Sunday 25 November 2012

Tim Tyler: Harms, Information and Meaning in Evolutionary Processes (review)

Transcript:

Hi. I'm Tim Tyler, and this is a review of this book:

Information and Meaning in Evolutionary Processes by William F. Harms
This book starts out with a 80-page critique of "replicator theories" - a term the author uses to cover the cultural evolution theories of Dawkins, Dennett, Blackmore, Hull and various other players. However, memes seem to attract most of the fire. We hear about how Dawkins backtracked apologetically after introducing memes, playing down their significance. How David Hull was only interested in memes to the extent that they helped him develop a scientific epistemology, and how Dennett got his memes second hand, and just wanted to use them to bolster his concept of the "intentional stance". Memes are based partly on G. C. Williams attempt to rechristen the gene. We hear that this rechristening never caught on, and the word "gene" today still has a totally different meaning in biology textbooks, leaving memes dependent of a dead definition. Further the definition of "gene" that Williams used makes little sense - since it defined genes in terms of selection pressures, which might fluctuate wildly in real life, causing genes and memes to flit in and out of existence. On page 67, Harms writes:

The reader cannot help be aware by now that I do not like the meme concept. It seems, in a word, "superstitious" to me - just the sort of concept that scientific progress will require us to abandon.
There's criticism of the concept of "selfishness" and criticism of the concept of "replication". Harms recognises the "meme's eye view" as a valid perspective, but claims that describing culture in terms of a symbiosis between memes and genes is "awkward". He writes, on page 51:

Methodologically, ontologically, the meme is a mess. For the purposes of popular appeal, however, it could not have been better designed by a Madison Avenue advertising exec. I must nevertheless urge that the only relevance that the meme and its shortcomings have to the application of evolutionary theory is as a distraction, or perhaps an embarrassment.
Harms does make some good points amidst the rhetoric. I agree that G.C. William's definition of a gene is not very usable - though not all information theoretic definitions of the term share the same problem. I don't much like the term "replicator" either. Many have been misled by its confusing connotations of high-fidelity copying. At best, it needs defining prominently by those who use it - to avoid misunderstandings.

However, I think his rejection of the meme is totally unwarranted. I think that all students of cultural evolution should find a sympathetic interpretation of memetics. If you don't understand memes you close yourself off from a lot of important literature on the topic. Further, you then have to find objections to memetics - and there aren't really any decent technical criticsms of memetics: it's a perfectly valid framework for studying cultural evolution with. Failing to understand memes isn't big or clever, it just means you didn't try very hard to understand them.

To respond to the specific criticism that "describing culture in terms of a symbiosis between memes and genes is awkward" - that's mostly nature's fault - the fact that creatures are symbiotic composites of multiple types of agent which exhibit partial genealogical independence is certainly a complication when constructing models. However, looking at the tangles the rival "inclusive phenotype" approach results in, a symbiotic union comes out looking like the simplest model which captures the observed behaviour. Such models of symbiosis have been widely, though sluggishly adopted by mainstream biology - though their impact on academic cultural evolution so far has been pretty minimal.

Having rejected the entire body of existing work on cultural evolution and memetics, Harms is, in his own words sent back to the drawing board in understanding cultural evolution. His idea of a replacement theory is one oriented around cells. He titles a section "cultural transmission as a cellular process", and explains that the cell is the basic explanatory nexus in biology. He says:

You can account for everything that memes are supposed to do in terms of the things that human beings do. The converse does not hold.

I think this proposed project has proved fruitless. Writing off most of the massive existing literature on cultural evolution was misguided. It isn't even clear that Harms is fully aware of the literature he is dismissing. Boyd, Richerson, Cavalli-Sforza and Feldman get one mention, and that says that these folk "seem to find memetics a timely label for an established and respected approach to the study of cultural evolution and transmission". Since I don't recall ever hearing a positive word about memetics from Cavalli-Sforza or Feldman, this raises the issue of whether Harms has actually looked at their work. It certainly isn't clear from the book that he has. Certainly these days, many in the field will roll their eyes at an attack on Dawkins, Dennett and Blackmore. They have a point: why not criticise some of the actual scientists working in the area - rather than their popularisers.

The proposed explanation of cultural evolution that Harms offers - namely "the cells did it" - seems to me like no explanation at all. It is like explaining the function of the brain by saying that it "computed" something. The point of memetics is that it allows you to use the existing theoretical framework of evolutionary biology to understand culture. Recombination, mutation, selection, adaptation, drift, frequency analysis, phylogenetics all apply to culture as much as to DNA-based organisms. We don't really have a corresponding theory of cellular function - since cells are flexible, diverse and can do many things. Cells aren't even really on an appropriate level to offer an explanation of culture. Nor do cells help much with the increasingly common phenomenon of cultural transmission via computers. Harms discards a lot of useful material - and doesn't offer much to replace it with.

I was expecting the rest of the book to expound on Harms' own theory of cultural evolution, but it doesn't. Next he has rather abstract chapters on populations, information theory and selection. The chapter on populations sets up a framework for a kind of universal Darwinism. He observes the generality of selection, discusses the sorting of pebbles on a beach and discusses general population-based models with variation and selection. Harms focuses on philosophical foundations.

Towards the end of the book there are two chapters about a naturalistic approach to meaning. The first chapter is about epistemology, and the second one is about morality. Harms is a philosophy instructor with an interest in ethics, and these seem like the punchline of the book. Harms objects to the idea that you can't go from is to ought. Instead he thinks that evolutionary theory informs morality, without endorsing any particular moral position. Rather he thinks that evolutionary theory helps explain why we adopt a diverse range of moral positions. Harms writes these chapters a bit more passionately than the rest of the book, and I don't disagree with his positions.

However, this whole book is pretty dry and tedious. Though the subject matter is dear to my heart in many places, Harms ladles on Kant, Quine, Locke and Hume in hefty doses, sprinkles on dry mathematics and dissects the philosophical minutae involved until the topics become dry and lifeless. Part of the problem is that I have a scientific background, and Harms is a philosopher - so we aren't speaking the same language much of the time.

Apart from the memetics critique, my favourite part of the book was where Harms discusses Otto Neurath's boat. Harms writes:

Neurath likened conceptual progress to rebuilding a ship on the ocean while traveling in it.
This is a beautiful image that I hadn't encountered before. However, the fact that this was a highlight reflects rather poorly on the rest of the book.

Enjoy,

Friday 23 November 2012

Tim Tyler: Positional inheritance

Transcript:

Hi. I'm Tim Tyler, and this is a video about positional inheritance.

In universal Darwinism, copying is found ubiquitously in nature, from spreading ripples to propagating cracks, from growing crystals to scattering radiation. Copying - in conjunction with variation and selection - forms the basis of Darwinian evolutionary theory.

The copying in universal Darwinism includes DNA, culture, and a whole bunch of other aspects of the environment. To give some examples of environmental inheritance, rabbits inherit rabbit warrens, corals inherit their coral reef - and so on. The most common form of environmental inheritance is positional inheritance. To give some examples of this:

  • Raindrops - split and produce offspring that inherit their parents' position.
  • Cracks - have dividing tips and offspring crack tips start their lives near to their parents.
  • Atoms - split during nuclear decay - and the offspring particles originate near the parent atom.
Because of the property of physics known as locality, any form of inheritance is also accompanied by positional inheritance. That makes positional inheritance the most widespread form of inheritance in existence.

The products of positional inheritance often form tree-like structures. The roots and branches of plants resemble trees - and actually are phylogenetic trees of plant cells, laid down in order during development - in a combination of phylogeny and ontogeny. Similarly, lightning, propagating cracks, fractal drainage patterns, and crystalline dendrites are all associated with prominent visual trees. In each case, these are family trees, that show the path of descent. Sometimes the associated phylogenetic trees are less obvious. For example, in a landslide, each moving boulder has been pushed into motion by collisions with one or more parent boulders. Though each boulder can trace its ancestry back to the first falling stone, the resulting family tree is not obvious to casual observers. It's the same with raindrops in clouds and vortices in turbulent fluid flow.

Positional inheritance also results in adaptation - another hallmark of Darwinian evolution. Cracks adaptively seek the weakest path through matter, streams adaptively trace out the boundaries of their associated drainage basins and turbulence selectively forms where there is the most energy to feed it.

One thing that evolving systems typically need, in order to exhibit complex adaptations, is high-fidelity copying. Excessive noise often results in inherited information getting lost - and this leads to the disintegration of complex adaptations. However, positional inheritance often has pretty high fidelity - allowing complex adaptations based on it to remain stable.

Though positional inheritance is a pretty central concept in Darwinian evolution, it is a curiously neglected idea. While some seem to appreciate that organisms inherit their parents' environment as well as their genes, simple inheritance of position gets practically no recognition as a component of evolutionary theory. Sad times for Darwinism.

Enjoy,

Note: This is an expanded version of a previous post on this topic. See also: velocity inheritance.

Sunday 18 November 2012

2012 interview with Peter Richerson

6:10 has Richerson's discussion of "cultural units". He says:

There's this idea that has been promoted, particularly by Richard Dawkins that inherited systems have to be digital - the way genes are supposed to be - that there have to be discrete units that are potentially infinitely long lived - and Rob Boyd and I think this is not correct actually - you can have "unitless" evolution without any trouble really.
I'm pretty sure this is a misunderstanding. What Dawkins said (in "River Out Of Eden", page 19) was:
Only a digital genetic system is capable of sustaining Darwinism over eons of geological time.
The idea that I think Dawkins was getting at here is that evolving systems "go digital" after a little while, and discover the advantages of digital transmission. Genes went digital this before DNA was invented, and many types of meme went digital before computers were invented - in what is known as the digital revolution. They key driver of these transitions is attaining better copying fidelity - and thus gaining the ability to maintain the integrity of a larger genome.

Richerson attacks the idea that everything that is inherited is digital - but that claim seems to be a very silly one. I don't think that is a claim that Dawkins ever intended - or would endorse. Indeed, Dawkins himself gives examples of low-fidelity cultural transmission in the same chapter, namely: audio tape recorders and photocopying (see page 16).

Tim Tyler: Cullen, Contagious Ideas (review)

Transcript:

Hi! I'm Tim Tyler, this is a review of this book:

Contagious Ideas: On Evolution, Culture, Archaeology and Cultural Virus Theory by Ben Sandford Cullen

Ben was writing this book when he died unexpectedly in 1995. His manuscript was subsequently combined with parts of his doctoral thesis on the topic, edited together and published in the year 2000.

In the book, Ben puts forwards his own theory of cultural evolution - which he calls the "Cultural Virus Theory". Though Ben says his theory was developed independently of memetics, the two theories are complementary - a fact that is recognised in the book.

Ben starts out with what he calls the cultural virus critique. He devotes his first four chapters to looking at alternative theories of cultural evolution, and explaining what's wrong with them. The first chapter criticises social Darwinism, the second criticises Ed Wilson's sociobiology, and the third and fourth chapters criticise cultural selectionism.

The critique of Ed Wilson's sociobiology is pretty accurate. Wilson's attempt to reduce cultural phenomena to things that benefit genes looks misguided retrospectively. However, since this book was published, Wilson's brand of sociobiology has mostly faded away - and been largely replaced by a more politically correct form: "evolutionary psychology". In avoiding all mention of differences between humans, this doesn't so much attempt to explain culture in genetic terms but rather belittles its influence and ignores it. Its practitioners typically don't have an understand cultural evolution.

The third and fourth chapters mostly look at what Ben calls "American Cultural Selectionism". This is characterised by explaining culture in terms of human traits, and tracing their passage between humans in terms of "oblique" and "horizontal" transmission of those traits. Ben classified these models as being "inclusive phenotype" models - since they include traits encoded by genes and memes in the phenotypes of human individuals. Rather than featuring distinct cultural individuals, phenotypes and populations - like memetics does - the "inclusive phenotype" models combine cultural and genetic influences within one human phenotype - in what Ben patronisingly refers to as a "bio-cultural muddle".

Of course, since cultural traits exhibit geneaolgical independence from the human germ line, modeling them in terms of a combined phenotype makes very little sense. It is rather like modelling a human population infected with smallpox by considering the DNA of the human hosts and the DNA of the smallpox virus as a single phenotype influenced by both sources, with the smallpox traits exhibiting horizontal and oblique transmission. Then you can consider one group of humans with the "smallpox trait" wiping out other groups of humans that lack it. However, biologists don't tend to do this much because of Occam's razor. Rather considering one phenotype, with multiple inheritance pathways, is simpler and neater to consider the smallpox virus to have one phenotype - and the human hosts to have another phenotype - and to consider their relationship to be a parasitic symbiosis between the two distinct entities. The exact same approach works very effectively with cultural inheritance.

Ben's critique of position of the American Cultural Selectionists is pretty devastating, in my opinion. It's much the same critique as I have previously offered, but spread out over two book chapters. I tend to use the term "extended genotype" instead of Ben's "inclusive phenotype" - but we are essentially talking about the same idea. Ben's critique remains relevant today - since the strain of American Cultural Selectionism involved is associated with the work of Cavalli-Sforza, Feldman, Boyd, Richerson, Durham and Rindos - and it has subsequently gone on to become the most popular form of cultural evolution within academia. Understanding where these folk originally went wrong is important to understanding the history of the field. A better model of cultural transmission invokes cultural symbionts. This has previously been previously described by Cloak and Dawkins - in 1975 and 1976 respectively.

While a single "inclusive phenotype" with genetic and cultural traits isn't all that much use for modeling cultural evolution with, one thing it is good for making byzantine mathematical models that appear to be sophisticated and original. The correct perspective on cultural evolution reuses the same concepts of symbiosis that already exist in the realm of organic biology. However it requires practically no new science - and so isn't a great source of revolutionary papers. By claiming culture exhibits the strange new phenomena of oblique and horizontal transmission that were absent from the organic realm, the American cultural selectionists had found an excuse to develop their own innovative models of the process of cultural transmission. If you think of academia as a place where affiliations with prestigious individuals are cultivated, these complex mathematical models make a lot of sense - since advanced mathematics is an impressive thing. However, the combination of an awkward, complex model and the associated reams of complex mathematics had an unfortunate side effect: few understood the work. As a result, the important science of cultural evolution has stagnated in academia for decades.

How has American Cultural Selectionism survived for so long - when it is based on such misguided models? Some of the practitioners involved do seem to have tipped their hats hat towards the perspective of symbiosis in the mean time. The books on the topic from the 1980s typically made no mention of symbiosis, viruses, parasitism, epidemiology - or any of the tools you actually need to understand cultural evolution. However, if you fast forward to the 21st century, you will see occasional mentions of these things in the associated academic literature - where they are often described as being "analogies". This takes the sting out of the type of criticism given here - since the defenders of the theory can point out these occasional passages related to symbiology and say: look, we actually have that covered. However the symbiology remains little more than a cosmetic veneer. The incorrect "inclusive phenotype" model is still there, misleading a new generation of researchers about the nature of cultural transmission in humans.

After explaining the problems with the competing theories, Ben goes on to lay out his own proposal: the "Cultural Virus Theory". It is a symbosis-based theory - a lot like memetics, but with its emphasis on cultural phenotypes - rather than heritable information. Ben explains that memes act as the genotype, while his cultural virus idea plays the role of phenotype in his model. He says he thinks the virus perspective is more palatable than considering artifacts to be organisms - since historically most artifacts have hijacked copying machinery inside humans to ensure their own reproduction. Ben doesn't seem too worried about the negative connotations of culture as a virus. There are some helpful viruses and many viruses go on to become part of the germ line of their hosts - so viruses aren't all bad.

For me one of the most interesting parts of the book was where Ben expounds on the idea of cultural kin selection and cultural eusociality. Ben has a whole chapter on cultural eusociality - where he explains that many of the cultural artifacts that we see that don't conspicuously reproduce are really sterile worker individuals. He says that things like chairs and tables don't directly reproduce, but instead they act to divert resources (particularly money) back towards the factory that produced them - and the factory plays the role of "queen". Ben's ideas about kin selection and eusociality in culture are mostly good - and were well ahead of their time. However I think that in places they may go a bit too far. When you have a distributed organism with sterile castes it is sometimes best to describe it as one big organism - rather than as an advanced form of social organisation. Also distributed phenotypes need not necessarily be classified as cultural individuals - they could be more like hair or nails, feathers or tusks. My nit picking aside, Ben is pretty clearly correct about the ubiquitous nature of cultural eusociality - and many of his examples of it are good ones.

Lastly I should probably say some bad things about the book:

  • The book starts off slowly - and I didn't get along with much of the material about evolutionary progress and Lamarckism in the first chapter.
  • Ben was an archaeologist and he uses examples from archaeology ubiquitously in the book - but it is a general book about cultural evolution, and he should probably have cast his net wider when selecting examples.
  • The idea that cultural entities act as viruses may be easier to swallow than the idea that they are partly-independent symbiotic fully-blown organisms. However, it is difficult to deny that memes create their own manufacturing facilities - rather than simply hijacking the brains of human hosts. Computers and the internet are the product of memes, much more than genes. The "cultural virus" perspective is too narrow - as well as having undesirable connotations of parasitism and host harm.
  • Ben talks quite a bit about cultural predators preying on humans. It turns out that his definition of a predator is that it is a large parasite - and it makes no mention of quickly devouring the victim for food. Most require predators to kill their hosts.
  • Most of the book is fairly readable - but with the material imported from his thesis, the readability goes through the floor.
However, overall, this book is pretty great. It's nice to have a thorough critique of American Cultural Selectionism available in print - and Ben was a pioneer in applying the theories of kin selection and eusociality to cultural evolution. If academics had been smart enough to follow in Ben's footsteps, the theory of cultural evolution wouldn't be in anything like the mess it is in today - but alas, that isn't what actually happened - and there's still much work to be done.

Enjoy,

Saturday 17 November 2012

Tim Tyler: Universal reproduction

Subscribers who share my interest in Universal Darwinism may be interested in my latest video:

Transcript: http://alife.co.uk/essays/universal_reproduction/

Friday 9 November 2012

Positional inheritance

In universal Darwinism, copying is found ubiquitously in nature, from spreading ripples to propagating cracks, from growing crystals to scattering radiation. Of course, copying, variation and selection are the basis of Darwinian evolutionary theory. The copying takes a variety of forms, but the most basic is positional inheritance - the topic of this post.

It is common knowledge that people inherit the environment of their parents - along with their parents genes. They inherit the local climate, the local language, government and religion - along with traits coded in DNA. These are examples of positional inheritance.

Many other organisms lack good dispersal strategies and exhibit the same kind of effect. Rabbits tend to inherit the warren of their parents. Corals inherit their parent's reef - and so on. Much the same is true of many inorganic natural forms. For example:

  • Splitting raindrops - produce offspring that inherit their parent's position.
  • Propagating cracks - when a crack tip divides the offspring crack tips start their lives nearby.
  • Nuclear decay - when atoms split, the offspring particles originate near the parent atom.
Because of multiverse locality, any form of inheritance is also accompanied by positional inheritance. That makes positional inheritance the most widespread form of inheritance in existence.

Since positional inheritance is so fundamental, how come you have never heard of it before? How come searching for the term just turns up this page? Well, there are some "nearby" terms, which have received more attention historically. Ecological inheritance, environmental inheritance and niche inheritance. These terms are fine - but they simply don't mean the same thing as positional inheritance. Positional inheritance is the inheritance of spatial position. That often comes with a bunch of other things as well - but not necessarily.

Helpful illustrations:

Resources

References