1.3 History as a Beaker of Mustard Seed 3 — A hall of great skulls

Posted on January 31, 2011


He set himself the job of measuring the interiors of skulls.  You would think it was straight-forward.  Yet somehow Samuel George Morton, rising with a bullet on the hit-parade of nineteenth century American scientists, messed it up utterly.

As noted previously.

And, looked at forensically, his data had the fingerprints of his own peculiar brand of scientific racism all over it.

What Morton failed to do, and which science says he should have done in order for his data to be scientifically defensible, was to create a consistent and objectively defensible set of criteria—applicable to all the groups he was studying—for deciding which skulls should be measured and which skulls should not.  Unless this was done, his conclusions could be too easily skewed by cherry-picking data, by engaging in a sort of statistical gerrymandering.

Political gerrymandering refers to the practice of manipulating political boundaries to ensure a government’s re-election.[1] Statistical gerrymandering, as I discuss it here, refers to a process of manipulating the input of data to support certain favoured outcomes.

As an example, let us say I want to prove that people in my city are taller than people in your city.  The sample group I generate from my city is 60% male and includes the rosters of every college and university basketball team within the city boundaries.  The sample group I generate from your city is 60% female and includes the full membership of the local jockey club as well as all the little darlings from Mrs. Semple’s kindergarten class.  Using these sample groups I ‘prove’ my point:  the people in my city are taller than the people in your city.

And is Jeff really taller than Mutt?

But that’s not fair, you say.  Men are usually taller than women.  Basketball players are almost inevitably taller than jockeys.  And why are there children in our sample group but not in yours?

Good point.  It isn’t fair.  The two groups really aren’t comparable.   But that was pretty much the case with the sample groups of skulls that Samuel George Morton used as well, when compiling his comparative tables of measurements.

As an example, Morton emphasized the largest skulls in his Caucasian sample group and de-emphasized the smallest ones, while following the opposite policy with his Indigenous American sample group.

When he removed 14 out of 17 “Hindoo” skulls from his Caucasian sample, he explained that he did so because their size was disproportionately small, and would have distorted the averages downward.  The three skulls he kept represented something over six percent of the final Caucasian sample.

The Inca skulls in Morton’s collection were almost identical in size to the South Asian, equally small, and, one would think, equally ‘anomalous’.  He included a generous number of Inca skulls in his final Indigenous American sample, 45% as compared to 6%.

Morton was apparently unconcerned with the downward effect this would have on the Indigenous American averages.

His heads I win, tails you lose policy was almost too successful in this particular case.  It brought the Indigenous American average down so low that it then stood perilously close to the African average.

Remember that Morton had certain theories that he wanted to prove.  He wasn’t only concerned with Caucasian superiority.  He wanted to demonstrate African inferiority as well.  He wanted his various ‘races’ to fit into a neat racial hierarchy with Caucasians on top, and the other groups arrayed below in proper order.  It didn’t suit his social and racial esthetics to have Indigenous Americans occupying the same bottom step that had been reserved for Africans.

So here, to the benefit of his thesis, Morton made an error in arithmetic—in favour of Indigenous Americans.  The error had the affect of bringing their average up by 1.8 cubic inches, and removing it a reassuring distance away from the lowly African average. (Which itself was the result of creative, if unconscious, manipulation.)

Morton managed and massaged his data in other ways.

His mathematical mistakes and inconsistencies were few, but in every case they favoured his personal theories.  He tended to round averages upwards for Caucasians and round them downwards for other groups.  In one case he abandoned logic altogether.

In his study of skulls found in Egyptian tombs (1844), Morton reported that the skulls of Greeks were significantly larger than those of Semites.  However, he had no way of really telling whether a skull was Greek or Semitic.  If a skull was large, he labeled it Greek on that basis alone.  Given such circular reasoning, all he was really proving—aside from his own prejudices—was that large skulls are larger than small skulls, a conclusion, I would suggest, of only limited long-term scientific significance.[2]

Morton also had a tendency to highlight findings that fit his social theories, and ignore the findings that contradicted them.

Morton’s hierarchical worldview placed Teutonic people—Germans, Anglo-Saxons, and so on—above everybody else, including other Caucasians.  He had in his collection an all-male (and thus unrepresentative) sampling of good-sized Anglo-Saxon skulls.  Their size was to Morton a welcome proof of Teutonic superiority, so he highlighted this information in his published charts under the label “Modern Caucasian Group–Teutonic Family–English.”

However, there were other equally impressive examples of large skulls among his collection of non-Caucasian groups, the Inuit and the Iroquois, for example, which Morton passed over without comment.  He was not interested in presenting evidence that peoples he regarded as “uncivilized” had skulls as large as Englishmen.  So there were no matching labels “American Group—Barbarous Tribes—Iroquois” or “American Group—Barbarous Tribes—Inuit.”[3]

Another thing was clear about Morton’s conclusions:  he never once entertained any explanations for his data except the ones he began with.  As a consequence he neglected to examine of any of the other variables.

Morton failed to do things as elementary as separating data by gender or comparing skull size to body height.  If he had done this, he would have seen that tall people on the average have big skulls, and small people have small skulls.  This simple correlation would also explain the difference in skull sizes between men and women—men as a gender are simply taller and larger than women.

So much for Morton’s theory about the relationship of skull-size to intelligence.

He would have had a hard time selling the notion that tall people were smarter than short people, which, if brain size really indicated intelligence, would have been his inescapable conclusion.

In fact, if there were that strong a correlation between brain size and intelligence, we would not be going to the great universities and places of learning to seek out answers to the world’s problems.

We would be going to the National Basketball Association.

We could also expect Andre the Giant’s portrait to have an honoured place in the Gallery of Great Minds, down the way from Jane Austen and Johannes Kepler, somewhere on the wall between Albert Einstein and Wilt “the Stilt” Chamberlain.

The story of Samuel George Morton has a postscript.

In 1873, Dr. Paul Broca, professor, surgeon and founder of the Anthropological Society of Paris, published a paper analyzing in detail Samuel George Morton’s method for measuring human skulls.  Broca himself was a careful scientist who admired and improved on Morton’s techniques.  Of Broca, Gould stated, “I believe his numbers and doubt that any better have ever been obtained.”  But despite Broca’s close and careful scrutiny of Morton’s methodology, a scrutiny attested to in a published document a hundred pages long, he failed to notice any of Morton’s errors, not even his errors of arithmetic.

Broca’s blindness, it seems, derived from the fact that he generally shared in Morton’s conclusions and assumptions, and so failed to challenge any of the evidence in support of them.  Without scientific or social motivation, he simply neglected to ask the right questions.  Unbiased scientific scrutiny of Morton’s data would have to await another day, a day when the standard social and scientific model of humanity did not accept European superiority as the proverbial truth.

More than we like to admit, we see what we expect to see.  This is no more than a sociological truism.  The parable of Dr. Morton shows just how much, and how embarrassingly, theory can shape data.

[1] Political gerrymandering could operate, for instance, when a party up for re-election finds that a marginal district borders on a district certain to fall into the hands of the opposition party.  The marginal district can be transformed into a certain one simply by extending its boundaries into the neighbouring district so that it includes neighbourhoods more likely to vote for the governing party.  Also, neighbourhoods likely to vote for the opposition can be transferred out.  Thus, from one certain loss plus one uncertainty, the government has fashioned a fifty-fifty situation.  The peculiar shape on the map of the political districts thus created is the origin of the name “gerrymander” – a peculiarly political sort of mythical beast.

[2] This kind of reasoning, tossing data out because you don’t like the results they give, is the entire basis of a recent book denying climate change, as the following review makes clear.


[3] Morton would more likely have used the term Eskimo or one of its variants, but since that term is generally regarded today as derogatory, in the absence of a direct quote, I prefer the term the Inuit use themselves.  I use the term “barbarous tribes” because that is Morton’s own favoured designation.

Posted in: Featherfolk