural evolution as we can. That is what makes this proposal a modest one: I am only asking for 50%.

I think the domain of this proposal should include publicly funded evolutionary science and those in teaching roles relating to evolutionary theory.

As with any such proposal I think there should be consequences for non-compliance. I think these should involve bad reviews, questioning of competence and other types of criticism. Nearer to the bottom of the barrel, there's poking fun and ridicule. The kind of thing we saw in Jerry Coyne on Cultural Evolution: What Does He Know? Looking at the current landscape, these strategies will probably prove necessary. Cultural evolution has been neglected for far too long. It is time to put a stop to the ignorance.

Thursday, 16 October 2014

Horizontal meme transfer

The term "horizontal gene transfer" refers to the transfer of genes between organisms a manner other than the transfer between parent and offspring that takes place during normal reproduction.

Such transfer normally involves a symbiotic relationship of some kind:

  • Parasitism and mutualism;
  • Direct injection of genes;
  • Sexual recombination;
Sexual recombination is not normally listed as a form of horizontal gene transfer - on the grounds that it is a case of ordinary reproduction. This seems like a misclassification to me: sexual recombination involves one organism injecting its genes into another one that is not its direct offspring. Like other forms of HGT, it creates a reticulated evolutionary tree.

Direct injection of genes is common among bacteria - where it is called bacterial conjugation.

In symbiology, when gene transfer is vertical (i.e. symbiont and host generations are synchronized) the interests of symbionts and hosts tend to become aligned - what is good for the host is good for the symbiont. By contrast, when gene transfer is horizontal the interests of symbionts and hosts are not aligned. This can result in stress and problems for the host. One classic (and topical) example of this is the ebola parasite. This cares little about the welfare of its host, and treats it as a reservoir of resources to be turned into more parasites as soon as possible - often killing the host in the process.

Horizontal gene transfer produces a "reticulated" (networked, web-like) phylogeny - rather than a classic tree-like shape.

Horizontal meme transfer works in a similar way to horizontal gene transfer. The symbionts involved are cultural, and their heritable material is memes. The hosts involved are often human beings - but they can also be computer systems or organizations. "Horizontal" is normally intended to be with respect to the lineages of the DNA genes of the human hosts.

Sometimes it is claimed that cultural evolution features more horizontal transfer than than is seen in the organic realm. However, actual studies of the topic tend to show that cultural evolution and DNA-based evolution are roughly similarly reticulated. Alex Mesoudi devoted a section to this topic in his 2011 book. It starts on page 90 and is titled: "But is cultural evolution treelike". He reports on work comparing the "retention index" of 21 phylogenetic trees and 21 phylomemetic trees. The cultural trees had a retention index of 0.59, whole the organic trees managed 0.61. His citation for this work is: Collard, M; Shennan, S.J.; Tehrani, J.J.; (2006) Branching versus blending in macroscale cultural evolution: A comparative study.

Another surprisingly-common claim is that cultural evolution features horizontal and oblique meme transfer - while the organic realm, gene transfer is always vertical. For examples of this claim, see here and here. This claim is patently false.

We are currently seeing a shift from vertical meme transfer (as seen in many traditional religions) to horizontal meme transfer. Human population densities are increasing, and additionally, new meme-transfer routes are appearing - in the form of wires, optical cables, and radio waves. Consequently meme interests and the interests of the host DNA genes are likely to become increasingly different - according to standard evolutionary theory.

Wednesday, 15 October 2014

Broward Horne's contributions to memetics

I haven't yet linked to Broward Horne's contributions to memetics. Here are three memetics-related videos of his:

Tuesday, 14 October 2014

Michele Coscia: Average is Boring: How Similarity Kills a Meme's Success

Michele also has a podcast on the importance of memetics here.

Update 2015-02-21: ArsTechnica coverage of Michele Coscia's work here: Why memes succeed.

Saturday, 11 October 2014

The expansion of the domain of Darwinian evolutionary theory

The last 50 years have seen a massive expansion of the domain of Darwinian evolutionary theory.

Though the idea that Darwin's ideas could be usefully applied to subject areas beyond the evolution of species over geological timescales has been around ever since Darwin's era, it wasn't until the 1970s that systematic exploration of the area began.

Karl Popper applied Darwinism to science. B.F. Skinner applied Darwinism to individual learning. Donald Campbell applied it to social learning. In 1976, Richard Dawkins devoted a chapter in "The Selfish Gene" to the idea that culture evolved.

Then things really began to get going. In the 1980s several serious books by scientists on the topic of cultural evolution appeared. Gerald Edelman discovered Darwinian evolution in the immune system - and in low-level brain structures - such as neurons and synapses. William Calvin also did pioneering work on Darwinian evolution in the brain.

Darwinian evolutionary theory was broadly applied to social sciences, with economics - often regarded as the most "scientific" of the social sciences - generally leading the way. Not things like evolutionary psychology - which is still mostly obsessed with DNA genes - but rather Darwinian theories based on cultural variation.

It now seems fairly clear that the basic concepts of Darwinian evolutionary theory, such as adaptation and drift, apply widely including to many physical and chemical systems outside of biology. There's some debate on whether we should call it "Darwinism" - partly because of the scale of the developments since Darwin's era. However, this is a relatively minor terminological skirmish; the point is that Darwin's basic principles apply.

Self-organizing systems were once seen as a challenge to Darwinism from physics. For example, the forms of tree branches and roots do indeed resemble the patterns made by lightning strikes and fractal drainage basins - and some speculated that the same physical principles were responsible for both. However an examination of how these systems are formed shows that the Darwinian principles of descent with modification and extinction is involved in lightning strikes and fractal drainage basins. Fractal drainage patterns adaptively fit the drainage basins that contain them - due to Darwinian adaptation. What started off as looking like an invasion of biology from physics has turned into a battle with territory being lost and gained on both sides.

Darwinian evolutionary theory also applies to observation selection effects - which are an important area of science. This is another case of Darwin invading territory which was once thought to belong to physics.

We also understand why Darwinian evolutionary theory is so broadly applicable. Darwinian evolutionary theory is based on the production of new forms from old ones via copying and the elimination of unfit ones. It depends on mutations being small. A theory with no restraints on the type of mutation is consistent with any sort of observation, and so is useless. We can see that copying is ubiquitous in nature. The observed fact that physics is simple and local works to keep mutation small. There's also a law of conservation of information. Information never comes out of nowhere - it always comes from somewhere else. Under these circumstances, an evolutionary theory along the general lines that Darwin described is practically bound to apply. Design can't come out of nowhere - because the universe is closed. Instead it evolves gradually.

Once the idea that the domain of Darwinian evolutionary theory was expanding dramatically took root in the minds of scientists, some questions developed:

  • How can we best characterise the general theory that applies to all these domains?
  • What are the limits of evolutionary theory? i.e. when does it turn into dynamical systems theory?
To some extent, these questions are still being worked on by scientists. However the expansion of Darwinian evolutionary theory is a big deal. Expansion of the domain of the theory has brought growth, new developments and refinements. We can see the evolutionary theory more clearly through having more varied data from a range of different processes - based on different kinds of heritable material. It is probably the biggest revolution that Darwinian evolutionary theory has seen to date.

A recent paper in Nature describes this revolution. It's written by a bunch of developmental biologists - who use their own terminology and have their own perspective on the issue. It also features some rather embarrassing straw man attacks on what they call "Standard Evolutionary Theory". However it is possible to make out that some of their points are a watered-down version of the one I am describing here. My coverage of that paper is here.


Friday, 10 October 2014

The evolution revolution makes Nature

The evolution revolution makes Nature:

  • Does evolutionary theory need a rethink? by Kevin Laland, Tobias Uller, Marc Feldman, Kim Sterelny, Gerd B. Müller, Armin Moczek, Eva Jablonka, John Odling-Smee, Gregory A. Wray, Hopi E. Hoekstra, Douglas J. Futuyma, Richard E. Lenski, Trudy F. C. Mackay, Dolph Schluter & Joan E. Strassmann.

I'm with the "yays". However, Their revolution isn't necessarily mine. They say:

In our view, this ‘gene-centric’ focus fails to capture the full gamut of processes that direct evolution. Missing pieces include how physical development influences the generation of variation (developmental bias); how the environment directly shapes organisms’ traits (plasticity); how organisms modify environments (niche construction); and how organisms transmit more than genes across generations (extra-genetic inheritance).
Their main points seem to be:

  • S.E.T. (Standard Evolutionary Theory) neglects development. Meh. Not really. I don't see any especially interesting revolution here.

  • There is more to inheritance than genes. Meh. Only if you define "genes" as being bits of DNA. However, genes are not bits of DNA. Genes are the units of genetics, and genetics is the study of heredity. Informational genetics neatly covers DNA, memes, environmental inheritance, and works fine as a basis for Darwinsm. If you define "gene" in this way (which you should!) "the idea there is more to inheritance than genes" is wrong. Genes are the units of heredity - by definition. If you use the "molecular biology" definition of "gene" then yes: a big revolution is needed. You need to replace genetics with whole new science. However, "genes" and "genetics" will do just fine as foundational concepts - if you understand these ideas correctly.

For me, the biggest revolution is associated with the radical expansion of the domain of evolutionary theory that has taken place in modern times. The dynamics of the new domains were a bit different - and evolutionary theory has been forced to evolve to encompass the new cases. However, phrasing this revolution as "there is more to inheritance than genes" adopts a particular terminological position - and it is one that I reject. "Gene" and "genetics" are fine and perfectly general terms. There is absolutely no need to have a revolution that replaces them. They just need to adapt a little. People need to understand that memes are genes too.

I think this expansion of the domain of Darwinism constitutes a revolution. I - and many others - have described it as a paradigm shift. It seems to obviously qualify as a revolution - because of the enormous number of players who so obviously don't grasp it. I think it is fairly clear that the naysayers in this paper haven't grasped the nature of this revolution. For example, they say:

Likewise, there is little evidence for the role of inherited epigenetic modification (part of what was termed ‘inclusive inheritance’) in adaptation: we know of no case in which a new trait has been shown to have a strictly epigenetic basis divorced from gene sequence.

That's an odd thing to say. Of course there are many cases where new traits are divorced from DNA sequences! Practically any culturally-transmitted trait fits the bill. Tattoos, piercings and surnames seem pretty "divorced from gene sequences" to me. What are these biologists thinking about? From their comments it seems clear that they just don't understand. They are on the wrong side of the paradigm shift.

They also question the "traits before genes" model, saying:

Much less clear is whether plasticity can ‘lead’ genetic variation during adaptation. More than half a century ago, developmental biologist Conrad Waddington described a process that he called genetic assimilation. Here, new mutations can sometimes convert a plastic trait into one that develops even without the specific environmental condition that originally induced it. Few cases have been documented outside of the laboratory, however. Whether this is owing to a lack of serious attention or whether it reflects a genuine rarity in nature can be answered only by further study.
Hang on: this is a well-known phenomenon. Our ancestors drank milk before they had lactase genes. They walked before they had locking knee joints. They threw rocks before they could aim them well. They used baby slings before their babies were so big and helpless. They swam before they had lost their fur - and so on. "Further study" - pffft: that's what these authors need to do.

As for going beyond the gene - the most popular attempt to replace the "gene" and "genetics" terminology stemmed from the work of Richard Dawkins. He proposed that DNA genes were a special case of an entity he described as being a "replicator" - an attempt to shift away from "gene" and "genetics" and into a more abstract, information-theoretic space. This was a sensible idea - but the terminology and execution were flawed. I've watched this proposed revolution in terminology trundle along over the decades - and eventually I rejected it. We have a science of heredity - genetics - and we don't need another one. We have a unit of heredity - the gene - and we don't need another one. Cultural inheritance can be encompassed within a generalised science of genetics.

The proposed "replicator" revolution hasn't gone especially well - for various reasons. It is still chugging along in some quarters. I don't foresee an especially favourable endgame for it.

Blogosphere commentary: Razib Khan | Larry Moran | Philip Ball | Reddit | Jarry Coyne

Thursday, 9 October 2014

Eusociality: the symbiont hypothesis - CFP

This is a call for programmers interested in the symbiont hypothesis of social evolution.

I've covered the symbiont hypothesis of eusociality before - e.g. see my 2011 book - or:

This is a big and important hypothesis relating to the origin of social organization and eusociality. However, it seems to have become absurdly neglected in modern times - where kin selection has got almost all of the limelight.

The field urgently needs more study. It needs computer simulation. In particular, I am thinking about simple agent-based models, or cellular automata - that illustrate its basic idea - that the introduction of symbionts can promote social behaviour. In a termite-like model - where the symbionts directly benefit the hosts, it is pretty obvious that this will happen. However there are an abundance of open questions in this area - associated with how model parameters affect the resulting evolutionary dynamics. Science needs a lot of models in this general area to resolve these questions. So far, as far as I can tell, very few of these models have been constructed.

A basic model might feature hosts, symbionts and walls - creating an environment with cave-like enclosures. Step 1 would be to find some parameter settings where the introduction of symbionts favoured cave-dwelling hosts that interacted with each other more frequently - more effectively spreading the symbionts to newborns.

If you are a programmer with a scientific bent, you can help to resolve the questions associated with these models. If you do a reasonable job, you will probably go down in history in the process for doing so.

This is work of large social and political significance. Humans have cultural symbionts, which make them cooperate. The cultural symbionts make the large difference between modern humans and primitive cavemen. However, because of low levels of scientific study of the symbiont hypothesis of eusociality, science still has a relatively poor understanding of exactly how and why they do this. The study of cooperation has been an active topic historically - partly as a result of this social and political significance. However the symbiont hypothesis of eusociality has been enormously neglected. I think most workers in the field don't understand it - or its importance.

Partner selection

Darwin famously originated the concept of "sexual selection". His idea at the time was that mate choice is an important factor in evolution.

Symbiolgy usefully generalizes this. Mating partners are only one of many types of partner in symbiotic relationships.

Sometimes, mutualists pick their partners, predators pick their prey and parasites pick their victims.

These processes are well-characterized as being forms of partner selection.

Paralleling "female choice" there are also "predator choice" and "parasite choice". Mates are selected - but also, so are hosts, prey, food, trading partners, assistants, targets of mimicry, bodyguards, slaves and tennents.

Sexual selection was famously involved in introducing minds into the evolutionary process - making a lie out of the idea that natural selection is "blind". However sexual selection rather unfairly gets all the attention - while other forms of partner selection are neglected - apparently through not having such a distinguished founder and such a catchy name.

Predation and parasitism likely existed before sexual recombination - giving these forms of partner selection a more primitive status. Their toll in modern times is impressive: far more organisms are lost to predation and parasitism than are neglected in the competition for mates. The term "prey selection" has seen some use. However, an umbrella term is also clearly needed. Evolutionists should make more use of the idea of "partner selection".

Natural selection is not an algorithm

The modern use of the term "natural selection" has some issues, from my perspective. The problems I see are:

  • Natural selection crudely combines the production and elimination aspects of Darwinism;
  • Natural selection is defined as being "non-random" change;
  • Natural selection is widely and frequently characterized as being an "algorithm";
To summarize the first objection, I think it is better to teach evolution by separating its productive and eliminative aspects. I've gone into this many times before, in particular:

This split seems like a simple matter of good epistemic hygene to me.

My second objection is about philosophical difficulties with the concept of randomness. I've gone into that in detail in:

My third objection is the topic of this article. Starting with Lewontin (1970), the term "natural selection" has been frequently characterised as being an "algorithm". Of course, there is a kind of abstract algorithm associated with Darwinism - involving the iterative application of both production and elimination. The issue described here is concerned with using the term "natural selection" to characterize this algorithm.

The term "natural selection" suggests that nature is choosing something. We can imagine nature choosing which organisms live, which die, which organisms are selected as mates, and which organisms are selected as victims. However all these usages omit the production of variation.

The Darwinian algorithm includes the production of variation. This is mutation. I think that mutation should be a different category from selection. It is possible - at a stretch - to characterize mutation as a form of choice: nature "chooses" which mutation to use. However, the biggest problem with combining them is that this conceptually muddles together two very different things.

IMO, even those who swallow the first two items I am objecting about should reject using the term "natural selection" to describe the Darwinian algorithm. It includes something else besides selection - namely the production of variation. Variation and selection should be conceptually separated out. If you are going to use "natural selection" to describe an algorithm involving selection and mutation, you've polluted the term for selection alone.

Wednesday, 8 October 2014

Darwinism detractors

This post is about objections to the term "Darwinism". It makes no attempt to cover objections to Darwinian theory.

One of the advocates of describing evolutionary change as being "Non-Darwinian" is Lianne Gabora. However, check out what she means by "Darwinian":

To make writings on these matters less awkward, the terms ‘Darwinian’ and ‘selectionist’ are used as a shorthand for ‘by means of natural selection or a process that is algorithmically equivalent to it’. It would be wrong to interpret this as implying that Darwin never gave thought to evolution by means other than natural selection. He was, of course, immersed in the views of his day, and considered several possible explanations for adaptive change (e.g., his ultimately unsuccessful theory of gemmules). Moreover, although Darwin did not use the term neutral evolution, he acknowledged that evolutionary change can involve fixation of variants that confer no selective advantage over previous adaptations; nevertheless, neutral evolution is commonly referred to as non-Darwinian [70,75,123]. Similarly, although Darwin was not committed to the idea that all life evolved from a single common ancestor, processes such as horizontal gene transfer (that is, genes transmitted between organisms in a manner other than through traditional reproduction) are commonly referred to as non-Darwinian [131].
With this definition of Darwinism, we are all non-Darwinians - including Darwin himself. Larry Moran is another advocate of this definition of "Darwinism". I am pretty sure that this is not a good way to define the term "Darwinism". This is straw man Darwinism. Only critics define Darwinism in this way: no advocates do so. Why is this bad? Critics have a weaker claim on defining terminology in a field than advocates do. They should mostly accept the terminology of advocates when criticising. If critics get to make up too much terminology, we get a polluted namespace, full of too many daft and useless terms.

Another complaint about "Darwinism" is that the term is used pejoratively by creationists:

Evolutionary biology owes much to Charles Darwin, whose discussions of common descent and natural selection provide the foundations of the discipline. But evolutionary biology has expanded well beyond its foundations to encompass many theories and concepts unknown in the 19th century. The term “Darwinism” is, therefore, ambiguous and misleading. Compounding the problem of “Darwinism” is the hijacking of the term by creationists to portray evolution as a dangerous ideology—an “ism”—that has no place in the science classroom. When scientists and teachers use “Darwinism” as synonymous with evolutionary biology, it reinforces such a misleading portrayal and hinders efforts to present the scientific standing of evolution accurately. Accordingly, the term “Darwinism” should be abandoned as a synonym for evolutionary biology.
I don't see why scientists should adopt terminology based on a bunch of idiocy about theology and god. I'm regularly told that I can't use the term "intelligent design" - because that term has already been taken - by a bunch of idiots. As you can probably imagine, I am not impressed by this: idiots don't get to decide what terminology scientists use.

"Darwinism" might not be an ideal term. However, it seems to me that it has fairly comprehensively trounced its competitors in the struggle for survival of words. In terms of volume, even neo-Darwinism fails to compete. The "don't call it Darwinism" paper fails to come up with a decent alternative. The nearest thing it manages to offer is "evolutionary biology". This seems to fail to me. "evolutionary biology" is an application domain. Darwinism is a refutable theory. Those things are not really in the same category. Also, "evolutionary biology" is a topic confined to biology. When I use the terms "Darwinism" and "Darwinian evolution" I am talking about something that also covers physics and chemistry - and is emphatically not confined to biology.

"Darwinism" has been a pretty spectacularly successful meme. I think we should make the most of it.

Sunday, 5 October 2014

Recent introductions to memes

A couple of introductions to memes were published recently:

Thanks to Brian and Dan for their efforts in this area. Brian's effort is longer and contains more controversial content. He offers a historical perspective on the development of memetics. One of the things he says is:

A “selfish meme” interpretation of a cultural universal, however, implies that it is maladaptive (harming fitness).

Of course selfish memes can and do harm fitness associated with their host's DNA. However, that isn't really part of the definition what "selfishness" is all about in this context. The idea is that selfish memes are in it for themselves. They behave as though they are self-interested - and as though they are indifferent to the fate of their hosts. That doesn't mean that they harm their hosts. For example, strawberry genes can be described as being "selfish". They (behave as though they) care about making copies of themselves - and not about any humans who might eat them. Does this "selfishness" on the part of the strawberry genes mean that they are bad for humans? No. In fact, strawberries are a health food.

Dan Zarella's article is also good. I'm still deeply critical of Dan's assertion that his "R0" differs significantly between cultural and organic epidemics, though. Both genes and memes can spread explosively and both genes and memes can reach saturation levels in their host population. What Dan is saying here is just wrong.

Dan says:

I’ve never found a meme with a sustained R0 above one. Given a large enough population and a long enough time, the R0 of every idea falls below one and the idea stops spreading.
This is not the normal meaning of r0. Here's how Wikipedia puts it:

In epidemiology, the basic reproduction number (sometimes called basic reproductive rate, basic reproductive ratio and denoted R0, r nought) of an infection can be thought of as the number of cases one case generates on average over the course of its infectious period, in an otherwise uninfected population.

Here r0 is a constant value. It doesn't change over time. That is because r0 is defined with respect to an otherwise uninfected population.

Friday, 3 October 2014

The curious idea that selection is Darwinian while mutations are not

I reviewed the discussion of what makes a process "Darwinian" in the paper:

Cultural transmission and the evolution of human behaviour: a general approach based on the Price equation

The authors advocate using the Price equation to divide evolutionary change into selection and transmission bias components. They then claim that a high selection component makes a process more "Darwinian" and a high transmission bias component makes it less "Darwinian".

I think that this is a pretty strange approach. Darwinian evolution depends heavily on both selection and mutation. One without the other generally results in a short-lived evolutionary process. For me, selection and mutation are both core components of Darwinism.

The mutation rate is highly variable in the organic realm. For example, around Chernobyl, the mutation rate was massively elevated in the 1980s. Does this mean that organic evolution suddenly became non-Darwinian in Chernobyl? I would say "no" - harsh environments are perfectly compatible with Darwinism. However, this paper's authors seem to think otherwise.

Worse for this proposed classification scheme, the split between selection and transmission bias depends on how you partition the population into generations before applying the Price equation. High levels of selection at one temporal scale (e.g. parasite lifecycle) gnerally look like transmission bias at a longer temporal scale (e.g. host lifecycle).

The authors give an example of what they see as a cultural process heavy in transmission bias:

At the other end of the continuum is the case where the transmission component is large and the selection component negligible. Such a case would be where individuals were exposed to ideas, but deliberately or subconsciously modified them through their cognition to such an extent that the trait they exhibit bore little resemblance to the traits of the previous generation. Changes in trait frequency would then be best explained with reference to the (genetically evolved) transformative properties of the human mind, not cultural natural selection. Such a situation does not possess the usual paradigmatic features of a Darwinian process
The evolution of ideas within minds is not Darwinian?!? Many neuroscientists would disagree. See Keeping Darwin in Mind for the details.

The problem here is that the authors have chosen a high level of partitioning before applying the Price equation - namely, social transmission between minds. High level partitioning favours the "transmission bias" component of the Price equation. However a more appropriate level for studying the Darwinian evolution of ideas would be to see how ideas are copied within minds. Or maybe how neural spikes are copied as axons split. If you partitioned on these scales, you would see less transmission bias, and more selection.

If you want to base a classification scheme on the Price equation, you need to decide how to divide the entities you are studying into discrete generations. Unfortunately, most real-world evolutionary processes have copying on multiple different scales - and so the are many ways of doing this and it matters which one you choose. Unless you keep the subjective aspect of the partitioning scheme you are using in mind, using the Price equation in this way will just result in reflections of your preconceptions.

Thursday, 2 October 2014

Wednesday, 1 October 2014

Darwin's razor

Occam's razor is the principle that one should prefer the simplest explanation that fits the data. It is one of the foundations of the scientific method. Occam's razor has turned out to be a flawed way of managing hypotheses - but it's still not too bad an idea.

Occam's razor is a special case of a more general principle: that the least useful may be profitably discarded. This idea is associated with evolution via natural selection. In honor of Darwin, I propose that we call it "Darwin's razor".

Definition of "Darwin's razor": the least useful is best discarded.

Occam famously discarded all but the shortest hypothesis. However, from a more Darwinian perspective, that's often a bad idea. For one thing you lose a diverse breeding population of hypotheses. For another, it is harder to adapt if new observations no longer fit the previously-selected hypothesis. Lastly, if you only consider the shortest hypothesis, it turns out that you can't estimate probabilities correctly.

"Darwin's razor" represents the winnowing, destructive side of Darwin's theory. I've previously referred to this as being "natural elimination". Of course, Darwinism also has a creative side: the production of new forms. However, that is hardly well described as being like the operation of a razor. "Darwin's razor" is all about eliminating the old, the unwanted and the unfit.

Occam's razor famously had only one blade. Darwin's razor is more like one of those modern multi-blade razors. This analogy works in two ways:

  • Occam's razor said to only keep one hypothesis, while Darwin's razor allows you to keep a breeding population of hypotheses;
  • Occam's razor only applied to one thing: hypotheses, while Darwin's razor is a universal razor that can cut through practically anything.
To give some examples of the universality of Darwin's razor in different domains:

  • Wallet full of small coins? Apply Darwin's razor.
  • Too much spam? Darwin's razor can help.
  • Fridge full of useless jars? Darwin's razor works there too.
  • Garden full of things you never planted? Use Darwin's razor.
Also, using Darwin's razor regularly makes you irresistibly attractive to members of the opposite sex.

Occam's razor has turned out to be a somewhat-inaccurate approach. However, Darwin's razor does a much better job of surviving critical scrutiny. It isn't possible to keep all unrefuted hypotheses around (as Solomonoff induction might suggest) because there isn't enough space in the universe. You have to discard something. Darwin's razor recommends discarding the least useful stuff. It's simple and pragmatic advice that's hard to disagree with. Darwin's razor.

The cultural Cambrian explosion

The Cambrian explosion has become synonymous with the idea that evolution is characterized by periods of explosive generation of diversity - followed by shake-outs - in which that diversity is winnowed.

I think we can see examples of this phenomenon in cultural evolution.

Economists are used to the idea that marketplaces regularly see periods of adaptive radiation followed by shake-outs. This is part of the idea of a boom-bust cycle.

Human languages seem to be past their "explosion" stage - and we are now in the "shake-out" stage - when diversity decreases.

By contrast, computer languages are still in the "explosion" stage.

TV channels are also still in their "explosion" stage.

Indeed, for most categories of memes you care to think of, we are currently seeing explosive growth. We are living through the cultural Cambrian era. Some explosions have clearly not started yet. We haven't yet seen much of an explosion of fusion reactor designs, or an explosion of machine intelligence. An explosion of machine intelligence is widely forecasted, though. Rodney Brooks titled his 1999 book "Cambrian Intelligence". More recently we have Dag Kittlaus claiming that A Cambrian Explosion In AI Is Coming It's not hard to believe.

What about the shake-out phase? We live in a world with a small number of search engines, a small number of social networks and a small number of popular operating systems. There's mostly only one internet. In some "winner-takes all" markets, competitors inflate - squeezing out the smaller players.

Biologists have long hoped for alien life to study - so that they can see what principles of biology are fundamental and which are accidents of history. Cultural evolution largely fulfills that dream - by providing new forms of life for biologists to study which are not based on DNA. In the case of the Cambrian explosion, speculation about multi-million year old events can be usefully supplemented by modern systems undergoing similar adaptive radiations followed by competitive winnowing. Evolutionary biology is likely to grow stronger and more mature by assimilating these new examples of the dynamics of evolving systems.

The most common misunderstanding of cultural evolution

A diagram has emerged in the coverage of the S.F.I. workshop that Daniel Dennett organized earlier this year. I think it neatly illustrates one of the most common misunderstandings about cultural evolution. First, here is the diagram:

The main problems here are not with the "cultural evolution" column. It appears that people understand cultural evolution! The problems are all in the "biological evolution" column. Every entry is wrong:

  • It is not true that - aside from culture - traits are only inherited from parents. Alice might have gotten her flea bites from her daughter, her warts from her grandmother and her chlamydia from her second cousin.
  • It is not true that transmission which is not down host generations happens in cultural realm and not in the organic realm. DNA-based traits can be inherited from offspring, as happened with Alice in the last example.
  • The speed of cultural evolution with respect to the organic realm is much exaggerated. If you look at areas where the generation time is comparable - such as medicine vs bacteria and viruses, the ideas simply aren't outstripping the pathogens in terms of the speed of their evolution. When people argue that cultural evolution is fast, they are mostly impressed by its short generation time and the large number of ideas. However, bacteria and viruses can reproduce quickly too. Plus they are far more numerous than ideas (and so they can perform many more experiments). Perhaps one day cultural evolution will unambiguously go faster.
  • Traits acquired during a human lifetime can be transmitted genetically. Fleas can be acquired during a human lifetime and are propagated genetically; AIDS can be acquired during a human lifetime and is propagated genetically - and so on.
  • People can choose which genetic traits they inherit. If you choose to never have sex, you won't inherit any syphilis genes, or any chlamydia genes. It's not true that you don't have a choice.
  • DNA is not the only medium of inheritance - aside from culture. Organisms inherit resource boluses from their parents. They inherit accommodation - as seen with rabbit warrens, beaver dams and dental carries. They inherit their location. The idea that DNA is the only medium of inheritance is wrong.
  • Lastly, even the column headings in this table are wrong. Culture is part of biology. Contrasting "cultural evolution" and "biological evolution" makes no sense.

If contrasting "cultural evolution" with "biological evolution" [sic] it is important not to use an out-dated characature of biological evolution that predates our understanding of the significance of symbiosis. If you mistkenly do that, you'll get the relationship between them all wrong.

I don't know who created this table. Anyway, they have helped to highlight a common misconception - so thanks are due.

Update 2014-10-15: I also saw this table on the "bioninja" web site. I suspect that is the original source.