Sunday 13 October 2019

Panmemetics?

An article by Maarten Boudry proposes that the term "panmemetics" be used for the memetics visualized by Susan Blackmore - which is contrasted with Daniel Dennett's presentation in his recent book on cultural evolution. Over to the author:

in Dennett’s current account, in particular his use of Godfrey-Smith’s (2009) Darwinian spaces to make sense of the “de-Darwinizing” of culture and the conscious domestication of memes, moves away from panmemetics, perhaps more clearly than his earlier Darwin’s Dangerous Idea. Even though memes initially evolved by blind and unguided evolution, claims Dennett in his new book, many were gradually domesticated by their hosts, who became more reflective and self-conscious about them. As culture gradually moves in the direction of the “intelligent design” corner of the Darwinian space (top-down, foresighted, directed), the meme’s eye view loses traction and becomes less interesting. Let me try to spell out the differences between the two approaches, as I see them. First, unlike Dennett’s account in BBB, panmemetics leaves little or no room for human autonomy and creativity. Panmemetics suggests that we are not in control of our thoughts; if anything, the reverse is true. In the words of Robert Aunger: “Do we have thoughts, or do they have us?” (Aunger, 2002, loc. 120) A second and related problem with panmemetics is that it threatens to level all distinctions between harmful and useful cultural inventions. If the whole of human culture is seen as just a set of swarms of viruses bent on exploiting our brains as a breeding ground for their own reproduction, we no longer have any theoretical resources to distinguish between good and bad memes, between beautiful folk songs and annoying earworms, between science and superstition.
The distinction might be an interesting one, but I don't think I can stomach rechristening memetics, "panmemetics". I am not a big proponent of the "de-Darwinizing” of culture that Dennett proposes. My conception of Darwinism is big enough to include all cultural evolution. We can have a picture involving degrees of intelligent design without dragging the rather vague and subjective "Darwin"-based terminology into it. IMO, panmemetics is just memetics. Dennett's idea that the "Darwinism" fades out as the intelligent design fades in doesn't seem interesting enough to warrant a big terminological shift. We can just use the "Darwinian" term for the more general framework that includes ID.

Regarding the criticisms of panmemetics: I don't see much impact of memetics on the issue of human autonomy and creativity (as mentioned in the quote). It is true that memetics suggests that we are not in control of our thoughts - but surely that is hardly news to most psychologists. I don't agree that memetics "threatens to level all distinctions between harmful and useful cultural inventions". The most obvious distinction between "good" and "bad" memes is between meme interests and the interests of DNA genes of the nuclear DNA of the host. The host is likely to have more than one gene, and there could be more than one host, but these issues can be deal with by averaging - as in the "parliament of genes" concept.

The author raises the issue of personal host interests - and where they fit into the picture. My favored approach to this is to have three spheres involving Darwinian dynamics:

  • DNA gene evolution;
  • Meme evolution;
  • Psychological evolution;
The last one is the least well understood - but information is copied with variation and selection in the brain, and we can use Darwinian models to represent the resulting dynamics. Psychological evolution is more obviously multi-level - since copying takes place at the level of branching axon impulses as well as at the level of ideas, and probably at multiple other levels in between. Personal interests can then be discussed in terms of copied information patterns - using the same modeling framework as with memes and genes.

Maarten Boudry goes into some more details about what is wrong with "panmemetics" in another paper titled "Attack of the Memes". Overall, I'm not seeing it.

Another exhiibit from the latter paper:

Memes that are parasitic from the perspective of my genes may simply be the outcome of deliberate human choices. Talk of "selfish memes" and "rogue culture" can be misleading here, as if humans are the hapless victims of the ideas of modernity. These are not novel memetic purposes but distinctively human ones.
Well, maybe, but maybe not. Selfish memes / rogue memes are a possible and valid answer too. It depends on the circumstances. One must also ask where those humans got their preferences. Some come from genes, some come from memes and some come from psychological forces (e.g. alcohol addiction). Memes are on the list. They can influence host preferences via manipulation from within hosts minds. Because a human has some preference, that doesn't mean it didn't ultimately come from memes.

References:

Wednesday 9 October 2019

The Organism-Centered Approach to Cultural Evolution

I found a paper from 2015 that offers a reply to the criticism of mine, that cultural evolution in academia is "host centric". I summarize that criticism in my article "The host-centric approach to cultural evolution".

Ben Cullen also summarizes this criticism in his "Contagious Ideas" book. To quote from my review:

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".

Here is a paper that offers a reply. It classifies approaches to cultural evolution as "organism centric" and "meme centric" - and argues for the superiority of the former. Overall, it is as though Dawkins had never written The Selfish Gene - explaining why Hamilton, Trivers and others had shifted focus away from organisms and towards genes.

The article says:

Instead of offering a set of criteria, we will pose several questions that need to be answered in order for there to be a coherent MC. First, can there be multiple copies of a meme within an individual’s head? If so, how are neuronal states individuated into distinct memes?
The claim that these questions need to be answered for memetics to be coherent is a mistake. It is quite possible to treat the entire human brain as a black box and conly consider how memes are transmitted into and out of it, via, talking, listening, etc. Memetics doesn't depend in any way on implementation details of neuroscientific representation. It would be nice to know all the details of what went on inside brains, but it is perfectly possible to discuss meme frequencies without this. That is, by and large, what memetics has done, because all the neuroscientific details are not there yet.

Another question it is claimed that meme theorists need to answer is:

And for the spork example above, is this a case of fork and spoon memes reproducing and finding themselves within the same utensil, instead of each occupying distinct utensils? Or is a spork a unique meme related to spoon and fork memes, but not an instance of either? These are just some of the sorts of questions that a MC needs to answer"
Those are alternative modelling scenarios within memetics. Researchers might choose between them depending on their research goals. However, memetics itself doesn't need to choose between them! It is the same situation as with genetics - the same section of DNA can be modelled as consisting of two or three genes, depending on what traits the researcher is interesting in studying. Genes can overlap, and that is just a fact of life, not a foundational challenge to genetics.

I've criticized host-centric cultural evolution as ignoring the dynamics of meme replication in machines, thereby missing the very possibility of a memetic takeover. Here is this paper's take on that. It claims that is actually an advantage to the organism-centric approach - because it makes things more palatable to social scientists:

If cultural evolutionists would embrace the MC andfocus on the number of memes (e.g. the number of pro-duced sporks), irrespective of whether these memes are actually adopted by humans, then this would decrease therelevance of cultural evolution for the social sciences andthe humanities. After all, these disciplines are mainly—if not solely—interested in culture in as much as culture affects the behavior and thinking of human beings

Cultural evolution within academia was developed by social scientists - and they went out of their way to make it more palatable to other social scientists, for example by stripping off biological references wherever possible, and exagerating the differences between cultural and traditional evolutionary theory. However, that is only a virtue if acceptance by existing social scientists is recognized as a valid aim. Existing social science failed to get to grips with evolutionary theory for 150 years. They typically think that "Social Darwinism" is a failed research program, and one which they do not wish to associate themselves with. It's not that pre-evolutionary social science is useless, it is that it lacked the firm foundation that evolutionary theory is now providing. Pandering to social scientists seems like a secondary goal to me. I favor going directly for correct and useful theories. If they seem revolutionary and foreign to the old guard, so be it.

The paper at least recognizes the issue of there being two rival approaches here. However, it offers a series of mischaracterisations of memetics and gives the wrong answer to the issue. It would be better to just say that organism-centric cultural evolution is holistic and meme-centric cultural evolution is reductionistic. For me, even that is too flattering. Part of the problem is the theory is not just organism-centric it is host-centric - and the only hosts considered are humans. Machine intelligence as a host of memes mostly isn't on the radar, and the possibility that "the Bible" might act like a cultural organism with its own largely independent genotype and phenotype is also largely neglected.

References

The Organism-Centered Approach to Cultural Evolution by Grant Ramsey