"The descent of Edward Wilson" (by Richard Dawkins) | Hacker News Reader

"The descent of Edward Wilson" (by Richard Dawkins)

prospectmagazine.co.uk

24 pointstokenadult25 comments

If the original thesis managed to get published in Science, then it is indeed scientifically acceptable (which is quite different from "accepted by the scientific community"), and worthy of debate. The fact that it is receiving strong criticism does not mean it should be discarded right away. It simply means that it goes again the mainstream conceptions.

When you have studied evolution the computational way --like I did-- you have an idea of how many hidden dynamics and synergies are going on. Evolution is one incredibly complex phenomenon. Truth is we don't understand that well what drives it --yet some of us seem supremely confident in the current mainstream theory (which seems a little too simple to me to account for all I've been witnessing).

In the case of Dawkins I suppose that confidence is not entirely unrelated to the fact that he considerably helped to propel that theory to its current acceptance level. Scientists should not cherish their own ideas that much, it's not healthy for scientific progress.

Dawkins' decades-long reductionist insistence on the supremacy of genes is incoherent. Organisms do not exist in a vacuum, they interact with other organisms in their environment in myriad cooperative and competitive ways, and that affects which organisms survive and reproduce and which do not

Beyond that, it's becoming increasingly clear the genes aren't some recipe for making an organism, they're much more like a complicated self-referential computer program for manipulating reality. This genetic reductionism is like refusing to think or talk about computer software as anything but 1s and 0s

Wilson says "Riak is succeeding because of its straightforward scalability" and Dawkins responds "show me which 1s and 0s make it scalable! I see no scalable 1s or 0s. you're wrong"

You must understand that Dawkins considers "Evolution" to be the name of God, and himself to be its Prophet.

Don't discredit his work simply because he disagrees with your religious views.

I said nothing about my own religious views - merely that Dawkins long ago abandoned science for his own brand of dogma.

> Organisms do not exist in a vacuum, they interact with other organisms in their environment in myriad cooperative and competitive ways, and that affects which organisms survive and reproduce and which do not

What's your point? In what way would Dawkins disagree with this statement?

> they're much more like a complicated self-referential computer program for manipulating reality

What does "manipulating reality" mean? What reality is being manipulated, if not an organism's cells?

I think Dawkins believes that genes are being manipulated not cells. But Dawkins disagrees that this genes (that cascadly manipulate reality) contain information about reality beyond cells, like for example different groups dynamics. While this other guy thinks the opposite. (Thats what i get from his squirels example.

Dawkins has hypothesis that evolution is "blind" process, so if this is true than it would probably collapse.

Is it your understanding that Dawkins and Wilson (or any other reputable scientist) disagree about the information content of genes? If so, can you provide a source for that?

1. yes 2. this article quote: "That’s differential group survival. But you’d never say of any part of a squirrel that it evolved to promote the welfare of the grey squirrel over the red. Wilson wouldn’t say anything so silly about squirrel"

Are you claiming that Wilson would in fact say something like that about squirrels? If so, why? If not, then what scientist agrees with you if neither Dawkins nor his rival Wilson?

More to the point, that quote doesn't refer to the information content of genes at all. You'll certainly agree that it doesn't do so explicitly. Can you find a quote that explicitly says that someone disagrees with Dawkins about the information content of genes? Maybe in the Wikipedia articles for gene selection and kin selection?

no i don't have explicit quote. but this is what it boils down to. What else could it mean?

No, it doesn't boil down to that. They agree on what information can be encoded genetically. They disagree on a detail of the mechanism of evolution: To what extent does selection occur at various "levels" (gene, individual, kin-group, species, etc)?

> This genetic reductionism is like refusing to think or talk about computer software as anything but 1s and 0s

Dawkin's book is called "The Selfish Gene". It is not called "Selfish Adenine" or "Selfish exon".

Your analogy is interesting. Are 'genes' for buffer overflows easily found via binary searches? Would that be a defective gene? Is input sanitisation an example of a beneficial gene being transmitted, via sex of open source and responsible diclosure?

It is intriguing that not one comment (here or on the site itself) actually addresses the technical content of the article. It's all foaming at the mouth about Dawkins.

The name 'Nowak' rang a bell, so I dug through my citation manager. His papers that I had stocked away are primarily evolutionary game theory in various mathematical settings (static graphs, lattices, or simply aggregate, unstructured populations). He dresses it up in various cute ways, but they're still textbook exercises in game theory with no biological relevance.

The argument from Dawkins is really simple: there are heritable traits (call them genes; forget DNA). We need a word besides just "trait" because of transmission of recessive characteristics. Naively, you might say that if an organism has more offspring on average than other organisms in the same population, then there will be more copies of the gene.

That's true, but it's not an adequate accounting. The genes of a rabbit that has fifty times as many babies, but they're all runts and die, the genes of that rabbit are not going to show up in the population in the longterm. So you have to talk about the expected long term frequency of the rabbit's genes in the population. And once you realize that, you start looking for all the copies of the gene. It doesn't matter whether the gene in a particular organism is identical by descent or identical by allele. Thus your bookkeeping has to include other paths of descent.

However, genes aren't traits (the whole recessive transmission thing). Thus any genetically encoded strategy can't depend on the genes of any other individuals, only on proxies. But proxies can be faked. Among proxies, relatedness is 1) hard to fake and 2) universal among sexually reproducing organisms. Thus it's a really important piece of bookkeeping to do when calculating out expected frequencies of genes. And that's kin selection. Not that it's the only proxy. It's just a really, really common one.

As for group vs kin selection, ideas of superorganisms, or such things, it's all a red herring. In the end, what I actually need is a theory describing masses of individual creatures out being born, eating, mating, and dying. Doesn't matter what abstraction you want to couch the case in question in initially, you eventually need to get back to that.

>both phenomena are more parsimoniously treated as emergent manifestations of individual phenotypes, and it is individual phenotypes that are directly influenced by genes. You may choose to treat a dominance hierarchy as a group phenotype if you insist, but it is better seen as emerging from each hen, say, being genetically programmed to learn which other hens she can beat in a fight and which normally beat her.

This sounds like an argument that applies equally well to the phenotypes of organisms - you can treat them as entities in their own right, but they're more truly seen as the result of particular gene expressions. Indeed, the end of the article seems to be acknowledging that families have phenotypes; there's no reason this shouldn't apply to groups.

You have skipped over the most important word in the quoted phrase: 'parsimoniously'.

Yes, there is no reason this shouldn't apply to groups, but is that the simplest explanation that fits the facts? The more parsimonious explanation is that genes (being the primary mechanism of propagating traits between generations) are the unit of replication, and closely-related kin have much higher gene similarity than arbitrarily-sized groups (due to the mutation/genetic drift that occurs all the time). Natural selection can only operate on replicators, and genetic replicators only create kin. These replicators may influence group formation and their subsequent natural selection, but is there any evidence for that hypothesis that isn't suitably explained by kin selection?

>Natural selection can only operate on replicators, and genetic replicators only create kin. These replicators may influence group formation and their subsequent natural selection, but is there any evidence for that hypothesis that isn't suitably explained by kin selection?

Territoriality and dominance hierarchies can't be explained in terms of kin selection (except insofar as they can be explained in terms of individual selection; my point is they have very little to do with kinship). Nevertheless they are evidently selected for, and therefore it makes sense to regard them as phenotypes.

> ...the end of the article seems to be acknowledging that families have phenotypes; there's no reason this shouldn't apply to groups.

You state that "there's no reason this shouldn't apply to groups", but that's the critical question under investigation! Namely, do the behaviours (that you're classifying as a part of the phenotype) provide evidence for group selection, or are they adequately explained by kin/individual selection? It's a given that group selection could occur, but where is the evidence that it does occur?

It's not a question of "adequately explained by". All evolution is explained by selection on genes, because that's what actually happens. There is no evidence that individual selection occurs, no gene propagation that can't be explained by gene selection and means we have to look to individual selection for an explanation. The notion of an individual phenotype is useful because it simplifies understanding why certain genes are more successful at replicating themselves, and this statement is also true for the group behaviours we're talking about.

Love how he spends a couple paragraphs arguing from authority, then says "scientists don't like arguing from authority"

If anyone is reading this comment without having read the article, note that he didn't do this inadvertently: "If it was authority that got the paper published, there is poetic justice in deploying authority in reply."

Funny how the comments rapidly descend into nonsense about memes. I wonder if he sometimes regrets introducing that concept.

Interestingly the commentators seem to be some sort of anti-Dawkins cabal which includes this rather charming fellow:

http://en.wikipedia.org/wiki/User:NBeale

http://starcourse.blogspot.co.uk/search/label/Richard%20Dawk...

Who obviously first started hating Dawkins because of his religious views. The cynic in me wonders if it was a co-ordinated attack, especially the silly tit-for-tat between him and the Munro guy. Given the frequency that he attacks him on the blog you'd think that would be the first book on his night-time reading list.

Dawkins doesn't have religious views. He has views on religions.

I think Dawkins' narrow concept of "the gene" is outdated, and he does not realize it because he is an old biologist, not an information scientist.

A "gene" was never a physical object, it is an information object--a mental placeholder for the smallest chunk of genetic information that determines a physical trait. But as our understanding of the nature of information has advanced, it seem like the gene is ultimately an obscuring concept, not a revealing concept.

A "gene" is not a specific, standardized DNA sequence. As a concept it is not physically specific enough to tie into the stack of rigorous scientific knowledge that starts with quantum mechanics and ascends through physics, past chemistry, to biology. As a result it cannot be treated wholly mathematically, which leaves open the risk of storytelling vs. actual rigorous hypothesizing and testing. The last few chapters of "The Selfish Gene" are filled with fuzzy, untestable storytelling for instance.

On the other hand, as a unit of information it is far too specific and limited, because of course a gene is not the only way that information can be passed from one generation to the next. Information is also passed through society and culture--and yes, that can lead directly to physical traits. For instance my wife has pierced ears, and so does my mother-in-law. This hole in the ear is a physically measurable trait of both those humans, and it was caused by the direct passage of information from one generation to the next. That chunk of information just does not happen to meet Dawkins' conception of a "gene". But why does he get to define which information "really matters", and which doesn't, when it comes to physical traits and survival of the fittest?

When I read this book review I see Dawkins attempting to defend an early, fuzzy conception of an information theory. Evolution is an information process, and information is information is information, regardless of whether it is coded in DNA, in societal behavior, or in the printed word. I have not read the book in question, but Wilson might be better equipped to break down old distinctions because of his focus on social behavior--which carries information.

Ironically Dawkins started down this road with his idea of the "meme" but has not been able to make the mental leap to a fully realized information-centric vision of life and evolution.

Biologists are in for a hard time in the next few decades. A lot of fuzzy concepts that they take for granted--like "gene"--are going to be broken down and reformulated rigorously by mathematicians and computer scientists. It is already well on its way--a friend who started his Ph.D. in evolutionary biology in the lab now spends the vast majority of his time at his laptop, coding hypotheses and tests against a deep data set of sequenced genomes.

Warning: the preceding has been the ravings of an interested amateur and should not be considered Real Authoritative Science.

edit: fixes

A gene was not conflated with a DNA locus until the 1960s. Classical genetics treated 'gene' as any trait that could be stably isolated and transmitted in breeding experiments.

I agree that a lot of the broken ideas are going to be revised, but in many cases they're going to be revised back to the original meaning before someone better at marketing than science forced a broken meaning on it.

Reply on news.ycombinator.com