Re: A Demand for the Kennewick Man's Remains
Sun, 24 Nov 1996 15:22:36 -0600

On 22 Nov 1996, Gerold Firl wrote:

> In article <>, writes:
> |> Which begs the question, "what is race"? It is a folk concept -- it
> |> certainly has no scientific currency, especially in anthropology.
> What do you make of the fact that the concept of race (in the sense of
> subspecies) is a vital concept in evolutionary biology, and yet it has
> no "scientific" currency in anthro? Maybe cultural anthropologists
> simply have more mana than biologists, so when they declare that the
> geographic variation of human physical types is not "racial" (i.e.,
> due to genetic variation caused by breeding isolation and consequent
> adaptation and drift) then any further discussion is taboo. Any other
> animal which displayed the geographic variation seen in humankind
> would be classified as having different races; the extreme
> ethnocentrism of anthropologists who declare a priviliged position for
> man is delightfully ironic.

I think, from the example you give below, that we are using the term
"race" in two different senses. I am using it to refer to the popular
conception of what is meant by "race" -- three major quasi-geographic
populations of humans, African (Negroid), Asian (Asiatic), and European
(Caucasoid). It seems from your example that you are referring to "race"
in a sense of "subspecies." This is a significant difference, because
the variation among any one of the quasi-geographic "races" is (I think)
as great as the variation between them -- so basically the "racial"
identifications are used as justifications of a classification by skin color.

"Race" is one of those dangerous terms with both technical and popular
meanings; people unfamiliar with the word's technical meaning see it used
in a technical sense but interpret it in line with their popular
understanding. You use it in a much more partitive sense. An analogy
might be languages -- we group languages into families, but the component
languages might be scattered across wide distances. Moreover, shared
language family does not necessarily mean shared identification with.

Similarly, the kinds of races of which you speak (rainforest pygmies and
nilo-sudanese) may be part of a larger "racial family" which is popularly
called "African" or "Negroid." That does not mean that the "family" is a
race; it is a means of classification that we infer to carry genetic
affiliation. It also does not mean that people descended from such
populations will identify more closely with "co-racial-family" people
than with anyone else.

> The
> |> indices by which the "race" of skeletons is determined are statistical
> |> indices. They do not give a probability that a skeleton is one "race" or
> |> another -- they give the frequency of occurrence of skeletal traits. It
> Well; the frequency of occurance of skeletal traits can be used to
> infer the probabilities concerning population origin for particular
> remains. If that's the only information you have, then that's the best
> you can do.

The key here in terms of NAGPRA is "the only information you have". We
have no evidence for populations of "caucasians" in North America circa
9,000 years ago. This is one individual. She or he might have come from
such a population. But by itself, it is not evidence for the presence of
such a population. Statistics don't mean anything without an explanatory
framework. Moreover, as probabilities, they allow for other
possibilities. Just because a skeleton manifests certain statistical
traits doesn't mean we automatically go with what the statistics suggest.

Besides, you are still assuming that the statistical indices we use are
valid for 9,000 years ago. Just because it's "the best you can do"
doesn't mean it's very good.

> |> Assigning "race" is therefore
> |> an educated guess, as much art as science.
> |> But you never really assign "race". Race presumes a particular
> |> historical circumstance (genetic relationship to a particular
> |> population). The statistical indices don't address questions of history,
> |> they address questions of associations between traits.
> Not necessarily. Suppose you had skeletal remains from the rwandan
> highlands, ca. 500 ad. This area has been an overlapping range for
> both rainforest pygmies and nilo-sudanese invaders since the
> introduction of iron working and domesticated cattle; a very
> unsophisticated forensic analysis will be sufficient to determine
> racial affiliation, unless the individual was a hybrid. You can argue
> that the racial distinction between a pygmy and a watutsi is trivially
> insignificant, of absolutely no import except as a means for imposing
> socio-economic/caste distinctions, but such a claim ignores the very
> real, long-term ecological adaptations of these two populations.

Again, you're using "race" in a very different sense from the one in
which I meant and from the popular conception. Moreover, you're not
justifying the use of statistical indices as a primary means of racial
identification very well. You are referring to two small (in comparison
to "African/Negroid") populations, which may not have nearly as much
internal variation as the larger ("African/Negroid") population.
Moreover, you acknowledge the importance of cultural factors between the
populations. The determination of racial affiliation of which you speak
is predicated upon a knowledge of the cultural environment.

But even so, if all you have is a skeleton, you are still only making an
educated guess. Statistical probabilities can never be more than
probabilities. There is always that small probability that it isn't what
you would expect. In the case of the Washington skeleton, the
pro-caucasian people are privileging the statistical indices over
archaeological evidence. A sample of one does not suffice to rewrite

In that sense, I would like to reword something I said above: we do
assign "race". Science is a human product; scientific categories are
therefore human ones. What we call "natural" is still really human,
because we define the groups into which we put things. Our categories
may closely parallel the really natural ones, but they can rarely (if
ever) exactly match them. The discussion of statistics and probablity is
a good illustration of why: our categories may be defined statistically,
but the fact that error is unavoidable (by the very nature of error, an
decrease in Type A error brings an increase in Type B, and vice versa)
means that our categories are unlikely to perfectly match the natural
ones. If we classify as Race A an individual who is a good statistical
match, but who is genetically Race B and is just exhibiting extreme
variation, then our category does not match the natural one. We assume
that phenotype correlates with genotype -- but the correlation is not
perfect, and as long as we classify by effect rather than cause, we are
going to make mistakes. So inasmuch as we put people into one race or
another, we are assigning race.

> Recent history should be sufficient to show that while our world might
> be a better place if such distinctions really were insignificant,
> wishing does not make it so.

True -- but what has this to do with statistical indices and NAGPRA?

Rebecca Lynn Johnson
Ph.D. stud., Dept. of Anthropology, U Iowa