08 November, 2019

Historians, Fear Not the Psychologists



This week Jonathan Schulz, Duman Bahrami-Rad, Jonathan Beauchamp, and Joseph Henrich had their big piece on WEIRD psychology and the Catholic Church published in Science. [1] Long term readers will remember that I wrote about this piece in the American Conservative when the pre-print was published last year, and then wrote a critique of the Schulz-Henrich research program as a whole on this blog shortly after.[2] I still feel like that critique is one of the better things I have written here; in a different world I would be a grad student trying to turn that critique into its own set of research papers.

My critique withstanding, I see the "Origins of WIERD Psychology" as a landmark paper in the fields of economics, psychology, anthropology, and history. It deserves that status precisely because it is one of the few attempts to use data and theory from all four fields in one place. This is how social science should be done—and increasingly, I believe, how it will be done. I was very happy to see a version of it published in Science. 

Not everybody was so happy. The paper was met with outcry on twitter. "This is a pile of hot-trash" declared one; "[none] of the authors even bothered to read a history book or talk to a historian" inveighed another. "It’s also almost unbelievably ethnocentric — what it says about Europe is wrong," we read, "and what it implies about the world beyond Europe is equally wrong." Discussion of the piece has been a crazed tumult of the contrived ("I am just very very tired. SIGHS IN ALL CAPS), the snarky ("Historians of the family are a whole field. With books. And classes that you can take. And experts and everything"), the crusaders ("its time for a heavyweight institutional response"), and the righteously enraged ("holy shit am I angry").

Most of the outcry came from historians or those who would fancy themselves such. Strip away the emotional bombast and we are left with one essential critique: Schulz and company did not do their proper research. Their historical knowledge is too thin to support their claims. If they had read more books on the history of medieval Europe then of course they would recognize that the chasm between the letter of Catholic marriage laws and the reality was too large to support their thesis.

This notion that Schulz and company do not take history seriously is silly. The Science paper splits the references up between the main paper and the supplements (which none of the critics seem to have read!). Instead of bouncing between those two documents, I am going to refer to the pre-print instead, which keeps everything in one neat 174 page mega-paper. At the bottom of the pre-print we find 242 sources. By my informal count, 43 of these were written by historians. I might I have misjudged one or two of these. But I am talking about books with titles like Reordering Marriage and Society in Reformation Germany, The Transformation of a Religious Landscape: Medieval Southern Italy 850-1150, Marriage, Family, and Law in Medieval Europe and From Sappho to De Sade: Moments in the History of Sexuality. To these sources are another 20 or so works by historical economists ("Girl Power: The European Marriage Pattern and Labour Markets in the North Sea Region in the Late Medieval and Early Modern Period,"  Institutions and the Path to the Modern Economy: Lessons From Medieval Trade, etc), and a slightly smaller number by anthropologists of a historical or comparative bent (The Evolution Of Human Societies: From Foraging Group To Agrarian State, Explanation of Ideology: Family Structure and Social System, etc.). Rounding us out are a few "big think" titles by people like Jared Diamond, Francis Fukuyama, and Michael Mann.  The remaining 140~ sources are devoted to research in cross cultural psychology,  development economics, demography, and evolutionary anthropology.

Sweeping big history books celebrated in the discipline (think the style of William McNiell) will deal with topics like these peripherally, and do so in three or four pages citing six or seven sources. Most article-length treatments in historical journals have less than 40 sources. Schulz et. al. have met that standard. Their study includes regional surveys of medieval society in every part of Europe, a dozen Europe-wide surveys of the history of European marriage, sex, and law, and even a few primary sources. No, they have not dived into the archives for this project; none of them began their career with a 300 page dissertation on marriage records in a single Welsh parish. But historians do not demand that standard even from other historians who set out to write trans-regional surveys. Their bibliography is as good as you would expect to find in any work of comparative history—which in a way, this is.

More importantly, those finer grained historical arguments already exist. The underlying thesis of the paper (Western individualism—"WEIRDness"—was an outgrowth of changes in the medieval family) was developed by a medievalist!  That theory was set forth by historian Michael Mitterauer in various journal articles a decade and a half ago.[3] Mitterauer expanded on these ideas his 2010 book Why Europe? The Medieval Origins of Its Special Path? To Mitterauer's work we could add books by anthropologists Emmanuel Todd and Alan MacFarlane, neither of whom were strangers to archives; their books also argued that family structure was key to understanding the individualist orientations of Western Europeans.[4]

Thus none of the underlying ideas here are novel. Schulz and company are not advancing some bold new thesis: they are testing an old one. The worst they have done is take an already existing theory in the field and asked, "How could we test this with statistical methods and experimental data? Does the Mitterauer-Todd-MacFarlane thesis predict actual cross cultural psychological variation today? If so, how would we know?" Their solution to this problem is clever and worth reading in full (including the supplements!).

That is all Schultz et. al. have done. They have tested and refined an existing thesis with new methods (and show how their findings accord with theories in economics and psychology). But why then, all this outcry?  Why the histrionics?

I have a charitable hypothesis for the historians' behavior and a less charitable one.

The charitable hypothesis is that many historians are reacting this way because of the paper's venue. This has been published in Science. Like most work in Science it stuffs all the caveats and sourcing into the supplements; the smaller summary brims with all the confidence of numbers and empirical fact. Mitterauer's theory was bold but contested. His work has thirty contenders in the "great divergence" literature. Among medievalists, his take on medieval family life has not attained general consensus. But here we see it proclaimed as SCIENCE. Perish the day one stray and contested medievalist theory becomes enshrined as scientific fact!

This reaction is understandable. We live in an age when 16 year old girls wield phrases like "unite behind the science" as a rhetorical bludgeon. In our world, "the science" has immense cultural authority. However, being published in Science is not the same thing as being a part of "the science." Tally up the number of studies in a newly emerging field (say, historical population genetics) that were published in Science only to have their interpretations of the data overturned a few years later. There are a lot!

The same is true for the work of our authors. Take the career of a prolific social scientist like Joseph Henrich (who is one of the co-authors of this paper). Lay out his publications out in one place, and you will see titles like Science, Nature, Behavioral and Brain Science, and Proceedings of the Royal Society B: Biological Sciences. But if you are familiar with the topics he researches you quickly realize that almost everything he has ever published has been at the center of some scholarly tussle. How universal are psychological experiments first tested on American university students? Can Darwinian models of change and inheritance be used to understand learned culture? When did  organized religion emerge, and has it had "pro-social" effects on the societies that adopted it? After two decades of fighting, Henrich has more or less won the battle on the first topic. But the other two topics are still very much live debates. And where do the people who disagree with Henrich and his fellows publish their critiques? They are published in.... Science, Nature, Behavioral and Brain Sciences, and so forth. Using statistics and being published in Science is not the same thing as being enshrined as part of THE SCIENCE. The advance of science is a story of model building, empirical challenge, debate, more data brought to bear on the question, and yet more model building until a consensus has been arrived at. We are in the beginning stages of this process: If other studies martial evidence or put together models that puts the Mitterauer thesis to doubt then this one will eventually fade away.

So there is no need for the upset. Unless the upset is about something else—which it might be. My less charitable hypothesis goes like this: When Mitterauer or another historian advances a historical thesis it is simply "one more contribution" to an argument. When a coalition of psychologists, economists, and evolutionary anthropologists advance a historical argument it is a threat.

It is jokingly said that too many proud physicists parachute out of their own discipline into others, convinced that all who research lesser subjects have lesser brains. I understand hostility towards the economist or evolutionary psychologist version of the same problem. Many of the social scientists who do this know nothing about the history they reduce to an equation. They deserve all of the hostility and ridicule that they get. It is frustrating to watch arrogant scholars try to colonize a field they have no knowledge of, especially when you have spent decades mastering its intricacies.

Yet historians who expect their field to remain un-colonized are blind. One of the most remarkable academic developments of the last twenty years is the slow blurring of what were once much tighter academic distinctions. I regularly see scholarly debates that draw in participants and ideas from the worlds of micro-economics, ethnology, neuroscience, evolutionary anthropology, cognitive science, sociology, genetics, computational social science, and psychology.  On a different scale I also see work that mixes freely from epidemiology, archeology, historical linguistics, climate science, biological anthropology, and population genetics, and network science. These fields have all 'colonized' each other. "Multi disciplinary" is often a gimmick. These emerging fields have achieved much more than that. [6] 

The work of historians is and will remain an important part of these debates. I do not know if the same thing can be said for the historians themselves. The essential problem was stated well by the pseudo-anonymous economic history blogger Pseudo-erasmus a few years back:
Cultural-social historians are ill-equipped for the age of “Big Data” that Guldi drones on about, but not because they are intellectually incapable. They can get trained in quantitative techniques and actually understand the various interdisciplinary debates that are mostly impenetrable to them right now. But such training would actually change who they are. It’s the historians’ hermeneutical and subjectivist instincts that alienate them from the big empirical debates amongst economists, psychologists, evolutionary biologists, climatologists, anthropologists, archaeologists, geneticists, etc. So the problem with historians is less any microhistorical preference, than an epistemological bias against positivism.[7]
Schultz and company are threatening because they use methods impenetrable to most historians and embody a positivist attitude uncongenial to these historians' broader beliefs. To receive Schultz et. al's work as a legitimate entry in the debate—as historians regularly do with the work of colleagues that they disagree with—would mean conceding that the methods psychologists and economists use to understand the world may be just as  useful for understanding medieval times as a trip to the archives. It would be mean recognizing the  importance of statistical literacy, and more terrifying still, accepting that the "subjectivist instincts" which rule so much of the history profession may be inadequate for answering the kind of questions social scientists may ask of them.

But the truth is that an economist and a psychologist do have methods that may be useful for understanding the medieval world. Medievalists should welcome the contributions of social scientists with the same warmness that Schulz and company embraced the research of historians. A past generation of empirical psychologists and economists would not have read 40 tomes on Medieval European society to inform their research. But they have—and created a worthy advance in the literature by doing so. They are not afraid of allowing historians into their debates. There is no reason for the historians to fear the reverse.

---------------------------------------------------------------------------------------
This post discusses both history and behavioral science. If you liked it for the history, you might be interested in some of the popular things I have written on that topic: "History is Written by the Losers," "Chinese Strategic Tradition: A Research Program," "Making Sense of Chinese History," or "A Study Guide For Understanding Human Society." On the other hand, I have also written about psychology: see "Public Opinion in Authoritarian States," "Taking Cross Cultural Psychology Seriously," "So Why Did They Publish Them?", and "Tradition is Smarter Than You Are." To get updates on new posts published at the Scholar's Stage, you can join the Scholar's Stage mailing list, follow my twitter feed, or support my writing through Patreon. Your support makes this blog possible.
             ---------------------------------------------------------------------------------------


[1] Jonathan Schulz, Duman Bahrami-Rad, Jonathan Beauchamp, and Joseph Henric, "The Church, intensive kinship, and global psychological variation,"Science, Vol. 366, Issue 6466 (2019).

[2] Tanner Greer, “How the Catholic Church Created Our Liberal World,The American Conservative (17 December 2018); "Taking Cross Cultural Psychology Seriously," The Scholar's Stage (21 Dec 2018).


[3] Michael Mitterauer, Why Europe? The Medieval Origins of its Special Path, trans. Gerald Chapple (Chicago: University of Chicago Press, 2010).


[4]Emmnauel Todd, The Explanation of Ideology: Family Structure and Social Systems (Blackwell Publishers, 1989); Alan MacFarlane, The Origins of English Individualism: Family Property and Social Transition (Hobeken, NJ: Wiley-Blackwell, 1991).


[6] To make this more concrete: think of the research programs of Joseph Henrich, Walter Scheidal, Peter Turchin, and Ara Norenzayan. There are many more scholars I could add to this list, but I this is sufficient for the argument.  

[7] Pseudoerasmus, "La Long Duree Puree," Pseudoerasmus (11 November 2014)

8 comments:

Matt said...

To some extent, I could be charitable and take it as that twitter encourages extremely short, poorly articulated responses to anything. Limitations of format!

Certainly the historians should go at the data though, and there's lots to go at:

1) For their largest and most robust comparison, of KII (Kin Intensity Index) vs Individualism, the pattern describes a decreased variance in Individualism-Collectivism, tending towards collectivism in high KII societies, not a simple shift in mean with no change in variance. Essentially it shows there are no high Individualism, high KII societies, but does not show that low KII societies are necessarily Individualist (e.g. Fig3a in the preprint, which is a bit more transparent about the actual societies involved).

The figure 2 in the published paper also makes it clear through its shading that the increased variation in Individualism in low KII societies is driven by a handful of countries with high length of time exposure to Western Church, which spatially cluster.
Suggesting that KII is actually redundant here and a pure geographical effect is more explanatory. (Otherwise, effectively "KII causes individualism.... but only in countries exposed to Western Church which happen to be geographically close to each other... but it's not any other shared cultural effect between these close countries, post-dating or pre-dating the church, it's definitely the Western Church's family program!" Hmm....).

2) The use of a log cousin marriage variable means that if there are small inaccuracies within standard errors in placing societies at 0.5% which are actually at 1.5%, or 3% when actually at 2%, then their regression will change. The cousin marriage comparisons are also relatively low powered in the pre-print, somewhat improved in the print, but seemingly saliently losing a lot of the low KII but collectivist dots (judging by their figure showing countries with reported cousin marriage rates).

A lot of their regressions are under-powered (Tightness / Asch Conformity for one) and have salient outliers (e.g. France on Asch Conformity) or not a lot of data points at one or the other end of their KII / log cousin marriage axis. Others, like Creativity are probably somewhat dubious as variables (Is there some meaningful way which is objective to response bias for Ghana to be at the low end of the Creativity scale and Uruguay at the high end? Or is it subjectivity in test/question framing all the way down?).

Another bigger methodological problem is that they have no means to show whether cousin marriage really declined *when* they believe it did and not more recently (as a consequence of urbanization etc), which is particularly a problem given that they seem to argue that the decisive changes in cousin marriage are on the log scaled, e.g. from 5% to 2.5% or 1% to 0.25%, not from 10% to 40%, etc. (This can be tackled with adna but taking seriously the marginal changes they seem to believe have had very large effects, probably only with very large sample sizes which may be infeasible to get robust to stratification).

Nor can they show whether changes in KII or cousin marriage directly preceded changes in individualism or those are better explained by pre- and post- changes in affluence, occupation, etc. In a sense they can't really do this, as you can't do opinion surveys on ancient, dead people who aren't around to ask. You can't really do psyhcology on the dead. (The best that can be done is these sort of sentiment meta-analyses of texts, which are fairly weakly correlated with actual opinion polls of sentiments in living people.)

Finally, should their sampling really be effectively biased to national level units rather than sheer population size? Should China really get one dot while multiple Western European countries split into different political jurisdictions get a fairly large number of dots?

Anonymous said...

I dare say there is even a less charitable way to look at it. In this case, the subjective instincts of the historians are really the product of their liberal arts education (indoctrination). The rhetoric of the outraged has made it beyond clear that their primary reason for the outrage is that the research paper has implicitly elevated the wrong kind of people. There were some studies (https://righteousmind.com/liberals-are-weirder-than-conservatives/) suggesting that liberalism is linked with WEIRD mentality. Added to the idea of western WEIRDness connection to Catholic Church actually has a genuinely dangerous implication: liberalism is a local phenomenon tied in particular time and place and its eventual victory is not guaranteed. But to our good liberal medievalists, who can even feel guilt for using the term "Anglo-Saxon",
the idea that liberalism and modernity are particularly linked to the West (the very Great Satan against whom the liberalism supposedly emerged) is beyond repugnant. They might as well be told they need to reconcil with their parents' murderers.
That said, look at the way our liberal intelligentsia--the WEIRDest of the WEIRDos--frantically apologize for their own WEIRDness, maybe WEIRD mentality is not so advantageous after all.

Lorenzo said...

Fantastic post. And, I agree, your previous post on cultural psychology and history is one of the best things you have done, which makes it a very fine thing indeed.

On another matter, I have invoked you against the view" that we just have to wait, markets and commerce will tame China".
https://www.econlib.org/politics-is-the-problem-trade-is-the-answer/

roundeye said...

I would like to see the fertility rate difference between WEIRD v non-WEIRD in the first world. We know in females in the biobank samples the more loci for EA means less fertility (and we can see that with our own eyes).

Dors said...

quote from this article: I regularly see scholarly debates that draw in participants and ideas from the worlds of micro-economics, ethnology, neuroscience, evolutionary anthropology, cognitive science, sociology, genetics, computational social science, and psychology.

I do not see them, and can't even think of where I'd be likely to find them. A pointer or two would be welcome.

T. Greer said...

Dors--

Check out any of the books on 'cultural evolution' or 'cultural epidemiology' to start. Take this paper and responses (as well as the f300+ citations it has racked up) as an example: https://www2.psych.ubc.ca/~ara/Manuscripts/2016%20Norenzayan%20cultural%20evolution.pdf

Or another recent paper: https://www.sciencedirect.com/science/article/abs/pii/S1090513817303239

The first paper has two anthropologists, a historian, and four psychologists as lead authors and cites things from all over the place; the second is balanced between psychologists and anthropologists. These researchers often have oddly interdisciplinary backgrounds themselves. Consider Quentin Atkinson:

Before coming to the University of Auckland in 2010 Quentin was a research fellow in the Institute of Cognitive and Evolutionary Anthropology at the University of Oxford and a Fellow of Wolfson College. He has also held research positions in the School for the Study of Religion at the University of Aarhus, Denmark, and with Professor Mark Pagel in the Zoology Department at the University of Reading, UK. Quentin teaches stage III and postgraduate courses on Evolutionary Psychology and in 2011 he started a new postgraduate course with Dr Niki Harré on Sustainability and Psychology

Here then we have an author who has held positions in Evolutionary Anthropology, Zoology, Religious Studies, and Psychology!

Harry Jones said...

Isegoria linked on this, and the discussion went all over the place, which is fine by me, but...

Here's the larger issue I see: people will be unreasonable and unfair when you threaten their narcissistic supply and their turf at the same time. There is a point of diminishing returns when it comes to refuting them. At a certain point you just need to give your critics the blowoff and get on with your thing.

Clipping Path said...




Yours article is very helpful. Thank you for providing this type of information.