The Social Evolution Forum published a piece of mine on international relations theory and cultural group selection.

This is a new response Pinker’s essay that makes different points than the ones I made in an earlier post on this blog, since those points were already covered by other commentators on the SEF.

Instead, I focus on the role that cultural selection plays (or might play) in international relations theory and that it would be hard to talk about selection in IR without reference to groups – specifically sovereign territorial states.

If you check it out, please tell me, in the comments, where I go wrong.

Update 2: Pete Richerson (my major professor) responds to Pinker.

The Social Evolution Forum has also published, so far, responses from Herb Gintis, Peter Turchin, x 2,   Michael Hochberg, and Harvey Whitehouse and Ryan McKay.

I think there will be more in the weeks to come.

Update: Herb Gintis responds to the Pinker Edge essay.  He writes about the evidence for altruistic behavior/preferences in humans, which is something I did not mention in my blog post.

Perhaps I should rename this blog “Why Famous Scientists are Wrong About Group Selection.”

This month’s post is a response to an Edge essay by Steven Pinker (author of The Blank Slate and The Language Instinct) called The False Allure of Group Selection.  My understanding is that a number of researchers in the field have been given the opportunity to respond to this essay, including Peter Turchin, and my PhD adviser, Pete Richerson.  My comments are likely to overlap, but perhaps this will be a useful summary of the issues.

First, I should reiterate, as I tried to do in first blog post, that multilevel selection and inclusive fitness are not different mechanisms of evolution, they are different ways of thinking about and modeling selection. So debates about which is the “right” or “wrong” mechanism are fundamentally misconstrued. David Queller makes this point in a comment:

 Pinker is therefore correct that multilevel selection results can usually be seen as restating things we already knew in a different language. But I am loath to say that just because I speak English, others cannot speak in (as homage to Peter Kropotkin) Russian.

Instead we should be, at best, reduced to debating which of these modeling paradigms are the most useful for answering particular questions – which is a worthwhile debate for mathematical theorists, but esoteric for almost everybody else.

Second, my comments here are about cultural group selection models. I think few would disagree with Pinker that selection on genes at the level of large groups of humans is unlikely to be very strong. But I think he fundamentally misunderstands the mechanisms of selection on cultural variation.

Pinker lists three problems he has with cultural groups selection.  I list them here followed by why I think his concerns are misplaced.

Pinker’s first and third criticisms, which I see as very similar:

(a) The criterion of success is not the number of copies in a finite population (in this case, the meta-population of groups), but some analogue of success like size, influence, wealth, power, longevity, territory, or preeminence.

(c) The “success” applies to the entity itself, not to an entity at the end of a chain of descendants. It was the Roman Empire that took over most of the ancient world, not a group that splintered off from a group that splintered off from a group that splintered off from the Roman Empire, each baby Roman Empire very much like the parent Roman Empire except for a few random alterations, and the branch of progeny empires eventually outnumbering the others.

Selection acts on cultural variation and the criterion of “success” is the number of copies of a cultural trait.  For example, if states with democratically-elected legislatures tend to do well, democratically-elected legislatures will tend to spread.  This does not require that autocratic states need to disappear and democratic states need to bud and reproduce.  It is just easier to think about this from a group-selection framework than an inclusive fitness framework because it is easier (at least for me) to account for democratically-elected legislatures as a property of groups.

Pinker’s second criticism:

(b) The mutations are not random. Conquerors, leaders, elites, visionaries, social entrepreneurs, and other innovators use their highly nonrandom brains to figure out tactics and institutions and norms and beliefs that are intelligently designed in response to a felt need (for example, to get their group to predominate over their rivals).

Selection works on variation.  It does not matter how that variation is generated.  It does not have to be random.  Although variation was essential to Darwin’s theory, it wasn’t even clear for decades after Darwin that source of genetic variation in natural populations (mutations) were (mostly) random.  Inventions do not have to be random for selection to act.  As long as all innovators are not inventing exactly the same thing (i.e., as long as iPods are different than Zunes), then selection can act on the inventions’ relative merits.  In his comment, Daniel Dennet makes this point:

Brainchildren die like flies, thousands or even millions of hopeful ventures extinguished for every innovation that truly”goes viral” and then goes on to fixation in the shared culture we pass on to our descendants. 

Pinker also asks:

But what does “natural selection” add to the historian’s commonplace that some groups have traits that cause them to grow more populous, or wealthier, or more powerful, or to conquer more territory, than others?

This question blew my mind.  It is equivalent to asking “what does ‘natural selection’ add to the naturalist’s commonplace that some organisms have traits that cause them to grow more populous, or live longer, or have more children than others?”  Following Pinker’s logic, gene-based evolutionary theory had nowhere to go after Malthus or, at best, Darwin.  Anyway, Peter Turchin, who has using cultural group selection to understand the rise and fall of empires, has a post explaining why this is a useful endeavor.

Finally, there are two more points of confusion in Pinker’s essay.

First, group-level institutions like punishment, coercion and reciprocity are evidence for the importance of considering selective forces on groups, not against it (Joe Henrich spends the bulk of his article on cultural group selection making this point).

Especially interesting, to me, was Pinker’s citing of Sarah Mathew’s work with the Turkana as evidence against cultural group selection (since I am writing a book chapter with Sarah that cites the same work as evidence for the importance of selection on cultural variation between groups).  The issue here is that Turkana males are punished by third parties if they do not participate in dangerous cattle raids of non-Turkana groups.  Raids often involve individuals from different villages that are not known to one another.  There is a large incentive to avoid the lethal consequences of raiding, but to obtain resources from the raid.  Individuals that shirk are punished by their group.  The important thing to notice here is that punishment is collective and costly and is, thus, a “second-order” collective action problem.  Therefore, invoking punishment is not good enough to explain raiding – you also have to explain why punishment happens.  That is, groups that punish do better than groups that do not.

Second, the Turkana will only raid villages of non-Turkana.  They will not raid Turkana villages, even when those villages are more convenient.  This is exactly what one would expect from thinking about things from a cultural group selection perspective.  Richard McElreath discusses the importance of Mathew’s Turkana work towards understanding cultural group selection here.

Second, Pinker seems to mistake individual psychology for selective forces.  Group selection is a framework for modeling selection – but the predictions of the models depend on what goes into the model.  Saying that a group-selection framework automatically predicts pro-social preferences or individual psychology that maximizes group fitness is like saying that using a calculator to solve algebra problems automatically predicts the answer is seven.  Multilevel selection models are tools that partition selection into different levels of analysis which help us see the conditions under which behaviors that benefit the group might be favored over ones that do not.  Therefore, pointing to the existence or non-existence of pro-social preferences does not help us determine whether the framework itself is useful.

One of my personal scientific heroes, Elinor Ostrom, passed away today.  She was a the pioneer in linking economic and with field observations of how people *actually* solve collective action problems.

I never met her, but was deeply inspired by her writing, especially the books Governing the Commons and Understanding Institutional Diversity.

An often explicit goal of her work has been an evolutionary theory of institutions – where the institutional rules under which people live result from a process of selection.  I think this is an important process to understand as we develop theories about how modern institutions might respond to large-scale environmental degradation.

If you are so inclined, here is a very recent article of hers on The Social Evolution Forum discussing this topic.

Richard Dawkins has a review of E.O. Wilson’s new book (which I blogged about here). After reading it, I was planning for my next blog post to be a discussion of how Dawkins makes many of the same mistakes as E.O. Wilson, but from the other direction.

However, David Sloan Wilson (no relation to E.O.) beat me to it.

Dawkins and Wilson are only two of dozens of scientists who have been working on the issues over a period of decades. This is in contrast to their outsized images on the public stage, as if they are the only two figures meriting attention and all the important ideas sprang from them.

I mean Dawkins and Wilson no disrespect by calling them two among many. I trust that they would agree and would defer to others especially when it comes to mathematical models, which is not their area of expertise. If the public is going to become literate on the issues at stake—as well they should, because they are fundamental to the study of human sociality—then they will need to realize that both Wilson and Dawkins get some things right and other things wrong. Moreover, the entire community of scientists is in more agreement than the infamous exchange in Nature seems to indicate.

I highly recommend the full post.

PS – I found Dawkin’s squirrel example particularly telling, for reasons D.S. Wilson explains:

 “The American grey squirrel is driving our native red squirrel to extinction…” Choosing an example of competitive exclusion between species to illustrate group selection is poorly informed. Even the concept of species selection, which is different than the concept of group selection within species, is not represented by competitive exclusion (14).

This is my review of Kevin Clarke and David Primo’s new book, A Model Discipline: Political Science and the Logic of Representations, which tackles the role of theoretical and statistical modeling in political science.  Phil Arena has a great review covering the book’s main points. My goal is to avoid duplicating his efforts – so I suggest reading his review.

Since I try to keep a foot in both evolutionary ecology and political science, I have been exposed to the methods of both. I am, to my knowledge, the only non-political-science grad student to attend an EITM summer school, so may have more insight into the context of their argument. But since most of my methods training has been in on the ecology side, this will be somewhat of an outsider’s view.

First, I really liked this book and agreed with most of the authors’ philosophy of modeling.  One of their main points is that the goal of theoretical models is rarely prediction.  Most often they are tools for reasoning about the world.  They might help evaluate the logical consistency of ideas, define terms and clarify arguments, or identify important new areas of research – but these many uses are not always clear to the non-modeler.  For example, in my inaugural blog post, I wrote about how Hamilton’s rule is the result of a simple model intended to demonstrate the mechanisms of kin selection and that the big misunderstanding of E.O. Wilson’s crusade against kin selection is mistaking it for a predictive model.

The standard story I keep hearing in political science is that the field has taken the “theoretical models as predictive models” to the extreme.  I have been warned, repeatedly, that the field’s top journal, APSR, will not publish purely theoretical models without some nod at an empirical “test” of the model.  Because of my experience at EITM, I agree with Phil that, the relationship between theory and statistics are hardly monolithic.

Primo and Clarke’s view of theoretical models would be very uncontroversial in ecology, population biology, population genetics, and related fields.  These, like political science, are fields of complex dynamics systems with important interactions at different levels and a heavy reliance on observational data for the large scale processes.  (International relations and ecosystem ecology are both hindered by a sample size of one planet and the difficulty of logistically and ethically conducting large-scale experiments.)

Clarke and Primo argue that theoretical models are like maps. Maps are designed for a specific purpose and what determines a good map is whether it is useful for the purpose for which it is designed.  In the same way that it is weird to argue about whether a map is “true” or “false” (since all maps are false in most respects), it is weird to argue about whether a model is true or false.*

I was surprised by their discussion of empirical models since Clarke and Primo did not take their argument to its seemingly logical conclusion by arguing for a model selection approach to empirical analysis (as opposed to the standard null hypothesis testing – NHT).  The NHT view of the world is that models are true and false and the job of science is to reject the false ones. In a null hypothesis test, the first step is to pick a “null” model (which, in practice, is almost always a very terrible model) and assume it is true.

The model selection view of the world assumes that all models are false – or incomplete views of the world – but that some models are better than others for specific purposes.  They try to distinguish between models based on some criteria of usefulness.  One criteria might be out-of-sample prediction.  Failing that, they can use criteria based on information theory. Another (important) criteria could be theoretical relevance based on a priori reasoning. 

I found this omission surprising, not because I’ve seen a lot of model selection statistics in political science** (I haven’t), but because, in ecology, as soon as you start reading about statistical models not being “true” or “false,” this seems to be next argument.  The standard reference for this approach is Burnham and Anderson’s Model Selection and Multi-Model Inference: A Practical Information-Theoretic Approach.   (Here is a short review.)

Overall, I recommend this book anyone interested in modeling in political science.  I especially found their taxonomy of model types helpful (see Phil’s review for details).  As someone who aspires to both a theoretical and empirical modeling program in political science, I hope that the authors manage to shift the field’s view of modeling more towards their own.

* – They are not the first to make this analogy.  I couldn’t remember whether I first saw it in McElreath and Boyd or Miller and Page – so I looked it up and it was in both.

** – Phil Arena pointed out to me that Kevin Clarke has written about model selection elsewhere.  But that makes it even more surprising to me that it was not in the book.

A few weeks ago I attended the one-day SkeptiCal Conference (partially organized by my friend Lauren). One of the day’s many interesting speakers was Alison Gopnik, a professor of psychology and author of The Scientist in the Crib and The Philosophical Baby.

Dr. Gopnik researches how babies and young children learn about the world.  While the bulk of her SkeptiCal presentation was similar to her TED talk, at the end she discussed recent experiments showing how children use two types of learning – individual experimentation and social learning (that is learning from others).  She also described these experiments in a recent interview:

A couple of recent studies show that preschoolers do something very different if they’re exploring a toy to figure out how it works than if they think somebody’s actually giving them the answer. In a nice experiment that was done at MIT, they gave children a toy to play with that could do lots of different things. You’d punch something and it squeaked, you’d push another button and something else happened, and so on. In one condition the experimenter came in and said, “Oh look, I’ve never seen this toy, let’s see what it can do,” and then bumped into it and it squeaked. In the other condition the experimenter said, “I’ll show you how this toy works.” In the first condition, the children then spontaneously explored everything else the toy could do. Whereas when the experimenter said, “I’m going to show you how this works,” the children just did exactly what the experimenter did, over and over and over again. The findings suggest that children and, presumably, adults, learn quite differently when they’re learning in this spontaneous exploratory way than when they’re learning from a teacher. Now, there are good things about having a teacher who just narrows the range of options you can consider, but there’s also the danger that you’ll wind up just essentially imitating the teacher.

Since I am interested in how the trade-offs between individual and social learning influenced the evolution of humans, I found this experiment interesting for its own sake.

However, this blog post is about last part of her talk in which she used this experiment to illustrate how children learn like “little scientists,” open-mindedly conducting experiments to discover things about the world.  That is, unless this exploration is constrained by others – at which point they cease to thinking like rational scientists and open the door to irrational pseudoscientific beliefs (which is the opposite of what most SkeptiCal participants would want).

What I found most interesting about this part of her talk is not that she gives so much credit to babies, but that she gives so much credit to scientists.*  Her conception of a scientist as the independent open-minded experimentalist as opposed to a socially-constrained learner seems, to me, old-fashioned and at odds with the view Thomas Kuhn made popular with his 1962 book The Structure of Scientific Revolutions.  In the book, Kuhn argued that scientists largely accept the received views of their field and only when it becomes overwhelmingly apparent that accepted views have problems do these views change (something he called a “paradigm-shift”).  As summarized in Kuhn’s NYT obituary:

Professor Kuhn argued in the book that the typical scientist was not an objective, free thinker and skeptic. Rather, he was a somewhat conservative individual who accepted what he was taught and applied his knowledge to solving the problems that came before him.

To me, this sounds a lot like Dr. Gopnik’s depiction of a baby…

* – Though, honestly, I think that both scientists and babies are best served by a mixture of individual and social learning.