{Reprinted from Science, N. S., Vol. IV., No. 103, Pages 901-908, December 18, 1896.

~\

SCIENCE
Editorial Committee S, Newcomb, Mathematics; R. S. Woodward, Mechanics; E. C. Pickering,
:

Astronomy; T. C. Mendenhall, Physics; R. H. Thurston, Engineering; Ira Remsen, Chemistry;

J. Le Conte, Geology; W. M. Davis, Physiography; O. C. Marsh, Paleontology; W. K.
Brooks, C. Hart Merriam, Zoology; S. H. Scudder, Entomology; N. L. Britton,
Botany; Henry F. Osborn, General Biology; H. P. Bowditch, Physiology;
J. S. Billings, Hygiene ; J. McKeen Cattell, Psychology;

Daniel G. Brinton, J. W. Powell, Anthropology.

THE LIMITATIONS OF THE COMPARATIVE their attention upon this purely historical
METHOD OF ANTHROPOLOGY *

problem, the tide has now completely

Modern Anthropology has discovered turned, so that there are even anthropolo-
the fact that human society has grown and gists who declare that such investigations
developed everywhere in such a manner belong to the historian, and that anthropo-
that its forms, its opinions and its actions logical studies must be confined to re-
have many fundamental traits in common. searches on the IaAVS that govern the growth
This momentous discovery implies that laws of society.
exist which govern the development of A radical change of method has accom-
society, that they are applicable to our soci- panied this change of views. While for-
ety as well as to those of past times and of merly identities or similarities of culture
distant lands that their knowledge will be a were considered incontrovertible proof of
;

means of understanding the causes further- historical connection, or even of common

ing and retarding civilization and that, origin, the new school declines to consider
;

guided by this knowledge, we may hope to them as such, but interprets them as results
govern our actions so that the greatest bene- of the uniform working of the human mind.
fit to mankind will accrue from them. The most pronounced adherent of this view
Since this discovery has been clearly formu- in our country is Dr. D. G. Brinton, in Ger-
lated, anthropology has begun to receive many the majority of the followers of Bas-
that liberal share of public interest which tian, who in this respect go much farther
was withheld from it as long as it was be- than Bastian himself. Others, while not
lieved that it could do no more than record denying the occurrence of historical con-
the curious customs and beliefs of strange nections, regard them as insignificant in re-
peoples; or, at best, trace their relationships, sults and in theoretical importance as com-
and thus elucidate the early migrations of pared to the working of the uniform laws
the races of man and the affinities of peoples. governing the human mind. This is the
While early investigators concentrated view of by far the greater number of living
*
Paper read at the meeting of the A. A. A. S. at anthropologists.
Buffalo. This modern view is founded on the ob-
2 SCIENCE.

servation that the same ethnical phe- ternal, that is founded in environment —

nomena occur among the most diverse taking the term environment in its widest
peoples, or, as Bastian says, on the appall- sense—or internal, that is founded on
ing monotony of the fundamental ideas of psychological conditions. The influence of
mankind all over the globe. The meta- external and internal factors upon elemen-
physical notions of man may be reduced to tary ideas embodies one group of laws
a few types which are of universal distribu- governing the growth of culture. There-
tion ; the same is the case in regard to the fore, our endeavors must be directed to
forms of society, laws and inventions. showing how such factors modify elemen-
Furthermore, the most intricate and ap- tary ideas.
parently illogical ideas and the most curious The first method that suggests itself and
and complex customs appear among a few which has been generally adopted by mod-
tribes here and there in such a manner that ern anthropologists is to isolate and classify
the assumption of a common historical ori- causes by grouping the variants of certain
gin is excluded. When studying the cul- ethnological phenomena according to exter-
ture of any one tribe, more or less close nal conditions under which the people live,
analoga of single traits of such a culture among whom they are found, or to internal
may be found among a great diversity of causes wdiich influence their minds or in-
;

peoples. Instances of such analoga have versely, by grouping these variants accord-
been collected to a vast extent by Tylor, ing to their similarities. Then the corre-
Spencer, Bastian, Andree, Post and many lated conditions of life may be found.
others, so that it is not necessary to give By this method we begin to recognize
here any detailed proof of this fact. The even now with imperfect knowledge of the
idea of a future life, the one underlying- facts what causes may have been at work
shamanism ; inventions such as fire and the in shaping the culture of mankind. Fried-
bow; certain elementary features of gram- rich Batzel and W J McGee have investi-
matical structure—these will suggest the gated the influence of geographical environ-
classes of phenomena to which I refer. It ment on a broader basis of facts than Ritter
follows from these observations that when and Guyot were able to do at their time.
we find an analogon of single traits of cul- Sociologists have made important studies
ture among distant peoples, the presump- on the effects of the density of population
tion is not that there has been a common and of other simple social causes. Thus
historical source, but that they have arisen the influence of external factors upon the
independently. growth of society is becoming clearer.
But the discovery of these universal The effects of psychical factors are also
ideas is only the beginning of the work of being studied in the same manner. Stoll
the anthropologist. Scientific inquiry must has tried to isolate the phenomena of sug-
answer two questions in regard to them : gestion and of hypnotism and to study the
First, what is their origin ? and second, effects of their presence in the cultures of
how do they assert themselves in various various peoples. Inquiries into the mutual
cultures ? relations of tribes and peoples begin to show
The second question is the easier one to that certain cultural elements are easily
answer. The ideas do not exist everywhere assimilated while others are rejected, and
in identical form, but they vary. Sufficient the time-worn phrases of the imposition of
material has been accumulated to show that culture by a more highly civilized people
the causes of these variations are either ex- upon one of lower culture that has been
 SCIENCE.

conquered are giving way to more thorough should be a distinction between the 1 we ’

views on the subject of exchange of cultural expressing the self and the person addressed
achievements. In all these investigations and the ‘we expressing the self and the
’

we are using sound, inductive methods in person spoken of, which distinction is found
order to isolate the causes of observed phe- in comparatively few languages only. The
nomena. lesser liability to misunderstandings in the
The other question in regard to the uni- plural explains this phenomenon partly but
versal ideas, namely that of their origin, is hardly adequately. Still more obscure is
much more difficult to treat. Many attempts the psychological basis in other cases, for
have beenmade to discover the causes which instance, in that of widely spread marriage
have led to the formation of ideas 1 that de- customs. Proof of the difficulty of this
velop with iron necessity wherever man problem is the multitude of hypotheses that
lives.’ This is the most difficult problem have been invented to explain it in all its
of anthropology and we may expect that it varied phases.
will baffle our attempts for a long time to In treating this, the most difficult problem
come. Bastian denies that it is possible to of anthropology, the point of view is taken
discover the ultimate sources of inventions, that if an ethnological phenomenon has
ideas, customs and beliefs which are of uni- developed independently in a number of
versal occurrence. They may be indigenous, places its development has been the same
they may be imported, they may have arisen everywhere; or, expressed in a different
from a variety of sources, but they are there. form, that the same ethnological phenomena
The human mind is so formed that it invents are always due to the same causes. This
them spontaneously or accepts them when- leads to thestill wider generalization thatthe
ever they are offered to it. This is the much sameness of ethnological phenomena found
misunderstood elementary idea of Bastian. in diverse regions is proof that the human
To a certain extent the clear enunciation mind obeys the same laws everywhere. It
of the elementary idea gives us the psycho- is obvious that if different historical devel-
logical reason for its existence. To exem- opments could lead to the same results, that
plify: the fact that the land of the shadows then this generalization would not be ten-
is so often placed in the west suggests the able. Their existence would present to us
endeavor to localize it at the place where an entirely different problem, namely, how
the sun and the stars vanish. The mere it is that the developments of culture so
statement that primitive man considers often lead to the same results. It must,
animals as gifted with all the qualities of therefore, be clearly understood that an-
man shows that the analogy between many thropological research which compares sim-
of the qualities of animals and human quali- ilar cultural phenomena from various parts
ties has led to the generalization that all the of the world, in order to discover the uni-
qualities of animals are human. In other form history of their development, makes
cases the causes are not so self-evident. Thus the assumption that the same ethnological
the question why all languages distinguish phenomenon has everywhere developed in
between the self, the person addressed and the same manner. Here lies the flaw in
the person spoken of, and why most lan- the argument of the new method, for no
guages do not carry out this sharp, logical such proof can be given. Even the most
distinction in the plural is difficult to an- cursory review shows that the same phe-
swer. The principle when carried out con- nomena may develop in a multitude of
sistently requires that in the plural there ways.
4 SCIENCE.

I will give a few examples Primitive use of masks is found among a great num-
:

tribes are almost universally divided into ber of peoples. The origin of the custom
clans which have totems. There can be no of wearing masks is by no means clear in
doubt that this form ofsocial organization has all cases, but a few typical forms of their
arisen independently over and over again. use may easily be distinguished. They are
The conclusion is certainly justified that used for deceiving spirits as to the identity
the psychical conditions of man favor the of the wearer. The spirit of a disease who
existence of a totemistic organization of so- intends to attack the person does not recog-
ciety, but it does not follow that totemistic nize him when he wears a mask, and the
society has developed everywhere in the mask serves in this manner as a protection.
same manner. Dr. Washington Matthews In other cases the mask represents a spirit
has shown that the totems of the Navajo which is personified by the wearer, who in
have arisen by association of independent this shape frightens away other hostile
clans. Capt. Bourke has pointed out that spirits. Still other masks are commemora-
similar occurrences gave origin to the tive. The wearer personifies a deceased
Apache clans, and Dr. Fewkes has reached person whose memory is to be recalled.
the same conclusion in regard to some of Masks are also used in theatrical perform-
the Pueblo tribes. On the other hand, we ances illustrating mythological incidents.*
have proof that clans may originate by These few data suffice to show that the
division. I have shown that such events same ethnical phenomenon may develop
took place among the Indians of the North from different sources. The simpler the
Pacific coast. Association of small tribes, observed fact, the more likely it is that it
on the one hand, and disintegration of in- may have developed from one source here,
creasing tribes, on the other, has led to re- from another there.
sults which appear identical to all intents Thus we recognize that the fundamental
and purposes. assumption which is so often made by mod-
Here is another example. Recent inves- ern anthropologists cannot be accepted as
tigations have shown that geometrical de- true in all cases. We cannot say that the
signs in primitive art have originated either occurrence of the same phenomenon is
from naturalistic forms which were gradu- always due to the same causes, and that
ally conventionalized or from technical thus it is proved that the human mind
motives, or that they were primarily geo- obeys the same laws everywhere. We must
metrical or that they were derived from demand that the causes from which it de-
symbols. From all these sources the same veloped be investigated and that compari-
forms have developed. Out of designs sons be restricted to those phenomena which
representing diverse objects grew in course have been proved to be effects of the same
of time frets, meanders, crosses and the causes. We must insist that this investi-
like. Therefore the frequent occurrence of gation be made a preliminary to all ex-
these forms proves neither common origin tended comparative studies. In researches
nor that they have always developed accord- on tribal societies those which have devel-
ing to the same psychical laws. On the oped through association must be treated
contrary, the identical result may have separately from those that have developed
been reached on four different lines of de- through disintegration. Geometrical de-
velopment and from an infinite number of signs which have arisen from convention-
starting points. See Eichard Andree. EthnographisclieParallelen
*

Another example may not be amiss: The und Yergleiche. Nene Folge, 1889, pp. 107 ff.
 SCIENCE. 5
alizedrepresentations ofnatural objects must at all; that they much rather point in the
be treated separately from those that have opposite direction. Therefore we must also
arisen from technical motives. In short, consider all the ingenious attempts at con-
before extended comparisons are made, the structions of a grand system of the evolu-
comparability of the material must be tion of society as of very doubtful value,
proved. unless at the same time proof is given that
The comparative studies of which I am the same phenomena could not develop by
speaking here attempt to explain customs any other method. Until this is done, the
and ideas of remarkable similarity which presumption is always in favor of a variety
are found here and there. But they pur- of courses which historical growth may have
sue also the more ambitious scheme of dis- taken.
covering the laws and the history of the It will be well to restate at this place one
evolution of human society. The fact that of the principal aims of anthropological re-
many fundamental features of culture are search, We agreed that certain laws exist
universal, or at least occur in many iso- which govern the growth of human culture,
lated places, interpreted by the assumption and it is our endeavor to discover these
that the same features must always have laws. The object of our investigation is to
developed from the same causes, leads to find the processes by which certain stages of
the conclusion that there is one grand sys- culture have developed. The customs and
tem according to which mankind has de- beliefs themselves are not the ultimate ob-
veloped everywhere ; that all the occurring jects of research. We desire to learn the
variations are no more than minor details reasons why such customs and beliefs ex-
in this grand uniform evolution. It is clear ist —in other words, we wish to discover
that this theory has for its logical basis the the history of their development. The
assumption that the same phenomena are method which is at present most frequently
always due to the same causes. To give an applied in investigations of this character
instance We find many types of structure
; compares the variations under which the
of family. It can be proved that paternal customs or beliefs occur and endeavors to
families have often developed from mater- find the common psychological cause that
nal ones. Therefore, it is said, all paternal underlies all of them. I have stated that
families have developed from maternal ones. this method is open to a very fundamental
If we do not make the assumption that the objection.
same phenomena have everywhere devel- AVe have another method, which in many
oped from the same causes, then we may respects is much safer. A detailed study of
just as well conclude that paternal families customs in their bearings to the total cul-
have in some cases arisen from maternal ture of the tribe practicing them, and in
institutions, in other cases in other ways. connection with an investigation of their
To give another example: Many concep- geographical distribution among neighbor-
tions of the future life have evidently de- ing tribes, afford us almost always a means
veloped from dreams and hallucinations. of determining with considerable accuracy
Consequently, it is said, all notions of this the historical causes that led to the forma-
character have had the same origin. This tion of the customs in question and to the
is also true only if no other causes could psychological processes that were at work
possibly lead to the same ideas. in their development. The results of in-
We have seen that the facts do not favor quiries conducted by this method may be
the assumption of which we are speaking three-fold. They may reveal the environ-
6 SCIENCE.

mental conditions which have created or independently. Therefore the investigation

modified cultural elements; they may clear must always demand continuity of distri-
up psychological factors which are at work bution as one of the essential conditions for
in shaping the culture ; or they may bring proving historical connection, and the as-
before our eyes the effects that historical sumption of lost connecting links must be
connections have had upon the growth of applied most sparingly. This clear distinc-
the culture. tion between the new and the old historical
We have in this method a means of re- methods is still often overlooked by the
constructing the history of the growth of passionate defenders of the comparative
ideas with much greater accuracy than the method. They do not appreciate the dif-
generalizations of the comparative method ference between the indiscriminate use of
will permit. The latter must pro- similarities of culture for proving historical
ceed from a hypothetical mode of develop- connection and the careful and slow de-
ment, the probability of which may be tailed study of local phenomena. We no
weighed more or less accurately by means longer believe that the slight similarities
of observed data. But so far I have not between the cultures of Central America
yet seen any extended attempt to prove the and of eastern Asia are sufficient and satis-
correctness of a theory by testing it at the factory proof of a historical connection. On
hand of developments with whose histories the contrary, analogy of other similarities
we are familiar. This method of starting make such a connection improbable. But,
with a hypothesis is infinitely inferior to on the other hand, no unbiased observer
the one in which by truly inductive pro- will deny that there are very strong reasons
cesses the actual history of definite phe- for believing that a limited number of cul-
nomena is derived. The latter is no other turalelements found in Alaska and in Siberia
than the much ridiculed historical method. have a common origin. The similarities of
Its way of proceeding is, of course, no inventions, customs and beliefs, together
longer that of former times when slight with the continuity of their distribution
similarities of culture were considered through a comparatively small area, are a
proofs of relationships, but it duly recog- satisfactory proof of this opinion. But it is
nizes the results obtained by comparative not possible to extend this area safely be-
studies. Its application is based, first of yond the limits of Columbia River in
all, on a well-defined, small geographical America and northern Japan in Asia. This
territory, and its comparisons are not ex- method of anthropological research is repre-
tended beyond the limits of the cultural sented in our country by Prof. F. W. Put-
area that forms the basis of the study. nam and Prof. Otis T. Mason in England
;

Only when definite results have been ob- by Dr. E. B. Tylor; in Germany by Fried-
tained in regard to this area is it permis- rich Ratzel and his followers.
sible to extend the horizon beyond its lim- It seems necessary to say a word here in
its, but the greatest care must be taken not regard to an objection to my arguments
to proceed too hastily in this, as else the fun- that will be raised by investigators who
damental proposition which I formulated be- claim that similarity of geographical en-
fore might be overlooked, viz; that when vironment is a sufficient cause for similarity
we find an analogy of single traits of cul- of culture, that is to say, that, for instance,
ture among distant peoples the presump- the geographical conditions of the plains of
tion is not that there has been a common the Mississippi basin necessitate the devel-
historical source, but that they have arisen opment of a certain culture. There are
 SCIENCE. 7

those who would even go so far as to be- tact with foreign tribes that they cannot
lieve that similarity of form of language be discovered without the most painstaking
may be due to environmental causes. En- isolation of foreign elements.
vironment has a certain limited effect upon The immediate results of the historical
the culture of man, but I do not see how method are, therefore, histories of'the cul-
the view that it is the primary moulder of tures of diverse tribes which have been the
culture can be supported by any facts. A subject of study. I fully agree with those
hasty review of the tribes and peoples of anthropologists who claim that this is not
our globe shows that people most diverse the ultimate aim of our science, because the
in culture and language live under the same general laws, although implied in such a
geographical conditions, as proof of which description, cannot be clearly formulated
may be mentioned the ethnography of East nor their relative value appreciated without
Africa or of New Guinea. In both these re- a thorough comparison of the manner in
gions we find a great diversity of customs which they assert themselves in different
in small areas. But much more important cultures. But I insist that the application
is this; Not one observed fact can be of this method is the indispensable condition
brought forward in support of this hypothe- of sound progress. The psychological prob-
sis which cannot be much better explained lem is contained in the results of the his-
by the well known facts of diffusion of cul- torical inquiry. When we have cleared up
ture; for archaeology as well as ethno- the history of a single culture and under-
graphy teach us that intercourse between stand the effects of environment and the
neighboring tribes has always existed and psychological conditions that are reflected
has extended over enormous areas. In the in it we have made a step forward, as we
Old World the products of the Baltic found can then investigate in how far the same
their way to the Mediterranean and the causes or other causes were at work in the
works of art of the eastern Mediterranean development of other cultures. Thus by
reached Sweden. In America the shells of comparing histories of growth general laws
the ocean found their way into the inner- may be found. This method is much safer
most parts of the continent and the obsidians than the comparative method, as it is
of the West were carried to Ohio. Inter- usually practiced, because instead of a
marriages, war, slavery, trade, have been hypothesis on the mode of development
so many sources of constant introduction actual history forms the basis of our deduc-
of foreign cultural elements, so that an tions.
assimilation of culture must have taken The historical inquiry must be consid-
place over continuous areas. Therefore, it ered the critical test that science must
seems to my mind that where among neigh- require before admitting facts as evidence.
boring tribes an immediate influence of en- By its means the comparability of the col-
vironment cannot be shown to exist, the lected material must be tested, and uni-
presumption must always be in favor of formity of processes must be demanded as
historical connection. There has been a proof of comparability. It may also be
time of isolation during which the principal mentioned that when historical connection
traits of diverse cultures developed accord- between two phenomena can be proved,
ing to the character and environment of the they must not be admitted as independent
tribes. But the stages of culturerepresent- evidence.
ing this period have been covered with so In a few cases the immediate results of
much that is new and that is due to con- this method are of so wide a scope that they
8 SCIENCE.

rank with the best results that can be at- growth alone, but whenever such is feasible
tained by comparative studies. Some phe- it must compare the processes of growth,
nomena have so immense a distribution and these can be discovered by means of
that the discovery of their occurrence over studies of the cultures of small geographi-
very large continuous areas proves at once cal areas.
that certain phases of the culture in these Thus we have seen that the comparative
areas have sprung from one source. Thus method can hope to reach the grand results
are illuminated vast portions of the early for which it is striving only when it bases
history of mankind. When Prof. Morse its investigations on the historical results
showed that certain methods of arrow of researches which are devoted to laying
release are peculiar to whole continents it clear the complex relations of each indi-
became clear at once that the common vidual culture. The comparative method
practice that is found over a vast area must and the historical method, if I may use
have had a common origin. When the these terms, have been struggling for
Polynesians employ a method of fire making supremacy for a long time, but we may
consisting in rubbing a stick along a groove, hope that each will soon find its appropriate
while almost all other peoples use the fire place and function. The historical method
drill, it shows their art of fire making has a has reached a sounder basis by abandoning
single origin. When we notice that the the misleading principle of assuming con-
ordeal is found all over Africa in certain nections wherever similarities of culture
peculiar forms, while in those parts of the were found. The comparative method,
inhabited world that are remote from notwithstanding all that has been said and
Africa it is found not at all or in rudi- written in its praise, has been remarkably
mentary forms only, it shows that the idea barren of definite results, and I believe it
as practiced in Africa had one single origin. will not become fruitful until we renounce
The great and important function of the the vain endeavor toconstruct a uniform sys-
historical method of anthropology is thus tematic history of the evolution of culture,
seen to lie in its ability to discover the pro- and until we begin to make our compari-
cesses which in definite cases led to the de- sons on the broader and sounder basis which
velopment of certain customs. If anthro- I ventured to outline. Up to this time we
pology desires to establish the laws govern- have too much reveled in more or less in-
ing the growth of culture it must not con- genious vagaries. The solid work is still
fine itself to comparing the results of the all before us. Franz Boas.