Wednesday, January 06, 2016

Dear Alex

  Here’s where I am.  All theory must be built on physical data, sensorily perceived.  Faraday did the experiments, Maxwell developed equations to describe those data, Einstein explored the equations and developed his two theories of relativity, these were confirmed by physical data, sensorily perceived.
  As I understand it cosmologists and particle physicists often observe data, develop a model, and go and bother friendly mathematicians saying, I know you’re busy, but have you got a bit of theoretical maths that looks about this shape, and they talk for hours, and the mathematician either has or hasn’t.  If they’ve got a bit of maths that might fit, the physicist takes it away, tries it out and if it continues to fit, tries to derive new predictions from it.  If further experimental evidence, (physical data, sensorily perceived) fits with these predictions, then the fit between maths and data is making progress.  This applies even to the ascertaining of the statistical near-certainty of the existence of something as totally off the map as the Higgs particle.
  Okay, it can be argued that that’s exactly what the large group of historians, anthropologists, sociologists and others like me who are exploring the idea of cultural evolution are doing.  All I can say is that I read a bit of this work (not all of it by any means, I’m not a professional) and I don’t think it’s what a lot of them are doing at all.  Put briefly, they are using quantum physics to prove that the sun shines.  People have known the sun shines for aeons.  The clever bit is to derive quantum physics from the shining of the sun (and other bits of the universe, obviously).
  Recently I was asked to comment on a bit cult evol work by an eminent practitioner.  The study proposed a match between a statistical procedure and a set of sociological observations which were only data in the sense that the results of opinion polls on voting intentions are data.  There was a kind of a match because the whole exercise was assembled to portray the match.  It was entirely hermetic and self-referential.  I showed the results to a non-specialist friend.  His observation was crude, but along the lines that this guy is pretending to use quantum physics to prove that the sun shines.
  It must be born in mind that just because a piece of mathematics looks from a distance, and if you screw up your eyes, like a good fit for a very loosely worded notion, that doesn’t mean it is.  I guess you could use the maths for getting the Ariane capsule hooked up to the ISS to predict the social attitudes of various groups to same-sex marriage as long as you knew what those attitudes were in the first place and were prepared to cut a few corners
  The attractor: one of the most memorable lectures I have been to was by a mathematician describing the Mandelbrot set.  Sitting just in from of me was the late great composer György Ligeti (this was at the Huddersfield Contemporary Music Festival), and the lecturer gave us a lot of information on Mandelbrot and fractals and so on.  But he started his lecture with a very simple bit of equipment, a bowled base, a pylon, a weight with a point on a line, I can’t remember the details, and demonstrated the relationships of complex planes, and how they could generate a pattern.  I am in no way a mathematician, but what I took away from the talk was a strong notion of what, in terms of physical data, sensorily perceived, a mathematical attractor is.  Looking now at how the Mandelbrot set is generated  and converted into that amazing pattern of colour reiteration I realise that my idea of an attractor is less than primitive, but I can still see in my mind’s eye the in-the-world existence that underlies all that theory, and its resultant physical expression.
  The “attractor” in recent discussion about, let us say, a formative something or other that has some effect at a distance on the content and sequencing of stories, let us say Cinderella, is not an attractor of this type.  So what actually is it?  And how does it attract something as amorphous as a story?  These are not only questions that haven’t been answered, I’m not sure that they have even been properly asked.
  What is required is a huge dose of reductionism.  It is the only way science works.  If those who talk about cultural evolution actually mean what Darwin meant, then we need at first to be able to describe, in a replication variation selection way, what Pitt-Rivers demonstrated visually, the evolution in lithic technology from leaf shape to tanged arrow head.  If we are incapable of doing that, then how on earth are we to deal with something as huge and almost infinitely extensible as a story.
  I read a paper on the cladistics of the Baltic Psaltery once.  It was honest and illuminating.  The writer outlined his methodology, described where it went, and concluded, fairly quickly, that every time he tried to abstract a characteristic it shot off into the distance in a cloud of pyrotechnic bifurcations.  So to speak.
  But that is what has to be done.  Starting at a very simple level.  And it’s not only possible, it’s productive, and it’s quite fun.
  If we don’t mean what Darwin meant, then the whole project is something else, a branch of sociology, very valuable and interesting, but only to do with evolution in the sense of “things change” with the rider, we still don’t know how that happens.  Eusociality! (points gun to own head and pulls trigger).  Now, film of one young male chimp smiling and offering food to another even younger male chimp; we could work on that.
  One of the things that puzzles me is the complete lack of interest on cult evol shown by two friends of a friend, Chris Stringer, and Professor Howard Morphy, who specialises in Australian Aboriginal art and culture.  I’ve tried to interest both of them in cultural evolution, but though Stringer has been instrumental in conferences on it, he’s says it’s a very peripheral interest.  I feel the fact that they are not interested is our responsibility, because if we were doing things right they would have to be, even though the theory would in some way disrupt their specialisms.
  I am concerned about all this because I am at the moment half way through a book in which I hope to engage the general reader in the “argument that human culture, which is to say the extended human phenotype, is a subset of the mass of the universe that has evolved, in the manner described by Darwin, step by step and alongside and in obligate symbiosis with the human organism”; and, despite the promise of Mesoudi Whiten Laland 2004, I seem to be on my own.  But I’m having fun writing it, and I’m still very confident that Darwin works.
  I envy you Falmouth.  I come from south of you, just above Cadgwith down on the Lizard.  I used to sail at Restronguet club.  One of my first loves, when I was about nine, lived in a laundry near Custom House Quay.  There seems to be a song in that.  Evolution of a song, that would at least be a bit easier than a story.

1 comment:

Alex Mesoudi said...

Hi James

Sorry it took a while to respond to your post - first week back at work after the break and all that.

I do agree with you about the challenges of studying cultural evolution, by which I mean the rigorous scientific study of cultural change in a manner that really explains change (e.g. in terms of inheritance pathways, selection pressures, sources of variation etc.), rather than merely describing it, as so often happens in the social sciences. You're right that it is very easy to construct plausible-sounding models that have interesting results but no connection to the real world. And I agree that many excellent social scientists, such as Chris Stringer, are put off by this abstraction - with good reason. Not enough engagement with real-world data.

But I think I am more optimistic than you, given recent work that much better ties together models and data. For example, have you seen the work of Peter Turchin, Tom Currie and others? They are amassing a huge historical database (Seshat) - in collaboration with many expert historians - that they will use to test specific cultural evolutionary hypotheses. They've already done some work (Turchin et al., PNAS, 2013) showing that the rise and fall of empires in Eurasia can be predicted by a few simple assumptions about within-group cooperation, geography and the spread of military technology. There's more here:

And there's now lots of work applying phylogenetic methods to reconstruct the history of artifacts and even stories, which you mention. My former colleague Jamie Tehrani at Durham did some great work on Little Red Riding Hood last year, showing how it spread across different parts of the world, with minor details changing but the core of the story remaining unchanged.

I envy you growing up on the Lizard. In the 6 months or so I've been down here in Cornwall, I've found it to be a remarkable part of the country. And there's still so much to see - every weekend we seem to be discovering a new cove or beach...