CTLF Corpus de textes linguistiques fondamentaux • IMPRIMER • RETOUR ÉCRAN
CTLF - Menu général - Textes

Boas, Franz. Race, Language and Culture – T20


Race and Character 1

At the present time so much is being written on the relations between
race and character that it is worth while to examine with
some care the line of thought that leads investigators to the conclusion
that racial descent is the determining cause for the character of a

It is a matter of observation that peoples located in different areas
are different both in bodily form and mental traits; and also that different
social strata differ in bodily build and in behavior. Using the
favorite terminology of modern literature we may also say that there
is a correlation between the bodily build and mental characteristics of
geographically or socially arranged groups of people. However, not
every correlation signifies a causal relation.

I may be allowed to illustrate this by means of a few examples. We
know that in a homogeneous population all anteroposterior measures
are more closely related among themselves than to transverse ones. For
this reason with increasing stature of adults the length of head increases
more rapidly than the width of head and in consequence the cephalic
index decreases with increasing stature. Every homogeneous population
shows a negative correlation between stature and head index which is
causally explained by the intimate interrelation between longitudinal
measures as over their loose relation to transverse measures. In this
sense the correlation expresses a causal relation.

If we examine the population of Italy from the same point of view
it is found to be locally strongly differentiated. There is a geographical
arrangement of fairly homogeneous groups of various types. In northern
Italy we find tall, round-headed types, in southern Italy short, longheaded
ones. If we compare the groups as such we find that the taller
the average stature of the group, the larger is its cephalic index; but we
may not conclude that this relation is determined organically. It is due
to the heterogeneous character of the material and the distribution of
various types. If the whole Italian population were investigated without
191reference to their location we should probably find a very weak positive
correlation, or perhaps no correlation whatever between these two measures.

This observation may perhaps be made still clearer by an artificial
example. I imagine a series of sticks of equal length, placed parallel,
side by side, so that their ends from left to right form a straight line at
right angles to the length of the sticks. Then I cut the other ends off
obliquely so that the length of the sticks decreases from left to right.
Next another person paints these sticks so that the larger ones, to the
left, are darkest and the intensity of color decreases towards the right.
Now there is an intimate relation between length and intensity of color,
but length and intensity are not causally connected. The correlation is
a result of the position of the sticks and of two unrelated actions. I may
not say: the intensity of color is determined by the length of the sticks,
but it is due to the fact that the sticks were in a certain order when they
were painted. If I had changed the position of the sticks before painting
them in the same way a different kind of correlation would have resulted;
if I should have shaken them, so that they lay in chance order
there would have been no correlation between length and intensity of

Matters would have been different, if knife and brush had been firmly
tied together. The position of the sticks would still have been decisive,
but every stick would have had a color and length which belonged together
even without consideration of position. A study of the length and
color would not clear up this point. It would require an examination of
knife and brush.

Now let us consider instead of the sticks lying in order, a number of
populations according to their geographical position; instead of length
of stick, bodily form; instead of color, mental character. We will also
imagine a continuous change in regard to both in a straight line. Then a
correlation will become apparent. Every people in a certain geographical
position has a characteristic bodily form combined with a characteristic
mental behavior. This, however, does not prove that both are
causally related, unless it can be proved by biological and psychological
methods that bodily form determines mental character.

The same consideration is valid when the distribution of populations
is discontinuous and irregular. In this case the relation cannot be expressed
numerically, but the phenomenon remains the same. The population
of each locality or every social group has certain traits of bodily
192form and mental behavior peculiar to itself, but this does not prove that
the two are causally related.

Bodily form and mental characteristics change each according to its
own laws and each in its own tempo, so that it is justifiable to ask
whether the population placed in another geographical location may
not retain its bodily form and change its mental character, analogous
to the change of the location of sticks of decreasing length before the
paint had been applied.

I repeat, the essential question, whether bodily form and mental character
are causally connected cannot be answered by means of the observation
that populations in different geographical location or in differing
social strata are different in both respects. The proof has to be
given by biological and psychological methods. We have here one of
the numerous cases in which the uncritical use of the concept of correlation
leads to unjustifiable conclusions.

It might be objected that the study of heredity and constitution has
proved the existence of partial, biologically determined relation between
bodily form and mental character which is not due to location
in a given order. This may be admitted. If behaviorists deny such
relations in the individual their claim is contradicted by the most elementary
facts of pathology. In how far there may be, nevertheless,
room for individual differences in mental character among individuals
of the same bodily form does not need to be discussed here.

On the other hand it is essential for our problem to differentiate between
individual character and the character of a population. I may
illustrate this problem also by the example of our sticks. We assume
that a large number of series, let us say. each of one hundred sticks, are
cut obliquely, as indicated before, but in such a manner that there are
a few only of the longest and shortest ones, while in the center of
the series the length changes slowly so that in this region there are
many of almost equal length. The absolute lengths of the sticks of the
first group extend from 1 to 80, that of the second from 2 to 81, and so
on; those of the last from 21 to 100. Next each series is painted separately,
as before. Knife and brush are supposed to be firmly connected
so that there is a causal relation between color and length. Now the
dark colors appear solely in the first few series, the light ones solely in
the last series, but all of them contain numerous sticks of middle length
and color. As groups the series will differ only slightly, although we
have assumed a causal relation between length and color. The degree
193of differences in color will depend upon the number of occurrence of
sticks of the same length in each series.

Let us transfer this to the question of relation between form of body
and mental traits. Length corresponds to bodily form, color to mental
traits, each series to a population. Now the populations differ slightly,
in regard to form of body and mental characteristics, although we assume
that individually mental characteristics are conditioned by form
of body. Since the relation between form of body and mental characteristics
is not absolute, their relation is still further weakened, even if
the overlapping of the series in regard to bodily form were less. The
questions to be answered are the following: How strong is the correlation
between bodily form and mental characteristics? And secondly, how
are the bodily characteristics which are important for the determination
of mental characteristics distributed between various groups of people,
and to what extent are the same bodily characteristics found among
various people? All these questions have to be treated without reference
to geographical or social position.

Let us return once more to the series of sticks which I have used as
an example. After they have been cut and painted we arrange bundles,
the first one is to contain sticks of the length 1 to 80, the second, those
of the length 2 to 81 and so on, the last one to contain sticks of the length
21 to 100. Now we place the bundles in a series and paint as before the
whole bundles from left to right, let us say, in twenty degrees of intensity,
dark to the left, light to the right. Then every bundle will have a
different color. After this the sticks of equal length taken from all the
bundles are placed together and it will be found that the distribution
of colors for each length shows very slight differences. Only the first
length has the color 1, only the two first ones contain the colors 1 and 2.
For the length 20 to 80 distribution of the colors is alike. In the lengths
from 81 to 100 the dark colors disappear gradually until finally the
last one remains with the lightest color. Since the very short and very
long sticks are few in number they do not influence the general picture
very much and the result is that the relation between length and color
is very weak.

Transferring this as before to form of body and mental characteristics
we find that populations arranged geographically or socially are different
like the bundles. It is impossible to determine from observation of
the distribution whether the differences are due to causal relations or to
the arrangement of the bundles. Individuals are so distributed that the
194relation between form of body and mental characteristics in the whole
mass is very slight. In the most numerous groups, all kinds of bodily
form and mental characteristics occur, providing the types are as similar
as those of Europe. This does not preclude the possibility of the hereditary
determination of the relation between bodily build and mental
characteristics in family lines, since the whole population consists of
numerous different lines.

In many cases in populations of similar bodily build and also among
different generations of the same people, mental characteristics of considerable
difference occur. For this reason it seems more likely that
differences between populations are rather due to position than to immediate
causal relations.

This may also be expressed in a simpler way. Assuming for the sake
of simplicity that position, bodily form and mental characteristics each
are distributed according to chance so that the ordinary method of
determining correlations can be used, then observations will show high
correlations between position and bodily build and between position
and mental characteristics. From this fact we may not infer how high
may be the correlation between bodily build and mental characteristics
unless this is determined by an investigation which does not take into
consideration position. According to the simple factual observation
they may be non-existent or may be very high. Or, stated in a still
more general form, a series of phenomena may be placed in a definite
order, then exposed to two causes, each of which has a certain influence
upon it. According to their character these two causes are entirely independent
of the principle of arrangement. Then every member of the
series will have two definite characteristics. Whether they are related
or not can only be determined by an investigation of the relation between
the two causes without any regard to the arrangement.195

1 Anthropologischer Anzeiger, vol. 8 (1932), pp. 280-284.