Much Ado About Francis Galton

Galton Dissed?

I was mystified by Douglas Medin’s aversion to Sir Francis Galton in his April Presidential Column (Medin, 2012). Medin said that Galton’s method of superimposing multiple headshots to study facial beauty “is not one of psychology’s success stories.” On the contrary, Galton’s method has inspired over 300 papers on using computer-averaging of headshots to study the roles of typicality, symmetry, and other geometric features in face perception and attractiveness, including several Nature papers by David Perrett and colleagues.

Recent research supports Galton’s suggestion that criminals tend to have some distinctive facial features: Alexander Todorov has shown that people do fast, automatic, amygdala-mediated evaluations of trustworthiness based on facial appearance, and several recent papers (e.g., Haselhuhn & Wong, 2011) show that people with certain facial geometries are more likely to deceive and exploit others.

Galton was cast as guilty by historical association with the more restrictive U.S. immigration policies of the 1920s and seemed to imply that the headshot is culturally oppressive in some analogous way. If Galton’s eugenics should make us wary of headshots as research stimuli, should we reject his other innovations: correlation, regression, scatterplots, percentiles, questionnaires, mental tests, sensory thresholds, reaction times, word association tests, mental imagery tests, twin studies, heritabilities, quantitative genetics, and the lexical approach to personality? What exactly would psychology have left?

Psychological science owes more to Galton than to any other researcher, and it is sad to see him dissed.

Geoffrey Miller
APS Fellow
University of New Mexico

Galton’s Techniques, Not His Values, Improved Science

My respected APS fellow member, Geoffrey Miller, takes strong exception to a recent Presidential Column, a footnote where I suggest that Galton’s composite photograph technique is not one of psychology’s success stories. He implies that I am blind to Galton’s many central contributions to our field. If truth be told, I am not a fan of the practice of using headshots, but I recognize that this is a losing battle.

I am in substantial agreement with Professor Miller with respect to Galton’s contributions. To his list, I would add that Galton recognized the importance of “regression to the mean,” which researchers ignore only at their peril. [1] The composite photograph technique is a precursor to modern work using morphing and averaging to address a whole range of issues in cognition and perception. [2] Galton’s work with this technique was circumscribed by both technological limitations and the common cultural practice of his day of men wearing beards. I’m not sure when I first read about Galton’s intriguing composite photograph technique, but I remember having a wonderful chat about it with Manny Donchin at the University of Illinois in 1977, when we discussed it as an empirical and conceptual tool for thinking about prototype theories of categorization.

Professor Miller is quite perceptive in detecting that my attitude towards Galton is less than uniformly positive. Although we could debate whether Darwin should be blamed for social Darwinism [3], in my opinion there is little doubt that Galton’s eugenics was closely tied to racist ideology and policies to favor higher social classes. His fawning [4] protégé Karl Pearson, another key figure in the development of correlation methods, in his biography of Galton, sums up the rationale for eugenics in this way:
“The garden of humanity is very full of weeds, nurture will never transform them into flowers; the eugenist calls upon the rulers of mankind to see that there shall be space in the garden, freed of weeds, for individuals and races of finer growth to develop with the full bloom possible to their species.” (Pearson, 1930, p. 220)

If you are inclined, take a look at Galton (1904) and come to your own conclusions. I think values often directly and indirectly shape what scientists do — and I do not think scientists can be excused from their values.

Douglas L. Medin
APS Immediate Past President
Northwestern University

Footnotes

[1] One example of this is studies that “match” scores for some cross-group comparison by selecting above average scores from one group [the group with a lower mean scores than the other group] and below average scores for the group with a higher mean. Because of regression to the mean, if the selected samples are simply retested, the sample from the lower scoring group should perform more poorly than the sample from the other group.
[2] Indeed my former student, Rob Goldstone, has used just these sorts of techniques very effectively in his own research (as have others), as a visit to his website will reveal. Forgive the nepotism.
[3] I’d be on Darwin’s side, but others might point to some of his “nonscientific” writings and make the case that he was far from innocent.
[4] If you doubt my use of this descriptor, read his discussion at the end of Galton (1904).