WI LEY

A M L

I C A N

ANTIIROPOIOGICAL
AS S

O C I

A T

I O N

"Eskimo
Words for
Snow": A
A Case
Case Study
Genesis and
and Decay
Decay of
"Eskimo Words
f o r Snow":
Study in
i n the
t h e Genesis
o f an
an Anthropological
Anthropological
Example
Example
Author(s): Laura
artin
Author(s):
Laura M
Martin
Source:
Anthropologist,
New
Series, Vo
Vol.
88, No.
No. 22 (Jun.,
(Jun., 1986),
1986), pp.
pp. 418-423
418-423
Source: AAmerican
merican A
nthropologist, N
e w Series,
l . 88,
Published by:
i l e y on
o f tthe
he A
m e r i c a n Anthropological
Published
by: W
Wiley
on behalf
behalf of
American
Anthropological Association
Association
Stable URL:
t t p : / / w w w. , i s t o r. o r g / s t a b l e / 6 7 7 5 7 0
Stable
URL: hhttp://www.jstor.org/stable/677570
Accessed: 15-03-2016
Accessed:
15-03-2016 21:08
21:08 UTC
UTC

Your use of the JSTOR archive indicates your acceptance of

of the
the Terms &
& Conditions
Conditions of
of Use,
Use, available
available at
at http://www.jstor.org/page/
http://wwwjstor.org/page/
info/about/policies/terms jsp
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars,
scholars, researchers,
researchers, and
and students
students discover,
discover, use,
use, and
and build
build upon
upon aa wide
wide range
range of
of content
content
in a trusted digital archive. We use information technology and
and tools to increase
increase productivity and
and facilitate
facilitate new
new forms
forms of
of scholarship.
scholarship.
For more information about JSTOR, please contact
contact support@jstor.org.

American
Anthropological
Association and
and Wiley
are collaborating
American A
n t h r o p o l o g i c a l Association
W i l e y are
collaborating with
with JSTOR
JSTOR to
to digitize,
digitize, preserve
preserve and
and extend
extend access
access to
to
American
Anthropologist.
American A
nthropologist.

STOR
http://www.jstor.org
http://wwwjstor.org

This
This content
content downloaded
downloaded from
from 170.140.26.180
170.14026.180 on
on Tue,
Tue, 15
15 Mar
Mar 2016
2016 21:08:12
21:08:12 UTC
UTC
All
All use
use subject
subject to
to JSTOR
JSTOR Terms
Terms and
and Conditions
Conditions

Research Reports

Research Reports

"Eskimo Words for Snow": A

Case Study in the Genesis and
Decay of an Anthropological
Example

the noncomparability of language structures, not to examine their cultural or

cognitive implications.
The example became i n e x t r i c a b l y
identified with Benjamin Whorf through
the popularity o f "Science and Linguist i c s , " h i s 1940 a r t i c l e (see C a r r o l l
1956:207-219) exploring the same ideas
that interested Boas, lexical elaboration
not chief among them. Although for Boas
the example illustrated a similarity between English and "Eskimo," Whorf reorients it to contrast them (1956:216). It is
a minor diversion in a discussion of pervasive semantic categories such as time
and space, and he develops it no further,
here or elsewhere in his writings.
O f particular significance is Whorl's
failure to cite specific data, numbers, or
sources. H i s English glosses suggest as
many as five words, but not the same set
given by Boas. Although Whorls source
is uncertain, if he did rely on Boas, his apparently casual revisions of numbers and
glosses are but the first mistreatments to
which the original data have been subjected.
Anthropological fascination with the
example is traceable to two influential
textbooks, written i n the late 1950s b y
members of the large group of language
scientists familiar with "Science and Linguistics," and adopted in a variety of disciplines well into the 1970s. One or both
of these were probably read by most anthropologists trained between 1960 and
1970, and by countless other students as
well during that heyday of anthropology's
popularity.
In the first, The Silent Language, Edward
Hall mentions the example only three
times (1959:107-108,110), but his treatment of it suggests that he considered it
already familiar to many potential readers. Hall credits Boas, but misrepresents
both the intent and extent of the original
citation. Even the data are misplaced.
Hall inexplicably describes the Eskimo
the noncomparability of language struc-

"Eskimo Words for Snow": A

tures, not to examine their cultural or

Case Study in the Genesis and

cognitive implications.

Decay of an Anthropological

The example became inextricably

Example

identified with Benjamin Whorf through

the popularity of "Science and Linguis-

LAURA MARTIN
Cleveland State University

LAURA MARTIN

tics," his 1940 article (see Carroll

Cleveland State University

1956:207-219) exploring the same ideas

that interested Boas, lexical elaboration

A common example purportedly documenting the inextricable linkage of language, culture, and thought refers to "Eskimo words for snow." According to this
example, undifferentiated " E s k i m o "
languages are credited with some variable number of unique words for snow and
are compared to English, which has but
one. As most commonly expressed, the
example refers to the power that cultural
interests or setting have on the structure
of language (e.g., Pyles 1964:16). A somewhat more sophisticated version applies
the putative Eskimo categorization o f
snow to theories of grammatical influence
on perception (e.g., Smith and Williams
1977:143). Other examples of vocabulary
elaboration are sometimes used for similar explanatory purposes, but none is as
widely cited as this one. Such popularity
is at once ironic and unfortunate because
the evolution of the example, a curious sequence of distortions and inaccuracies,
offers both a case study in the creation of
an oral tradition and an object lesson on
the hazards of superficial scholarship.
The earliest reference to Eskimos and
snow was apparently made b y Franz
Boas (1911:25-26). Among many examples of cross-linguistic variation in the
patterns o f form/meaning association,
Boas presents a brief citation of four lexically unrelated words for snow in Eskimo:
aput 'snow on the ground', gana 'falling
snow', pivirpoq ' d r i f t i n g snow', a n d
gimugsug 'a snow drift'. I n this casual example, Boas makes l i t t l e distinction
among " r o o t s , " " w o r d s , " and " i n d e pendent terms." He intends to illustrate

not chief among them. Although for Boas

A common example purportedly doc-

the example illustrated a similarity be-

umenting the inextricable linkage of lan-

tween English and "Eskimo," Whorfreo-

guage, culture, and thought refers to "Es-

rients it to contrast them (1956:216). It is

kimo words for snow." According to this

a minor diversion in a discussion of per-

example, undifferentiated "Eskimo"'

vasive semantic categories such as time

languages are credited with some varia-

and space, and he develops it no further,

ble number of unique words for snow and

here or elsewhere in his writings.

are compared to English, which has but

Of particular significance is Whorf's

one. As most commonly expressed, the

failure to cite specific data, numbers, or

example refers to the power that cultural

sources. His English glosses suggest as

interests or setting have on the structure

many as five words, but not the same set

of language (e.g., Pyles 1964:16). A some-

given by Boas. Although Whorf's source

what more sophisticated version applies

is uncertain, if he did rely on Boas, his ap-

the putative Eskimo categorization of

parently casual revisions of numbers and

snow to theories of grammatical influence

glosses are but the first mistreatments to

on perception (e.g., Smith and Williams

which the original data have been sub-

1977:143). Other examples of vocabulary

jected.

elaboration are sometimes used for simi-

Anthropological fascination with the

lar explanatory purposes, but none is as

example is traceable to two influential

widely cited as this one. Such popularity

textbooks, written in the late 1950s by

is at once ironic and unfortunate because

members of the large group of language

the evolution of the example, a curious se-

scientists familiar with "Science and Lin-

quence of distortions and inaccuracies,

guistics," and adopted in a variety of dis-

offers both a case study in the creation of

ciplines well into the 1970s. One or both

an oral tradition and an object lesson on

of these were probably read by most an-

the hazards of superficial scholarship.

thropologists trained between 1960 and

The earliest reference to Eskimos and

1970, and by countless other students as

snow was apparently made by Franz

well during that heyday of anthropology's

Boas (1911:25-26). Among many exam-

popularity.

ples of cross-linguistic variation in the

In the first, The Silent Language, Edward

patterns of form/meaning association,

Hall mentions the example only three

Boas presents a brief citation of four lexi-

times (1959:107-108, 110), but his treat-

cally unrelated words for snow in Eskimo:

ment of it suggests that he considered it

aput 'snow on the ground', qana 'falling

snow', piqsirpoq 'drifting snow', and

already familiar to many potential read-

ers. Hall credits Boas, but misrepresents

qimuqsuq 'a snow drift'. In this casual ex-

both the intent and extent of the original

ample, Boas makes little distinction

among "roots," "words," and "inde-

citation. Even the data are misplaced.

Hall inexplicably describes the Eskimo

pendent terms." He intends to illustrate

418

418

This content downloaded from 170.140.26.180

170.14026.180 on
on Tue,
Tue. 15
15 Mar
Mar 2016
2016 21:08:12
21:08:12 UTC
LUC
All
All use
use subject
subject to
to JSTOR
JSTOR Terms
Terms and
and Conditions
Conditions

RESEARCH REPORTS

RESEARCH REPORTS 419

data as "nouns" and, although his argument implies quite a large inventory, specific numbers are not provided. Hall introduces still another context for the example, using it in the analysis of cultural
categories.
At approximately the same time, Roger
Brown's Words and Things (1958) appeared, intended as a textbook in the
"psychology of language." Here the example is associated with Whorf and thoroughly recast. Brown claims precisely
"three Eskimo words for snow," an assertion apparently based solely on a drawing
in Whorl's paper. Psychological and cognitive issues provide still another context
in Brown's discussion o f a theory about
the effects of lexical categorization on perception (cf. Brown and Lenneberg 1954).
Brown's discussion illustrates a creeping carelessness about the actual linguistic facts of the example; this carelessness
is no less shocking because it has become
so commonplace. Consider Brown's application of Zipfs Law to buttress arguments about the relationship between lexicon and perception. Since Z i p f s Law
concerns word length, Brown's hypothesis must assume something about the
length of his "three" "Eskimo" "snow"
words; his argument stands or falls on the
assumption that they must be both short
and frequent. Eskimo words, however,
are the products of an extremely synthetic
morphology in which all word building is
accomplished by multiple suffixation.
Their length is well beyond the limits of
Z i f f calculations. Furthermore, precisely identical whole "words" are unlikely to recur because the particular
combination o f suffixes used w i t h a
"snow" root, o r any other, varies by
speaker and situation as well as by syntactic role (Sadock 1980).
A minimal knowledge of Eskimo grammar would have confirmed the relevance
of these facts to the central hypotheses,
and would, moreover, have established
the even more relevant fact that there is
nothing at all peculiar about the behavior
or distribution of "snow words" in these
languages. T h e structure o f Eskimo
grammar means t h a t the n u m b e r o f
"words" for snow is literally incalculable,

a conclusion that is inescapable for any

other root as well.
Any sensible case for perceptual variation based on lexical inventory should,
therefore, require reference to distinct
"roots" rather than to "words," but this
subtlety has escaped most authors.2
Brown, for example, repeatedly refers to
linguistic units such as "verbal expression," "phrase," and " w o r d " i n a way
that underscores the inadequacy o f his
understanding of Eskimo grammar. His
assumption that English and "Eskimo"
are directly comparable, together with his
acceptance of pseudo-facts about lexical
elaboration in an unfamiliar language,
cause him to construct a complex psychocultural argument based on cross-linguistic "evidence" related to the example
with not a single item of Eskimo data in
support (1958:255). This complete absence of data (and of accurate references)
sets a dangerous precedent because it not
only p r e v e n t s d i r e c t e v a l u a t i o n o f
Brown's claims but suggests that such
evaluation is unnecessary.
As scholarship in linguistic anthropology, this treatment is wholly inadequate.
It is particularly unfortunate, then, that
this particular treatment was perpetuated and disseminated to a new generation of students in Carol Eastman's 1975
survey of linguistic approaches in anthropology, Aspects of Language and Culture.
Eastman summarizes the Sapir-Whorf
hypothesis, which she calls "the worldview problem," entirely by reference to
the snow example, q u o t i n g Brown's
"modifications'' o f Whorf's ideas
(1975:76). Even more striking than the
distortion o f Whorfs writing and thinking, which is implicit in the association of
him with it, is the powerful influence the
snow example exerts even on an experienced linguistic anthropologist. W i t h
Brown's reference to "three words" only
six lines away, Eastman still asserts that
"Eskimo languages have many words for
snow."
Thus is the complexity of the interrelations of linguistic structure, cultural behavior, and human cognition reduced to
"Eskimo words for snow." These and
other textbooks have disseminated mis-

data as "nouns" and, although his argu-

a conclusion that is inescapable for any

ment implies quite a large inventory, spe-

other root as well.

cific numbers are not provided. Hall in-

Any sensible case for perceptual varia-

troduces still another context for the ex-

tion based on lexical inventory should,

ample, using it in the analysis of cultural

therefore, require reference to distinct

categories.

At approximately the same time, Roger

"roots" rather than to "words," but this

subtlety has escaped most authors.2

Brown's Words and Things (1958) ap-

Brown, for example, repeatedly refers to

peared, intended as a textbook in the

linguistic units such as "verbal expres-

"psychology of language." Here the ex-

sion," "phrase," and "word" in a way

ample is associated with Whorf and thor-

that underscores the inadequacy of his

oughly recast. Brown claims precisely

understanding of Eskimo grammar. His

"three Eskimo words for snow," an asser-

assumption that English and "Eskimo"

tion apparently based solely on a drawing

are directly comparable, together with his

in Whorf's paper. Psychological and cog-

acceptance of pseudo-facts about lexical

nitive issues provide still another context

elaboration in an unfamiliar language,

in Brown's discussion of a theory about

cause him to construct a complex psycho-

the effects of lexical categorization on per-

cultural argument based on cross-linguis-

ception (cf. Brown and Lenneberg 1954).

Brown's discussion illustrates a creep-

ing carelessness about the actual linguis-

tic "evidence" related to the example

with not a single item of Eskimo data in

support (1958:255). This complete ab-

tic facts of the example; this carelessness

sence of data (and of accurate references)

is no less shocking because it has become

sets a dangerous precedent because it not

so commonplace. Consider Brown's ap-

only prevents direct evaluation of

plication of Zipf's Law to buttress argu-

Brown's claims but suggests that such

ments about the relationship between lex-

evaluation is unnecessary.

icon and perception. Since Zipf's Law

concerns word length, Brown's hypothe-

sis must assume something about the

As scholarship in linguistic anthropol-

ogy, this treatment is wholly inadequate.

It is particularly unfortunate, then, that

length of his "three" "Eskimo" "snow"

this particular treatment was perpetu-

words; his argument stands or falls on the

ated and disseminated to a new genera-

assumption that they must be both short

tion of students in Carol Eastman's 1975

and frequent. Eskimo words, however,

survey of linguistic approaches in anthro-

are the products of an extremely synthetic

pology, Aspects of Language and Culture.

morphology in which all word building is

Eastman summarizes the Sapir-Whorf

accomplished by multiple suffixation.

Their length is well beyond the limits of

hypothesis, which she calls "the world-

view problem," entirely by reference to

Zipf'? calculations. Furthermore, pre-

the snow example, quoting Brown's

cisely identical whole "words" are un-

"modifications" of Whorf's ideas

likely to recur because the particular

combination of suffixes used with a

"snow" root, or any other, varies by

(1975:76). Even more striking than the

distortion of Whorfs writing and think-

ing, which is implicit in the association of

speaker and situation as well as by syn-

him with it, is the powerful influence the

tactic role (Sadock 1980).

snow example exerts even on an experi-

A minimal knowledge of Eskimo gram-

enced linguistic anthropologist. With

mar would have confirmed the relevance

Brown's reference to "three words" only

of these facts to the central hypotheses,

six lines away, Eastman still asserts that

and would, moreover, have established

the even more relevant fact that there is

nothing at all peculiar about the behavior

or distribution of "snow words" in these

languages. The structure of Eskimo

grammar means that the number of

"words" for snow is literally incalculable,

419

"Eskimo languages have many words for

snow."

Thus is the complexity of the interre-

lations of linguistic structure, cultural be-

havior, and human cognition reduced to

"Eskimo words for snow." These and

other textbooks have disseminated mis-

This content downloaded from 170.140.26.180

170.14026.180 on
on Tue,
Tue. 15
15 Mar
Mar 2016
2016 21:08:12
21:08:12 UTC
LUC
All
All use
use subject
subject to
to JSTOR
JSTOR Terms
Terms and
and Conditions
Conditions

420

420 AMERICAN ANTHROPOLOGIST [88, 1986]

interpretations o f the example throughout the educated American population

since the late 1950s. Boas's small exampleironically, one intended as a caution
against superficial linguistic comparisonshas transcended its source and become part o f academic oral tradition.
Like folk beliefs about English vowels
(Walker 1970), tenaciously held folk theories about Arctic snow lexicon are not
easily contradicted. Unlike the vowel example, however, this folklore has not been
promulgated by secondary-school teachers but by anthropologists and linguists
who should know better and by professors
in other fields who first learned i t from
them.
Textbook references to the example
have reached such proliferation that no
complete inventory seems possible, b u t
examination o f a representative set reveals several common features: lack or inaccuracy o f citations; application o f the
example to diverse (and contradictory)
theoretical purposes; wholesale reanalysis of the example and its history. Thus,
according to a text on acoustics and
speech physiology (Borden and Harris
1980:4f.), the Whorfian hypothesis "was
based on comparative linguistic data
which show that languages differ in the
number of terms for such things as color
or snow." Even a recent introductory anthropology text cites the example as typical of those upon which Whorf founded
his conclusions about the effects o f linguistic categorization on thought (Cole
1982:69). From time to time, linguists
and anthropologists have attempted to
restore a sensible interpretation and
proper context to the example (e.g.,
Hymes 1967:213; Lyons 1981:306), b u t
these efforts have probably only succeeded in increasing its visibility. References in serious texts are testimony to the
example's widespread acceptance, b u t
they are only the most easily traceable of
its manifestations. Casual classroom use
is startlingly frequent and much more
often accompanied by apocryphal numbers, which usually range from about a
dozen to more than one hundred.
Even if academic use were suddenly to
cease, years o f carelessness have taken
interpretations of the example through-

out the educated American population

since the late 1950s. Boas's small exam-

ple--ironically, one intended as a caution

against superficial linguistic compari-

sons-has transcended its source and be-

come part of academic oral tradition.

Like folk beliefs about English vowels

(Walker 1970), tenaciously held folk the-

ories about Arctic snow lexicon are not

easily contradicted. Unlike the vowel ex-

ample, however, this folklore has not been

promulgated by secondary-school teach-

ers but by anthropologists and linguists

who should know better and by professors

in other fields who first learned it from

them.

Textbook references to the example

have reached such proliferation that no

complete inventory seems possible, but

examination of a representative set re-

veals several common features: lack or in-

accuracy of citations; application of the

example to diverse (and contradictory)

theoretical purposes; wholesale reana-

lysis of the example and its history. Thus,

according to a text on acoustics and

speech physiology (Borden and Harris

1980:4f.), the Whorfian hypothesis "was

based on comparative linguistic data

which show that languages differ in the

number of terms for such things as color

or snow." Even a recent introductory an-

thropology text cites the example as typ-

ical of those upon which Whorf founded

his conclusions about the effects of lin-

guistic categorization on thought (Cole

1982:69). From time to time, linguists

and anthropologists have attempted to

restore a sensible interpretation and

proper context to the example (e.g.,

Hymes 1967:213; Lyons 1981:306), but

these efforts have probably only suc-

ceeded in increasing its visibility. Refer-

ences in serious texts are testimony to the

example's widespread acceptance, but

they are only the most easily traceable of

its manifestations. Casual classroom use

is startlingly frequent and much more

often accompanied by apocryphal num-

bers, which usually range from about a

dozen to more than one hundred.

Even if academic use were suddenly to

cease, years of carelessness have taken

ANTHROPOLOGIST

1986]

their toll. Although awareness of the example is largely an artifact of higher education, the process of its transmission as
a folk myth no longer depends on that
context. The gradual filtering o f the example into the educated lay population
has established its vitality beyond university walls. Consider a diverse random
sample o f recent references: " m a n y
words" in the Journal of American Photography 3:1.19 (March 1984); "fifty" in Lanford Wilson's 1978 play The Fifth of July;
"nine" in a trivia encyclopedia called The
Straight Dope: A Compendium o f Human
Knowledge (Chicago Review Press, 1984),
which includes a droll explanation for the
variety: "[Eskimos] have a limited environment to talk about, so they have to
make up a lot of words to fill up their conversations"; a New York Times editorial
(February 9, 1984), citing Whorf in reference to a "tribe" distinguishing "one
hundred types o f snow"; Time's July 1,
1985, comparison of the Beirut glossary of
descriptive terms for shelling to the Eskimos' "many" words for snow; and the inevitable local television references t o
" t w o hundred w o r d s " d u r i n g winter
snow forecasts (e.g., WEWS-Cleveland,
1984).
How may we account for such remarkable persistence and ubiquity? No doubt
exoticism plays some role. Arctic peoples,
among the most easily recognized ethnographic populations, remain a poorly
understood group about whom other easy
generalizations are routine: they eat only
raw meat, they give their wives as gifts to
strangers, they rub noses instead of kissing, they send their elderly out on ice floes
to die. We are prepared to believe almost
anything about such an unfamiliar and
peculiar group. (See Hughes [1958] for
another example of scholarly misinterpretation of Eskimo culture.)
The context of such generalizations is
not altogether negative. There is in them
an element o f respect for the creative
adaptability of people who live in the almost unimaginably harsh Arctic environment. The tendency to inflate the numbers associated with the snow example is
a reflection o f admiration, not simply of
linguistic creativity but of human variatheir toll. Although awareness of the ex-

ample is largely an artifact of higher ed-

ucation, the process of its transmission as

a folk myth no longer depends on that

context. The gradual filtering of the ex-

ample into the educated lay population

has established its vitality beyond univer-

sity walls. Consider a diverse random

sample of recent references: "many

words" in the Journal ofAmerican Photogra-

phy 3:1.19 (March 1984); "fifty" in Lan-

ford Wilson's 1978 play The Fifth of July;

"nine" in a trivia encyclopedia called The

Straight Dope: A Compendium of Human

Knowledge (Chicago Review Press, 1984),

which includes a droll explanation for the

variety: "[Eskimos] have a limited envi-

ronment to talk about, so they have to

make up a lot of words to fill up their con-

versations"; a New York Times editorial

(February 9, 1984), citing Whorf in ref-

erence to a "tribe" distinguishing "one

hundred types of snow"; Time's July 1,

1985, comparison of the Beirut glossary of

descriptive terms for shelling to the Eski-

mos' "many" words for snow; and the in-

evitable local television references to

"two hundred words" during winter

snow forecasts (e.g., WEWS-Cleveland,

1984).

How may we account for such remark-

able persistence and ubiquity? No doubt

exoticism plays some role. Arctic peoples,

among the most easily recognized ethno-

graphic populations, remain a poorly

understood group about whom other easy

generalizations are routine: they eat only

raw meat, they give their wives as gifts to

strangers, they rub noses instead of kiss-

ing, they send their elderly out on ice floes

to die. We are prepared to believe almost

anything about such an unfamiliar and

peculiar group. (See Hughes [1958] for

another example of scholarly misinter-

pretation of Eskimo culture.)

The context of such generalizations is

not altogether negative. There is in them

an element of respect for the creative

adaptability of people who live in the al-

most unimaginably harsh Arctic environ-

ment. The tendency to inflate the num-

bers associated with the snow example is

a reflection of admiration, not simply of

linguistic creativity but of human varia-

This content downloaded from 170.140.26.180

170.14026.180 on
on Tue,
Tue. 15
15 Mar
Mar 2016
2016 21:08:12
21:08:12 UTC
LUC
All
All use
use subject
subject to
to JSTOR
JSTOR Terms
Terms and
and Conditions
Conditions

RESEARCH REPORTS

RESEARCH REPORTS 421

bility and survival as well. Taken in this

way, a self-evident observationwhat is
in our environment is likely to be reflected
in our languagehas become imbued
with exaggerated meaning. Through repetition in print or in lecture, the snow example has become cloaked i n scholarly
importance. I t s patina o f sophistication
reflects on the lecturer who appears to be
in possession o f specialized knowledge
and impresses any listener to whom close
attention to the details of language or culture m a y be a novel enterprise. I t s
"meaning" remains vague but seems generally simple: human beings are very different from each other. (Or, depending on
the version o f the example that is used
and the theoretical matrix in which it is
grounded, human beings are much the
same.)
Students constantly seek such simplicities and are abetted i n their quest b y
their teachers. Many facts about Eskimo
languages are fascinating and even astonishing. However, providing the detail, the
careful reasoning, and the technical sophistication required to draw conclusions
about language or culture or psychology
from those facts is a demanding task. Too
often the search for shorthand and simple-minded ways to talk about the complexities of language and culture results
in excessive reliance on inadequately detailed illustrations. I n the case o f the
snow example, sheer repetition reinforces
it, embedding it ever more firmly in folk
wisdom where i t is nearly immune to
challenge. Whenever issues in language,
culture, and thought are raised, a substantial proportion of listeners are unwilling to abandon the notion that " I t ' s all
just like Eskimos and snow."
Such a trivialization of the complexity
inherent in linguistic structures, linguistic behaviors, a n d t h e relationships
among them distorts the requirements of
research into these relationships by implying that counting words is a suitable
method of pursuing such investigations.
It may not be excessive to speculate that,
through this process, the example has
come to substantiate for some the bias
that these investigations are either impossible, irrelevant, or unscientific.
bility and survival as well. Taken in this

way, a self-evident observation-what is

In this twisted form, the snow example

duced as "proof' that Whorfs ideas were

superficial or lacked insight (cf. Lehman

1976:267). At a time when Whorfian hy-

etition in print or in lecture, the snow ex-

potheses are receiving renewed attention

ample has become cloaked in scholarly

among serious scholars whose ap-

importance. Its patina of sophistication

proaches to them are of exemplary rigor

in possession of specialized knowledge

and impresses any listener to whom close

(e.g., Bloom 1981), it is especially unfor-

tunate that the frivolousness of the snow

example should continue to be so promi-

attention to the details of language or cul-

nent and to obscure the true dimensions

ture may be a novel enterprise. Its

of such research problems. Relying in-

"meaning" remains vague but seems gen-

creasingly on the dubious value of surveys

erally simple: human beings are very dif-

and summaries instead of on original

ferent from each other. (Or, depending on

sources, even graduate students may

the version of the example that is used

never understand that Whorf's work-

and the theoretical matrix in which it is

like that of other linguistic anthropolo-

grounded, human beings are much the

gists-is not only not primarily con-

same.)

Students constantly seek such simplic-

cerned with snow words, but not even pri-

marily concerned with vocabulary. Such

ities and are abetted in their quest by

misunderstandings are especially hurtful

their teachers. Many facts about Eskimo

when they underpin much of the training

languages are fascinating and even aston-

given to today's students about the role of

ishing. However, providing the detail, the

linguistic investigations in anthropology.

careful reasoning, and the technical so-

Certainly, we have little control over

phistication required to draw conclusions

the processes of folklorization that can re-

about language or culture or psychology

move scholarly statements from their

from those facts is a demanding task. Too

rightful context and cause misinterpreta-

often the search for shorthand and sim-

tion. However, greater alertness to the

ple-minded ways to talk about the com-

dangers inherent in careless disregard for

plexities of language and culture results

the essential requirements of responsible

in excessive reliance on inadequately de-

scholarship might have prevented the

tailed illustrations. In the case of the

sorry evolution of the snow example

snow example, sheer repetition reinforces

within our own discipline. Now that we

it, embedding it ever more firmly in folk

have its history before us, perhaps it is not

wisdom where it is nearly immune to

too late to introduce yet another-and,

challenge. Whenever issues in language,

we may hope, final-context for it: the

culture, and thought are raised, a sub-

cautionary tale that serves to remind us of

stantial proportion of listeners are unwill-

the intellectual protection to be found in

ing to abandon the notion that "It's all

the careful use of sources, the clear pres-

just like Eskimos and snow."

Such a trivialization of the complexity

In this twisted form, the snow example

returns to the academic context and is adduced as "proof' that Whorl's ideas were
superficial or lacked insight (cf. Lehman
1976:267). A t a time when Whorfian hypotheses are receiving renewed attention
among serious s c h o l a r s whose a p proaches to them are of exemplary rigor
(e.g., Bloom 1981), it is especially unfortunate that the frivolousness of the snow
example should continue to be so prominent and to obscure the true dimensions
of such research problems. Relying increasingly on the dubious value of surveys
and summaries instead o f on original
sources, even graduate students may
never understand that Whorl's work
like that o f other linguistic anthropologistsis not only not p r i m a r i l y concerned with snow words, but not even primarily concerned with vocabulary. Such
misunderstandings are especially hurtful
when they underpin much of the training
given to today's students about the role of
linguistic investigations in anthropology.
Certainly, we have little control over
the processes of folklorization that can remove scholarly statements from their
rightful context and cause misinterpretation. However, greater alertness t o the
dangers inherent in careless disregard for
the essential requirements of responsible
scholarship might have prevented the
sorry evolution o f the snow example
within our own discipline. Now that we
have its history before us, perhaps it is not
too late to introduce yet anotherand,
we may hope, finalcontext for it: the
cautionary tale that serves to remind us of
the intellectual protection to be found in
the careful use of sources, the clear presentation of evidence, and, above all, the
constant evaluation of our assumptions.

in our language-has become imbued

reflects on the lecturer who appears to be

returns to the academic context and is ad-

in our environment is likely to be reflected

with exaggerated meaning. Through rep-

entation of evidence, and, above all, the

constant evaluation of our assumptions.

inherent in linguistic structures, linguis-

Notes

tic behaviors, and the relationships

Notes

among them distorts the requirements of

research into these relationships by im-

plying that counting words is a suitable

method of pursuing such investigations.

It may not be excessive to speculate that,

through this process, the example has

come to substantiate for some the bias

that these investigations are either impos-

sible, irrelevant, or unscientific.

Acknowledgments. A n earlier version o f this

paper was presented at the 1982 annual meeting of the American Anthropological Association. Since then, many students and colleagues have contributed new data, examples,
and editorial comments; although indebted to
them all, I owe particular thanks to Jill Brody,
Nora C. England, and Victor Golla. Any flaws
are, of course, my own responsibility.
Acknowledgments. An earlier version of this

paper was presented at the 1982 annual meet-

ing of the American Anthropological Associ-

ation. Since then, many students and col-

leagues have contributed new data, examples,

and editorial comments; although indebted to

them all, I owe particular thanks toJill Brody,

Nora C. England, and Victor Golla. Any flaws

are, of course, my own responsibility.

This content downloaded from 170.140.26.180

170.14026.180 on
on Tue,
Tue. 15
15 Mar
Mar 2016
2016 21:08:12
21:08:12 UTC
LUC
All
All use
use subject
subject to
to JSTOR
JSTOR Terms
Terms and
and Conditions
Conditions

422

422 AMERICAN ANTHROPOLOGIST [88, 1986]

'Various debates exist over the proper terminology and classification for languages o f
the Arctic. What is usually referred to as the
Eskimo language family encompasses several
important dialect divisions, most prominently
those of Yupik and Inuit-Inupiaq; for details
regarding both linguistic classification and description, consult Woodbury (1984) and the
extensive body of references cited therein.
"Eskimo" is used here in reference to the snow
example in recognition of the fact that those
who perpetuate it, like those who originated it,
fail to make any linguistically significant distinctions among speakers.
'There seems no reason to posit more than
two distinct roots that can be properly said to
refer to snow itself (and not, for example, t o
drifts, ice, storms, or moisture) in any Eskimo
language. I n West Greenlandic, these roots
are ganik 'snow in the air; snowflake' and aput
'snow (on the ground)' (Schultz-Lorentzen
1927; cf. Boas's data). Other varieties have
cognate forms. Thus, Eskimo has about as
much differentiation as English does for
'snow' at the monolexemic level: snow and
flake. T h a t these roots and others may be
modified to reflect semantic distinctions not
present in English is a result of gross features
of Eskimo morphology and syntax and not of
lexicon. A n y consequences that those grammatical differences may have for perception or
cognition remain undocumented.
'Various debates exist over the proper ter-

minology and classification for languages of

the Arctic. What is usually referred to as the

Eskimo language family encompasses several

important dialect divisions, most prominently

those of Yupik and Inuit-Inupiaq; for details

regarding both linguistic classification and de-

scription, consult Woodbury (1984) and the

extensive body of references cited therein.

"Eskimo" is used here in reference to the snow

example in recognition of the fact that those

who perpetuate it, like those who originated it,

fail to make any linguistically significant dis-

tinctions among speakers.

2There seems no reason to posit more than

two distinct roots that can be properly said to

refer to snow itself (and not, for example, to

drifts, ice, storms, or moisture) in any Eskimo

language. In West Greenlandic, these roots

are qanik 'snow in the air; snowflake' and aput

'snow (on the ground)' (Schultz-Lorentzen

1927; cf. Boas's data). Other varieties have

cognate forms. Thus, Eskimo has about as

much differentiation as English does for

'snow' at the monolexemic level: snow and

flake. That these roots and others may be

modified to reflect semantic distinctions not

present in English is a result of gross features

of Eskimo morphology and syntax and not of

lexicon. Any consequences that those gram-

matical differences may have for perception or

cognition remain undocumented.

References Cited

References Cited

Bloom, Alfred H.
1981 T h e Linguistic Shaping of Thought:
A Study in the Impact of Language on
Thinking in China and the West. Hillsdale, NJ: Laurence Erlbaum Associates.
Boas, Franz
1911 I n t r o d u c t i o n to The Handbook o f
North American Indians. Smithsonian
Institution Bulletin 40, Part I (Reissued by
the University of Nebraska Press, 1966.)
Borden, G l o r i a " , and Katherine S. Harris
1980 Speech Science Primer: Physiology,
Acoustics, and Perception of Speech. Baltimore: Williams & Wilkins.
Brown, Roger W.
1958 W o r d s and Things. New York: Free
Press.
Brown, Roger W., and Eric H. Lenneberg
1954 A Study in Language and Cognition.
Journal of Abnormal and Social Psychology 48:454-462.
Carroll, John B., ed.
1956 Language, Thought and Reality: Selected Writings of Benjamin Lee Whorl
Cambridge, MA: M I T Press.
Bloom, Alfred H.

1981 The Linguistic Shaping of Thought:

A Study in the Impact of Language on

Thinking in China and the West. Hills-

dale, NJ: Laurence Erlbaum Associates.

Boas, Franz

1911 Introduction to The Handbook of

North American Indians. Smithsonian

Institution Bulletin 40, Part 1. (Reissued by

the University of Nebraska Press, 1966.)

Borden, Gloria J., and Katherine S. Harris

1980 Speech Science Primer: Physiology,

Acoustics, and Perception of Speech. Bal-

timore: Williams & Wilkins.

Brown, Roger W.

1958 Words and Things. New York: Free

Press.

Brown, Roger W., and Eric H. Lenneberg

1954 A Study in Language and Cognition.

Journal of Abnormal and Social Psychol-

ogy 48:454-462.

Carroll, John B., ed.

1956 Language, Thought and Reality: Se-

lected Writings of Benjamin Lee Whorf.

Cambridge, MA: MIT Press.

ANTHROPOLOGIST

1986]

Cole, Johnetta B.
1982 Anthropology for the Eighties: I n troductory Readings. New York: Free
Press.
Eastman, Carol M.
1975 Aspects o f Language and Culture.
San Francisco: Chandler.
Hall, Edward T.
1959 T h e Silent Language. Garden City,
NY: Doubleday/Anchor Books.
Hughes, Charles Campbell
1958 A n o m i e , the Ammassalik, and the
Standardization o f Error. Southwestern
Journal of Anthropology 14:352-377.
Hymes, Dell
1967 Objectives and Concepts of Linguistic Anthropology. In The Teaching of Anthropology. David G. Mandelbaum, Gabriel W. Lasker, and Ethel M . Albert,
eds. Pp. 207-234. Berkeley: University of
California Press.
Lehman, Winifred P.
1976 Descriptive Linguistics. 2nd edition.
New York: Random House.
Lyons, John
1981 L a n g u a g e and Linguistics: A n I n troduction. New York: Cambridge University Press.
Pyles, Thomas
1964 T h e Origins and Development of the
English Language. New York: Harcourt,
Brace & World.
Sadock, Jerrold M.
1980 N o u n Incorporation in Greenlandic
Eskimo. Language 56:300-319.
Schultz-Lorentzen, C. W.
1927 D i c t i o n a r y of the West Greenlandic
Eskimo Language. Meddeleser om
land, 69. Copenhagen: Reitzels.
Smith, Dennis R., and L. Keith Williams
1977 I n t e r p e r s o n a l C o m m u n i c a t i o n :
Roles, Rules, Strategies and Games. 2nd
edition. Dubuque, IA: Wm. C. Brown.
Walker, Willard
1970 T h e Retention o f Folk Linguistic
Concepts and the T O & Caste i n Contemporary Nacireman Culture. American Anthropologist 72:102-105. (Reprinted in Nacirema: Readings on American Culture, James P. Spradley and
Michael A. Rynkiewich, eds., pp. 71-75,
Boston: Little, Brown, 1975.)
Whorf, Benjamin Lee
1940 Science and Linguistics. Technology
Review ( M I T ) 42:229-231, 247-248.
(Reprinted in Language i n Action, S.
Hayakawa, ed., pp. 302-321, 1941; Readings in Social Psychology, T. Newcomb
and E. Hartley, eds., pp. 207-218, 1947;

Cole, Johnetta B.

1982 Anthropology for the Eighties: In-

troductory Readings. New York: Free

Press.

Eastman, Carol M.

1975 Aspects of Language and Culture.

San Francisco: Chandler.

Hall, Edward T.

1959 The Silent Language. Garden City,

NY: Doubleday/Anchor Books.

Hughes, Charles Campbell

1958 Anomie, the Ammassalik, and the

Standardization of Error. Southwestern

Journal of Anthropology 14:352-377.

Hymes, Dell

1967 Objectives and Concepts of Linguis-

tic Anthropology. In The Teaching of An-

thropology. David G. Mandelbaum, Ga-

briel W. Lasker, and Ethel M. Albert,

eds. Pp. 207-234. Berkeley: University of

California Press.

Lehman, Winifred P.

1976 Descriptive Linguistics. 2nd edition.

New York: Random House.

Lyons, John

1981 Language and Linguistics: An In-

troduction. New York: Cambridge Uni-

versity Press.

Pyles, Thomas

1964 The Origins and Development of the

English Language. New York: Harcourt,

Brace & World.

Sadock, Jerrold M.

1980 Noun Incorporation in Greenlandic

Eskimo. Language 56:300-319.

Schultz-Lorentzen, C. W.

1927 Dictionary of the West Greenlandic

Eskimo Language. Meddeleser om Gr6n-

land, 69. Copenhagen: Reitzels.

Smith, Dennis R., and L. Keith Williams

1977 Interpersonal Communication:

Roles, Rules, Strategies and Games. 2nd

edition. Dubuque, IA: Wm. C. Brown.

Walker, Willard

1970 The Retention of Folk Linguistic

Concepts and the Tiyvtr Caste in Con-

temporary Nacireman Culture. Ameri-

can Anthropologist 72:102-105. (Re-

printed in Nacirema: Readings on Ameri-

can Culture, James P. Spradley and

Michael A. Rynkiewich, eds., pp. 71-75,

Boston: Little, Brown, 1975.)

Whorf, Benjamin Lee

1940 Science and Linguistics. Technology

Review (MIT) 42:229-231, 247-248.

(Reprinted in Language in Action, S. I.

Hayakawa, ed., pp. 302-321, 1941; Read-

ings in Social Psychology, T. Newcomb

and E. Hartley, eds., pp. 207-218, 1947;

This content downloaded from 170.140.26.180

170.14026.180 on
on Tue,
Tue. 15
15 Mar
Mar 2016
2016 21:08:12
21:08:12 UTC
LUC
All
All use
use subject
subject to
to JSTOR
JSTOR Terms
Terms and
and Conditions
Conditions

RESEARCH REPORTS

RESEARCH REPORTS 423

Collected Papers on M e t a linguistics,

Foreign Service Institute, 1952; and Language, Thought and Reality, John B.
Carroll, ed., pp. 207-219,1956.)
Woodbury, Anthony C.
1984 E s k i m o and Aleut Languages. I n
Handbook o f North American Indians,
Vol. 5: Arctic. David Damas, ed. Pp. 49
63. Washington, DC: Smithsonian Institution.
Collected Papers on Metalinguistics,

Foreign Service Institute, 1952; and Lan-

guage, Thought and Reality, John B.

Carroll, ed., pp. 207-219, 1956.)

Woodbury, Anthony C.

1984 Eskimo and Aleut Languages. In

Handbook of North American Indians,

Vol. 5: Arctic. David Damas, ed. Pp. 49-

63. Washington, DC: Smithsonian Insti-

tution.

423

enlarged breasts evolved as a signal t o

alert males to the nutritional status of females. Females who were better able to
build up fat and maintain it would have
more reserves to convert to parental investment during pregnancy and lactation. Breasts, with their high concentration of fat, would act to signal to males the
potential parental investment o f the female.
These scenarios are based on the assumption that permanently enlarged female breasts have arisen p r i m a r i l y
through the process o f sexual selection
whereby males choose female mates
based on their "attractiveness." Breasts
are seen as attractive either because they
mimic the buttocks, signal ovulation, or
signal potential female parental investment. As Hamilton points out, such scenarios view females as subject to selection
based o n m a l e e r o t i c r e a c t i o n s
(1984:658).
The validity of the concept of sexual selection to explain sexual dimorphism has
generated considerable debate since Darwin developed this idea over a hundred
years ago. Despite this controversy, Fedigan points out that,

enlarged breasts evolved as a signal to

alert males to the nutritional status of fe-

males. Females who were better able to

build up fat and maintain it would have

more reserves to convert to parental in-

vestment during pregnancy and lacta-

tion. Breasts, with their high concentra-

tion of fat, would act to signal to males the

potential parental investment of the fe-

male.

These scenarios are based on the as-

Evolutionary Perspectives on
Permanent Breast Enlargement
in Human Females
Evolutionary Perspectives on

Permanent Breast Enlargement

in Human Females

sumption that permanently enlarged fe-

male breasts have arisen primarily

through the process of sexual selection

whereby males choose female mates

based on their "attractiveness." Breasts

FRANCES E . M A S C I A - L E E S
FRANCES E. MASCIA-LEES

Social Science Division

Simon's Rock of Bard College

Social Science Division

Simon's Rock of Bard College

JOHN H . RELETHFORD
JOHN H. RELETHFORD

Department of Anthropology
State University of New York College
at Oneonta
Department ofAnthropology

State University of New York College

at Oneonta

are seen as attractive either because they

mimic the buttocks, signal ovulation, or

signal potential female parental invest-

ment. As Hamilton points out, such scen-

arios view females as subject to selection

based on male erotic reactions

(1984:658).

The validity of the concept of sexual se-

T O M SORGER

TOM SORGER

lection to explain sexual dimorphism has

Institute of Aging
Temple University

Institute ofAging

Temple University

generated considerable debate since Dar-

win developed this idea over a hundred

years ago. Despite this controversy, Fe-

Compared to the other primates, one

unique characteristic of Homo sapiens is the
existence o f p e r m a n e n t l y e n l a r g e d
breasts in human females. While several
authors have attempted to account for the
evolution of this anomaly, these explanations are invariably based on the problematic concept of breasts as sexual signals.
Morris's early account, for example,
suggests that with the advent of bipedalism, female breasts acted to shift the interest of the male to the front by acting as
a sexual signal that mimicked "the ancient genital display o f hemispherical
buttocks" (1967:75). More recently, Gallup (1982) has proposed that breasts signal ovulation, thus selecting for males
who could synchronize copulation with
ovulation. Short (1976) has suggested
that as hominid females became increasingly constrained in their movements due
to increased infant dependency, male parental investment became increasingly
necessary. Breasts became objects of attraction ensuring pair-bonding even before the female reached maturity. Finally,
Cant (1981) suggests that permanently
Compared to the other primates, one

unique characteristic of Homo sapiens is the

existence of permanently enlarged

breasts in human females. While several

authors have attempted to account for the

evolution of this anomaly, these explana-

tions are invariably based on the proble-

matic concept of breasts as sexual signals.

digan points out that,

Indeed, one of the most curious aspects of

the application of sexual selection principles
in the behavioral sciences is that in spite of
widespread discussions of the shortcomings
of this theory, the major tenets still operate
as hidden assumptions, or even axioms, in
much o f the writing on social behavior in
animals. [1982:271]

Indeed, one of the most curious aspects of

the application of sexual selection principles

in the behavioral sciences is that in spite of

widespread discussions of the shortcomings

of this theory, the major tenets still operate

as hidden assumptions, or even axioms, in

much of the writing on social behavior in

animals. [1982:271]

Morris's early account, for example,

suggests that with the advent of bipedal-

In terms of these shortcomings, many

authors (e.g., Fedigan 1982; Sayers 1982)
have pointed to the ethnocentrism inherent in these kinds o f explanations since
they quite consistently use a recent pattern o f sexual relations as a model for
early hominid sexual interactions (Hamilton 1984). For instance, enticing males
to contribute more in the way of parental
investment is central to several of the theories reviewed above. Such enticement,
gained through sexual appeal, is seen as
necessary since this model contends that
females invest more at the moment o f
conception than males due to the larger
size o f the ovum. Such a disproportionately large investment on the part of females portends greater parental investIn terms of these shortcomings, many

ism, female breasts acted to shift the in-

authors (e.g., Fedigan 1982; Sayers 1982)

terest of the male to the front by acting as

have pointed to the ethnocentrism inher-

a sexual signal that mimicked "the an-

ent in these kinds of explanations since

cient genital display of hemispherical

they quite consistently use a recent pat-

buttocks" (1967:75). More recently, Gal-

tern of sexual relations as a model for

lup (1982) has proposed that breasts sig-

early hominid sexual interactions (Ham-

nal ovulation, thus selecting for males

ilton 1984). For instance, enticing males

who could synchronize copulation with

to contribute more in the way of parental

ovulation. Short (1976) has suggested

investment is central to several of the the-

that as hominid females became increas-

ingly constrained in their movements due

to increased infant dependency, male pa-

rental investment became increasingly

necessary. Breasts became objects of at-

traction ensuring pair-bonding even be-

fore the female reached maturity. Finally,

Cant (1981) suggests that permanently

ories reviewed above. Such enticement,

gained through sexual appeal, is seen as

necessary since this model contends that

females invest more at the moment of

conception than males due to the larger

size of the ovum. Such a disproportion-

ately large investment on the part of fe-

males portends greater parental invest-

This content downloaded from 170.140.26.180

170.14026.180 on
on Tue,
Tue. 15
15 Mar
Mar 2016
2016 21:08:12
21:08:12 UTC
LUC
All
All use
use subject
subject to
to JSTOR
JSTOR Terms
Terms and
and Conditions
Conditions