Monday, June 13, 2011

Did Stephen J. Gould Fudge The Data ?

I will assume that most archaeologists have read some works by the late biologist and public intellectual Stephen J. Gould. His works on history and evolution (Gould 1986, 1987, and 1989) are of fundamental importance for those who (like me) consider archaeology as a historical science. History, archaeology, geology, paleontology, astronomical cosmology, and other historical sciences share a number of conceptual problems concerning how we can understand what happened in the distant past. Reading what Gould says about how these problems have been addressed in geology and paleontology helps archaeologists think about our own methods and data. And while I am on this theme, Darwin's book on worms is also of great use here (Darwin wondered why ancient Roman ruins became buried, and did fieldwork to investigate the actions of worms in the burial of old ruins -- great stuff!!), and Gould's introduction to that book is very good also. Don't forget Toulmin and Goodfield (1965), another important work in this area. And by coincidence, today I am wearing a T-shirt with Darwin's portrait, wearing a Che Guevara-style beret, with the logo, "Viva la evoluciĆ³n!"

Anyway, back to Gould possibly fudging data. In Mismeasure of Man, Gould uses Samuel Morton's measurements of human cranial capacity (in the 19th century) as an example of how a scientist's personal bias can influence his work and skew one's results. Morton wanted to show that Causians had the biggest brains, and Gould claims to show that this desire led to subtle "fudging" of Morton's data in favor of his preconceptions. Even since 1981, this has become a "textbook example" of bias in research. Well, now some anthropologists have gone back and remeasured a bunch of Morton's skulls at the University of Pennsylvania Museum, and re-analyzed Morton's and Gould's data. They concluded that Morton's results were NOT biased or skewed, but that Gould's analyses were biased, with some aspects evidently made up. Wow, this is disheartening -- Stephen Jay Gould has long been one of my intellectual heroes.

The article is:

Lewis, Jason E., David DeGusta, Marc R. Meyer, Janet M. Monge, Alan E. Mann, and Ralph L. Holloway
    2011    The Mismeasure of Science: Stephen Jay Gould versus Samuel George Morton on Skulls and Bias. PLOS Biology 6(6):1-6.  .


The University Museum is using the study to generate publicity. There is information on their website, including a short video about Morton and his collection. The video, narrated by Janet Monge, does not have the best discussion of the collection and the new research, however.

For a nice summary of the new article, see John Hawks' blog.

References:

Darwin, Charles
1985    The Formation of Vegetable Mould, Through the Action of Worms, with Observations on Their Habits. Orig. pub. 1881 ed. University of Chicago Press, Chicago.

Gould, Stephen Jay

1981    The Mismeasure of Man. Norton, New York.

1986    Evolution and the Triumph of Homology, or Why History Matters. American Scientist 74:60-69.

1987    Time's Arrow, Time's Cycle: Myth and Metaphor in the Discovery of Geological Time. Harvard University Press, Cambridge, MA.

1989    Wonderful Life: The Burgess Shale and the Nature of History. Norton, New York.

Toulmin, Stephen and June Goodfield
1965    The Discovery of Time. University of Chicago Press, Chicago.

10 comments:

Ryan Anderson said...

"Wow, this is disheartening -- Stephen Jay Gould has long been one of my intellectual heroes."

I am right there with you. Gould has always been one of my favorites, especially his collections of essays.

Jason Antrosio said...

Thank you for this nice summary of both the study and Gould's contributions. This reassessment of Gould is indeed disheartening, especially as it will inevitably be used to support the race-resurgence. I've just posted about this as "Mismeasuring Gould": http://www.livinganthropologically.com/2011/06/14/mismeasuring-gould/

RP said...

Jason, I just checked out your webpage/blog and found a lot to like. I share your concern that this research will be misused to promote racist ideologies. Your foregrounding of this concern is valuable in making me consider how to best combat this.

On your webpage, I especially like how you have challenged anthropologists to get involved in more contemporary debates rather than sticking to their old and easy references from the 70s and 80s. I see our disappointment at Gould in this light. That is, we share some guilt in building the anthropological case on such infirm ground. We should have discovered these problems earlier, rather than accepting them because they were convenient and building so much up on them.

After reading your comment, and some of your webpage, I am curious about the ‘race resurgence’ you refer to. Do you mean this to refer simply to the debates that have been occurring after the 70s and 80s which we are guilty of not engaging enough in? Or are you referring to a greater resurgence of racism?

Chris said...

I find this article troubling in its science is always apolitical vision. I am a scientist and an ardent materialist, but I am led to wonder, why take on this study? What are the authors' owns biases here if, as the recognize, scientists are biased but science is not. Moreover, even if science as an ideal is apolitical, its results, driven by the interests of interested scientists, certainly are not.

Also, I do not think there is any question that Morton was practicing shabby science, a point the authors, for one reason or another, seem to resist, especially if some of Gould's errors were based on not recognizing Morton's. Mostly, however, Morton is working without the notion of clines and distributional variation of traits. This goes back to a well-know critique of the race concept made by Boas and some of his students, especially Alain Lock: races cannot be said to exist if there is more intra-group variation than inter-group variation. What this means is that Morton pre-defined his groups based on social and racial biased of his times rather than try and document groupings in the data themselves. This leads me to ask, unless I missed something in this article, where are the reportings of variance and deviation? Funny how the authors do not focus on this issue whatsoever. The only place I can see discussion of deviation is between using seed and shot to measure cc. At no place in this article do the authors use this issue to question the analytical framework that drove Morton's study. In fact, for me, that is one of the major points to critique 19th century bioanth for students. In this regard, the authors' assertions against biased results are frankly disheartening to the point of being sickening. They write "The Morton case, rather than illustrating the ubiquity of bias, instead shows the ability of science to escape the bounds and blinders of cultural contexts." Really?

I keep asking myself, why this article? Why this study? Why, why, why? What is more important, tearing down the work of Gould or providing a better, more solid critique of the work of Morton? One would think the latter, no?

Anonymous said...

Chris,

Having spent a lot of time thinking about the notion that, "races cannot be said to exist if there is more intra-group variation than inter-group variation", I think we anthropologists tend to too readily accept this idea and repeat it to each other. I usually see this idea attributed to Lewontin and will have to look into Alain Lock's take on the topic (do you recommend starting with any particular work of his?). While having more variance in a trait or gene within a group than between groups suggests that factors within the group are more important in explaining variation within the group than factors that differ across groups, I've (so far) found no good basis to clearly equate this to a statement about race.

One good review of where this skepticism comes from is:

Pigliucci, M. and J. Kaplan (2003). "On the concept of biological race and its applicability to humans." Philosophy of Science 70(5): 1161-1172.

I think one of the issues is that race is a slippery word, with different meanings in different contexts and for different people. As such, I worry when we too often make blanket statements about what race is and isn't without explaining how we happen to operationalize the term race.

I'm still thinking through what ignoring pre-defined groups based on social categories and documenting groupings in the data themselves would look like in contemporary studies of organismal variation. I suppose it depends on ones particular research question. Although little appreciated or grappled with by most anthropologists, when using principal components analysis on human genetic data from people around the world, continental origins are very easily discernible, even though PCA doesn't consider these groups in coming to its results.

However, for the same reason talking about variance within versus between groups is misleading, talking about groupings from principal components can also be misleading. While both might tell us something about phylogeny (one way race is defined), neither tell us a lot about phenotypic differences between groups or what their causes are.

RP

Chris said...

I imagine that starting with the data themselves would show a great deal of variation overlapping the racial groups Morton pre-defined, thereby highlighting the social nature of his categories. Would Morton not be guilty of circular analytical reasoning: if one pre-defines a racial group and then uses data within each group to define key differences between the groups?

I imagine that idea has spilled into my brain via Lewontin at some point, but I tend to remember it from a book 100 year old essays of Alain Locke's I read as an undergrad (which I never sold back to the bookstore and happen to have handy), entitled Race Contacts and Relations. In it, he directly makes reference to the social nature of race and the circular reasoning of early biological anthropologists:

"the variations between individuals of the same race...more than outspan the maximum variability between what are regarded as cognate races of mankind. To put it in terms of a concrete example, one can find more variability in the anthropological differences between one class of Frenchmen and another class of Frenchmen than when you take an average European and compare him with an average African or Malay...So the point in the modern view is that, because there is so considerable a degree of variation within a race group, many anthropological differences between the races are lost. And the older view of the racial significance of anthropological differences must give place, if anthropological science is to progress, to simply the description of anthropological factors as clues to the environmental and biological influences which have been at work during long periods of history upon this or that group of people (5-6)"

Of course, this is old hat for those who feel that race is entirely a social construct and an example of humans' ability to classify gone insane with power. But I think it is remarkable still, as it is from an essay that is 100 years old this year. But, to go on, he has some nice things to say also about how this causes analytical problems:

"For that reason we can believe that, in spite of the accuracy of its methods, such science is largely pseudo-scientific both in its postulates and in its conclusions. It has set out, therefore, to prove something which has already been made a basic assumption of the science in question" (3).

None of this, I think, brings up the question of why the authors of this study chose this study? They say that scientists are biased but science's results are not...or at least Morton's were not. Perhaps, as Locke says "in spite of the accuracy of its methods." I still hold to my guns on questioning why they chose to re-evaluate Morton's work this way rather than do a better job then Gould. Even if science's results are not political and biased (and I am unconvinced the authors demonstrated Morton was free of bias), the political uses of science's results certainly are biased. I am just waiting for their article to be linked on David Duke's blog.

Jason Antrosio said...

Chris writes, "I am just waiting for their article to be linked on David Duke's blog."

Thank you, Chris, and you don't have to wait. Fire up the browser, go to Google and type in "Mismeasure of Science" (include the quotes so the search results refer directly to the latest article). At least four of the top ten results are to racist, white supremacist, or anti-immigrant blogs. On my website, one of the co-authors said, "the moral of the story: science is and must continue to be a self-correcting organism." Does he really think that this "moral of the story" is getting out there?

Keep up the good work,
Jason

Chris said...

Jason, thanks for showing how fast the fascists flock to research like this. Hopefully anthropologists and other researchers will not let the question of "why this study" be buried under the supposedly non-biased nature of science. As a materialist and a scientist, saying I am disappointed in how such ethically problematic research is undertaken without interrogating the bigger picture is an understatement.

CM

Anonymous said...

Chris,

Given your concern about not acknowledging variation within and between group, I think this paragraph in the piece is particularly important:

"In reevaluating Morton and Gould, we do not dispute that racist views were unfortunately common in 19th-century science [6] or that bias has inappropriately
influenced research in some cases [16]. Furthermore, studies have demonstrated that modern human variation is generally continuous, rather than discrete or ‘‘racial,’’ and that most variation in modern humans is within, rather than between,
populations [11,17]. In particular, cranial
capacity variation in human populations
appears to be largely a function of climate,
so, for example, the full range of average
capacities is seen in Native American
groups, as they historically occupied the
full range of latitudes [18]. It is thus with
substantial reluctance that we use various
racial labels, but it is impossible to discuss
Morton and Gould’s work without using
the terms they employed."

I realize this doesn't address all of your concerns, especially why a better analysis wasn't conducted. I hope the authors are able to make other better contributions to the literature as well.

If the choice was between Lewis et al publishing this imperfect and incomplete piece versus anthropologists and others continuing to build our scholarship around false allegations of Gould..which would you prefer?

RP

Anonymous said...

Funnily enough, one of the authors of this latest study (DeGusta) had previously written another one which was pretty much pro-Gould on the same issue exactly, defending his conclusions (not to the fullest extent though) from an earlier criticism.

http://www.stanford.edu/~lewisjas/Morton_files/Michael%20Analysis.pdf

Or on google cache:

http://scholar.googleusercontent.com/scholar?q=cache:0dltg_cDfg0J:scholar.google.com/&hl=en&as_sdt=0,5&sciodt=0,5



Quoting the conclusion:

"[...] Michael’srecalculations [1] of the Morton means [4] are of questionable value given theincongruent samples but, overlooking that, reveal the racial pattern of errors expected byGould [7], contra Michael [1]. Michael’s defense of Morton against Gould’s claimsoverlooks the most relevant charges made by Gould. While Michael is correct in pointingout the biological invalidity of the race concept as applied to humans, this does not renderMorton’s data invalid, or trump Gould’s criticisms of Morton. [...]"

This guy is luck he didn't day and wasn't famous, having time to reevaluate his own findings, otherwise he would have been just another one in this post-mortem lynching, as Gould's "accomplice" in hiding "politically incorrect scientific truths".