Does studying boat design show us that culture is subject to natural selection?

The creative PNAS study that Razib pointed out the other day is out, and I’ve read it. It comes from two biologists at Stanford, Deborah Rogers and Paul Ehrlich, who studied the canoe design of 10 Polynesian groupsPolynesian Canoe and one Fijian group for functionality and symbolism. The paper, “Natural selection and cultural rates of change,” is open access, so please download a copy and read what the authors have to say first hand. They decided on these Oceanic cultures because they were colonized by one cultural group that radiated and became relatively isolated from one another. In other words, little outside influence, or noise, from other cultures has theoretically impacted Oceanic canoe design.

They find that functionality (traits that affect whether or not the occupants of the canoe will survive or not) changes very little, whereas symbolism in canoe design (such as aesthetic, spiritual, and decorative) changes much faster. From this observation, they conclude that “cultural change, like genetic evolution, can follow theoretically derived patterns,” and that natural selection is the driving force behind the evolution of cultural traits. The authors rationalize boats are being selected naturally through a quote from the French philosopher Alain,

“Every boat is copied from another boat… Let’s reason as follows in the manner of Darwin. It is clear that a very badly made boat will end up at the bottom after one or two voyages and thus never be copied… One could then say, with complete rigor, that it is the sea herself who fashions the boats, choosing those which function and destroying the others.”

This quote doesn’t make much sense to me. It sure is a simplified analogy… one that is supposed to help us digest and reduce… But I’m left a bit confused on how the authors found evidence of natural selection out observing intelligent boat design. Unlike boat designers, who learn and modify boat design based upon past experiences, natural selection does not operat with conscious understanding of trial and error. Natural selection in nature is independent. Boat design is not independent. Boat designers can go study past designs, learn from their mistakes, and make modifications… So is not functional boat design just selection?

That’s one flaw I see in this paper, one that I haven’t really completely thought out. I do appreciate how the authors are one of the first, in my knowledge, to attempt to integrate how ‘cultural things’, such as items that serve a life or death purpose in everyday life, have a functional constraint. Often we’re taught that ‘cultural things’ are arbitrary but we know we want our things to work, and we’ve will always design them to work better.

I’ll leave you with words from, Nina Jablonski, an anthropologist that I really respect and whose work I appreciate immensely. She christened this paper as, “one of the most significant papers to be written in anthropology in the last 20 years.” Man, that’s a pretty amazing endorsement to have, but just cause Jablonski casted her vote doesn’t mean I’m sold, nor does this make it into the trophy cabinet for classical anthropology papers.

    Rogers, D.S., Ehrlich, P.R. (2008). Natural selection and cultural rates of change. Proceedings of the National Academy of Sciences DOI: 10.1073/pnas.0711802105

17 thoughts on “Does studying boat design show us that culture is subject to natural selection?

  1. I very much agree, and i think It`s a factor in the evolution of objects, that the nature laws and the enviroment somehow judges what will last and what will be made again and not lose interest. Of course we can try to predict what the mother nature will accept, and go two steps ahead, but then you did not really test it against the natural selection process. There is so many versons of designs tried out, that in general it is random variations. Many(all) designs are random search and findings, within logical chosen thinking paths. And this is a random reply, I know..

  2. One could then say, with complete rigor, that it is the sea herself who fashions the boats, choosing those which function and destroying the others.

    I haven’t yet read the paper (browsing Google Reader on the sly at work), but the first thing that comes to mind is that the “sea herself” (to use the author’s original anthropomorphism) doesn’t design features in ships that are functional to the sailor’s purpose (fishing, war, cargo, etc.); nor does the sea’s influence include anticipating or predicting what features might work. In other words, natural selection isn’t predictive whereas human selection is. I might note that a short keel provides stability and consider a longer keel -or predict that where one sail is useful, multiple sails might be more useful

  3. With apologies to Nina Jablonski, I found this piece remarkably naive. First of all, as you both remind us, human invention is NOT a form of “natural selection.” Humans didn’t simply try this and that at random, as though their ideas were simply mutations, with certain solutions “winning out” over thousands of years simply because they happened to be better “adapted” to the environment. Boats are created and modified via a process better described as, literally, “intelligent design,” NOT natural selection. They were designed based on their designer’s knowledge of the sea, a knowledge gained by trial and error, certainly, but NOT via natural selection, that’s absurd. It’s one thing to challenge the notion that evolution was guided by the hand of “God” and another thing entirely to suggest that culture wasn’t guided by human hands.

    Also troubling is their willingness to draw such far reaching conclusions from such a slim body of evidence. To decide that decorative elements “evolve” more rapidly than practical ones on the basis of such a small sampling of boat designs and decorations is absurd. Also, what, exactly, is meant by “change” in such a context? In what sense can it be claimed that the decorative designs change, by whose standard of similarity and difference? No two things are exactly alike, there will always be differences between them. My guess is that the overall style of these decorations may well have remained consistent over great lengths of time, despite certain variations of detail.

  4. I think there is a sense of unfulfillment in many of the “human sciences” when it is impossible to reduce the actions of human intelligence or consiousness to strictly materialistic phenomenon. (And that is probably where the irrelevant “natural” in the title is comming from.)

    To accomplish this reduction has been an outspoken goal in some philosophical and/or psychological schools. Reductionism is a common unspoken assumption.

    On a general level materialistic views on the mind/body problem is totally dominant today. But as yet there has been no satisfactory modell to sort of explain away “the gost in the machine”.

    Personally I think this attitude is misinformed. You can have a purely materialistic view on the mind/body problem without the need to reduce its functions. Intelligence, in these kinds of contexts, provide a phenomenon that is totally sui generis. There is no need to explain it away – only to acount for it.

  5. I think that this article does hold some validity. The authors are simply making a comparison in the mechanistic structures that control and develop existence. Darwin himself drew upon these mechanisms for the entire concept of natural selection when he looked to Hutton and Lyell for the concept of uniformitarianism and actualism. The authors are simply implying that the same forces of natural selection act upon cultural traits. Though several of you have mentioned trial and error as erroneous when considered under the pretenses of natural selection, I argue that trial and error of human ideas when tested against the environment, is a highly similar system of selection to the processes of natural selection on genetic inheritance. Ideas are equally variable, some are prone to success and others to failure.

  6. I agree totally with Chris on this. Our culture and our technology constantly evolve. And “trial and error” are certainly a part of biological evolution.

  7. Conscious trial and error is not a part of biological evolution. When a boat design fails, the boat makers think about the failed design and improve upon it. In biology, this doesn’t happen.

  8. Hi Terry,

    Yeah I caught that in the original paper and I think it is flawed to consider it so absolute. Boat making is highly structured and often relies on more than one person. The boat maker is also not always the boat user.

    That matter aside, when a boat design fails, sure the sailors perish but knowledge of the failed design doesn’t. People learn remarkably quickly to change up a boats design when members of their group don’t return. That’s what I mean by conscious trial and error.


  9. I’d assume we all agree that boat evolution doesn’t evolve through random variations but we’d have to agree that boat design has evolved. The original Australians probably arrived there on logs or bamboo rafts whereas the first people to cross the wider distances to New Britain, New Ireland and the Northern Solomons presumably had a more advanced boating technology. Again those who reached Fiji, Tonga and Samoa had a yet more improved boating technology, possibly outrigger canoes, and those who eventually moved beyond those islands into the wider Pacific must have had an improved sail. This indicates boating has evolved, probably through the spread of new technologies. Biological evolution also seems to work most often through the spread of new genes. Therefore I maintain there are parallels.

    I realise we are all nervous about providing quotes for the Intelligent Design brigade but perhaps these sorts of parallels may help us explain exactly where their arguments fall down.

  10. The example of boat design ‘evolving’ is definately not incorrect, but it is incorrect to call it natural selection. If anything it is artificial selection, a process by which favorable traits considered desirable by humans are systematically favored over unfavorable traits. There’s not much ‘nature’ involved when we consider there’s a conscious element here that’s ultimately doing the selection.

  11. Kambiz. Perhaps “it is incorrect to call it natural selection” however it does demonstrate similarities. The main point I’d disagree with in the original article is their statement, “little outside influence, or noise, from other cultures has theoretically impacted Oceanic canoe design”. I believe they underestimate the influences from outside that have influenced canoe design. I also believe we underestimate the genetic influences from outside that have influenced regional human evolution. All sorts of influences bounce around the world.

  12. Let’s be clear about what the authors actually do here: they compare how much variability there is in the functional versus symbolic data set (or rather, how different or similar islands appear in each dataset) and then use this as a proxy for ‘rate of evolution’. In order for this to work you have to assume canoes are phylogenetically related, and that both the ‘functional’ and ‘symbolic’ traits were conjoined in the evolutionary process. That is, the symbolic traits aren’t part of multiple other evolving systems (e.g. plant use, wood carving) and don’t get tacked on to canoes randomly. Otherwise you could simply be testing whether people are more likely to actively share/borrow functional traits or symbolic traits, rather than how quickly each evolves. Or worse: you could be comparing variation in canoes versus variation in arbitrary variables.

    The authors don’t seem particularly familiar with the prehistory of Oceania, previous research done on canoes and voyaging, or the context in which Haddon and Hornell gathered their data. For example many of the islands included had ceased long distance voyaging ca.1400AD, and it is widely believed lost canoe making technology, others such as Fiji had recently adopted Micronesian canoe designs.

    But whatever. The main problem that I see is the way in which traits are defined and classified as either functional or symbolic. Several of the functional traits appear to be stylistic variations, with probably very little functional import – the outrigger attachment techniques for example. The authors speculate variance in this trait was due to “types of waves encountered” (!) or “constrained by availability of materials”. But speculation is not proof. Haddon and Hornell themselves regarded outrigger variation as stylistic/cultural. With Darwin’s finches it is demonstrable that beak shape is functional and associated with ecological niche. But how can you separate traits into ‘functional’ and ‘symbolic’ when you don’t know (much less demonstrate) the source of variation? Ehrlich and Rogers seem to define functional traits using some common sense idea of ‘integral bits of a canoe’ and contrast this with the ‘frilly bits’. If you look at their trait lists this becomes really apparent. The functional traits are elements that determine the basic canoe form, whilst the symbolic traits consist of seemingly arbitrary add-ons – whether a canoe had feathers attached to the bow, or leaves on the mast. I immediately ask: did all canoes have feathers in a given region? Did Haddon and Hornell (or more properly their sources: traders, explorers, missionaries, paintings, toy models) record that kind of thing systematically? Or were they more concerned with basic techniques of canoe construction? (actually, Yes). Is variation more evident in the symbolic traits because these contain a large portion of arbitrariness in both their definition and recording?

    What really makes me laugh about this kind of research however is the fact that they once again mine an ancient colonial era book for data. And then they (and Shennan in his commentary) go and complain about how anthropologists and sociologists don’t collect this stuff anymore. But if their findings are to have any wider significance whatsover they could very easily go out and record any similar dataset in their own neighbourhood. Maybe cellphone traits? Or how about Cars? Or Desktop computers? Or are they really arguing this kind of thing can only be detected in small scale societies, living on isolated islands, having diverged from an ancestral population, and abandoned ties with the world?

    As it is now all they have done is proved that they can generate data from a book and make it appear patterned.

Comments are closed.

A Website.

Up ↑

%d bloggers like this: