In a moderate tone that belies the seriousness of their allegations, a group of anthropologists last week made the case that the giant of evolutionary biology, historian of science, much-loved author and Harvard professor Stephen Jay Gould, had falsified criticisms of skull measurements originally reported by 19th century American physician Samuel Morton.
Like many 19th century intellectuals, Morton held repugnant racist views. But in a 1978 paper in Science and later in his book Mismeasure of Man, Gould, who died in 2002, also cited Morton’s work as a classic example of bias influencing scientific results. Gould wrote in Science that Morton’s measurements were subjective and his data finagled to give the answer he wanted; that the cranial capacity of Caucasians is larger than that of other worldwide populations represented in his unique collection of 1000 skulls. Contrary to biologists today, Morton posited a straightforward correlation between brain size and intelligence.
Gould’s Science paper stopped short of accusing Morton of deliberate fraud, and Gould’s critics, who published a paper June 7 in PLoS Biology, phrase their concerns about Gould’s work with similar caution. Perhaps because of this civil approach, the paper hasn’t yet prompted the widespread attention that it probably deserves, given Gould’s stature (although evolution blogger John Hawks has jumped in to express his outrage at what he terms “perfidy” by Gould).
Among the claims by anthropologist Jason Lewis of Stanford University and his colleagues are:
-Gould was wrong when he claimed that Morton’s measurements of skull capacity by packing seeds were subconsciously influenced by his racial prejudice. In fact, Morton’s measurements show no statistically significant deviation in the direction of his hypotheses when compared to independent measurements that Lewis and his colleagues’ have now carried out using Morton’s own samples, which are located at the University of Pennsylvania’s Museum of Archaeology and Anthropology in Philadelphia.
– Gould’s calculation of average skull size for Native American populations, and his statement that Morton wishfully underestimated this value because of his racial bias, were unduly influenced by Gould’s decision to exclude 34 skulls out of a sample of 144, on the basis that they came from populations with less than 4 skulls per population. When the omitted data are restored, the average calculated using Gould’s preferred method is actually slightly lower than the average calculated using Morton’s method.
– While Gould alleged Morton biased his results by breaking down some populations into subsamples while amalgamating others, a detailed review of Morton’s data shows that he did report the subsample values. Furthermore, Gould himself misdefined the Native American samples which served erroneously to inflate the mean he calculated.
While the paper stops short of accusing Gould of fraud, it does say “most of Gould’s criticisms are falsified or poorly supported” and one author on the paper, physical anthropologist Ralph Holloway of Columbia University in New York City, adds that he believes Gould’s alleged misstatements were intentional, that is, motivated by the case he was trying to build against Morton. “I don’t think it was an innocent mistake,” he says. Holloway notes that there is no known link between brain size and intelligence. That appears to make the data essentially irrelevant for Morton’s racist purposes.
David Wake, a biologist at the University of California, Berkeley, who has written about Gould’s legacy, says “the authors do make a pretty convincing case that if there was any data manipulation [by Morton] it seems to have been minor and perhaps inconsequential.” He adds that he thinks Gould would have appreciated one of the authors’ discoveries; that several proof copies of Morton’s book Crania Americana contain corrections of a typographical error in one of Morton’s tables that Gould had claimed Morton failed to check because it supported his prejudice.
Ian Tattersall, a curator at the American Museum of Natural History who knew Gould, says he has no idea how he would have responded to the PLoS paper, “though I have no doubt he would had a ready answer.” He adds that if Gould really did bias his own analysis of Morton’s data, then he proved his own point; that science is susceptible to unconscious bias. Tattersall says that cranial capacities are notoriously variable within human populations, and hard to correlate with any other quality, so it is difficult to say what ordering populations by brain size would mean, even if it could be done.
Lewis and his colleagues point out, and even Gould briefly acknowledges in his Science paper, Morton’s careful reporting of his raw data and methods, with no attempt to “cover his tracks,” enables later scientists to dispute his interpretation, or, as in this case, defend it. That may suggest Morton’s work is closer to objective science than, as Gould charged, the opposite.
Image: Stephen Jay Gould/ Jon Chase, Harvard News Office