Journal of Linguistics

http://journals.cambridge.org/LIN

Additional services for Journal of Linguistics:

Email alerts: Click here

Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

Brent Berlin and Paul Kay, Basic color terms:

their universality and evolution. Berkeley and
Los Angeles: The University of California
Press, 1969. Pp. xi + 178.

W. R. Merrield

Journal of Linguistics / Volume 7 / Issue 02 / September 1971, pp 259 - 268

DOI: 10.1017/S0022226700002966, Published online: 28 November 2008

Link to this article: http://journals.cambridge.org/

abstract_S0022226700002966

How to cite this article:

W. R. Merrield (1971). Review of D.W. Rajecki, M.E. Lamb, and P.
Obmascher 'Toward a general theory of infantile attachments: a comparative
review of aspects of the social bond' Journal of Linguistics, 7, pp 259-268
doi:10.1017/S0022226700002966

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/LIN, IP address: 138.251.14.35 on 12 Apr 2015

 REVIEWS
REFERENCES
Bh«t, D. N. S. (1970). A new hypothesis on language change. IL 31. 1-13.
Grammont, M. (1948). PhotUtique du grec ancien. Lyons.
Kiparsky, P. (1965). Phonological change. Unpublished doctoral dissertation, MIT.
Kiparsky, P. (1968). Tense and mood in Indo-European syntax. FL 4. 30-57.
Kurylowicz, J. (1967). A remark on Lachmann's law. HSCP fz. 295-299.
Thomas, A. R. (1967). Generative phonology in dialectology. TPhS 1967. 179-203.
Watkins, C. (1970). A further remark on Lachmann's law. HSCP 74. 55-65.

Reviewed by N. E. COLLINCE,
Center for
Linguistic Studies,
(Received 18 November 1970) University of Toronto.

Brent Berlin and Paul Kay, Basic color terms: their universality and evolution.
Berkeley and Los Angeles: The University of California Press, 1969. Pp. xi+178.

This book is made up of four sections and four appendices. Section one outlines
the experimental procedure, sets forth the definition of a 'basic' colour term, and
presents the general findings of the authors. Section two discusses a proposed
evolutionary hypothesis in terms of seven stages of structure. Section three
presents the colour terminology of each language used in the study, moving
from stage I languages to stage II languages, and so forth, through languages
exhibiting stage VII terminology. Section four is a summary of results and
'speculations'. Three of the appendices present technical details of the study,
while the fourth is a historical sketch of theory concerning colour vocabulary.
The main conclusions of the authors are best expressed in their own words
(104):

'First, there exist universally for humans eleven basic perpetual color
categories, which serve as the psychophysical referents of the eleven or fewer
basic color terms in any language. Second, in the history of a given language,
encoding of perceptual categories into basic color terms follows a fixed
partial order. The two possible temporal orders are:
purple
white"] pink
red -+ green -» yellow -» blue -> brown
blackj orange

and
'purple
whitel pink
red -• yellow -• green -* blue -» brown
blackj orange

259
 JOURNAL OF LINGUISTICS

Third, the overall temporal order is properly considered an evolutionary one;

color lexicons with few terms tend to occur in association with relatively
simple cultures and simple technologies, while color lexicons with many
terms tend to occur in association with complex cultures and complex
technologies (to the extent that complexity of culture and technology can be
assessed objectively).'

Only time will tell if these conclusions will be verified independently through
further experimentation; but whether the conclusions stand or not, the book
before us is a welcome study of colour perception and colour terminologies
across cultures. The material is presented in clear concise terms which allow
the reader to follow the argument, as well as see where problems of interpre-
tation lie. If the authors have 'accentuated the positive* in presenting their inter-
pretation of the facts, they seem to have made an honest effort to inform the
reader of counter-evidence as well.
The findings are based on experimental data collected by students and the
authors from native-speaking informants of twenty languages from a number of
unrelated language families. In most cases, apparently, only one informant was
used for each language, Tzeltal being the exception, where Berlin obtained data
from forty informants. The experimental data were supplemented by data from
seventy-eight other languages reported in the literature.
In order to collect the experimental data, a colour chart was prepared by
mounting 329 colour chips from the Munsell Color Company on stiff cardboard.
The set of chips are the same as those used by Lenneberg & Roberts (1956), with
the addition of nine chips of neutral hue (white, black, and greys). After eliciting
the colour terms from the informant (using a monolingual approach as far as
possible), the colour chips were overlaid with clear acetate, and the informant
was given a grease pencil and asked to map both the focal point and the outer
boundary of each of his basic colour terms on the colour chart. A printed facsimile
of this chart accompanies each copy of the book, folded in a pocket inside the
back cover.
Defining 'basic' colour terms is not without problems. The authors define a
basic term as one which exhibits four characteristics: it is monolexemic, it does
not map a colour area mapped by a more generic term, its use is not limited to
denote the colour of a small set of cultural objects, and it is a salient term across
informants. Additional criteria are invoked in doubtful cases. A doubtful term
should enter into the same syntactic constructions as established ones. A term is
suspect if it names an object such as gold, ash, etc. Recent foreign loan words may
be suspect. Morphological complexity is suspicious in cases where lexemic
status is not clear.
A difficulty that arises in using this definition when handling data is in dis-
tinguishing synchronic facts from diachronic facts, both of which are of interest
260
 REVIEWS

to the authors. Language change - phonological, morphological, semological -

is a very complex phenomenon. The result is simplification of a pattern at one
point, added complexity at another, and skewing across systems. The members
of a semantic set, for example, need not exhibit the same morphological
structure, since the semantic reference of a linguistic form is a function of
linguistic and non-linguistic context, and of language history.
It is not uncommon in the literature to find statements given in support of a
synchronic argument, say, concerning the structure of a certain lexical domain,
which cannot support the argument because of the skewing that is inevitably a
part of language change. An example that comes to mind is Goodenough's
argument (1965: 268) that differences in juncture and stress accent between
forms like grand nephew and grandson 'reveal' that such pairs of terms 'are not to
be equated', and that a different componential analysis of 'Yankee' kinship is
required if these facts are ignored (cf. his fn. 10).
To the outsider, who does not understand the Yankee system of kinship
terminology, facts about juncture and stress accent can only be etic (pre-systema-
tic) clues to help him discover the system. Such clues may or may not prove to
be relevant to the system because of semantic or phonological changes in the lan-
guage. This being the case, they cannot be invoked as more than peripheral
confirmation of an already established analysis.
On occasion it is necessary to use whatever clue is available. I myself have used
phonological clues in support of an historical hypothesis (Merrifield, 1966). This
is as problematical as using them to support hypotheses concerning present-
day systems. The unreliable nature of such clues must be kept well in mind, and
they must be set aside in favour of better evidence whenever possible.
This can be illustrated for the definition of BASIC colour terms by the Palantla
Chinantec term for blue (see end of this review). Palantla Chinantec has five
monomorphemic roots which map white, black, red, green, and yellow, but a
bimorphemic root which maps blue. The term for blue is a basic colour term.
Its anomalous structure alone is insufficient evidence to warrant either throwing
it out of the present system, or assuming the system has evolved from a five-term
system to a six-term system by the addition of this term. Now, all four Chinantec
languages for which I have experimental data exhibit six terms which map the
six terms in question; but in all the languages, the terms for the first five colours
are cognate, while no pair of languages shares cognate terms for blue! Add to this
the suggestion of Berlin and Kay that a term for blue does not appear in a system
until and unless there are five terms for white, black, red, green, and yellow.
Support the suggestion by data from a wide sampling of languages that five-
term systems exclude blue while six-term systems do not. Only then, does the
bimorphemic nature of the Palantla term for blue take on the nature of evidence
that it may have developed after the other terms. Note that in Tlacoatzintepec
Chinantec, the term for yellow is anomalously bimorphemic rather than that for

261
 JOURNAL OF LINGUISTICS
blue, although the theory and supporting evidence indicate that the yellow term
has been part of the system of colour terminology longer than the blue term.
An important theoretical point made by the authors centres in an attack against
what they call 'extreme linguistic relativity' (i), referring to the Sapir-Whorf
hypothesis. Their experimentalfindingsthat colour foci are fairly constant cross-
culturally has confirmed for them their initial intuitive feelings concerning the
translatability of colour terms. In reading their book, I get the feeling that they
have made something of a straw man of the Sapir-Whorf view. Gleason's book
(1961) is described (fn. 1) as 'perhaps the most influential of standard linguistics
texts in the United States', and Eugene Nida as 'perhaps the leading American
authority on translation'. Then quotations from the text and an article by Nida
(1959) are presented in a context which implies that the two men espouse the
Sapir-Whorf view. But neither man is in fact addressing his remarks to that view
as a glance at the original contexts will suffice to show.
But then, I am not an adherent of the Sapir-Whorf view, and it may be more
widely accepted than I imagine. One often sees references to Whorf's writings,
but more often than not these can be written off as attempts to show erudition
through knowledge of the literature. Longacre (1956) expressed quite well some
of the weakness of the Whorfian view, and said some interesting things about
colour terms which anticipate the work under review (303^:'... we should not
assume without investigation that there may not be certain color contrasts almost
universally posited. To controvert the latter it would not be sufficient to show
that a given language uses the same term, say, for "blue" and "green", if in
phrase formation or compounding "grass-X" is distinguished from "sky-X" or
something of the sort. Such studies might indicate that Language (with a capital
L) segments certain phases of reality in a rather consistent manner . . . and that
this points to a "natural" segmentation of reality perhaps inescapable for us as
human beings.'
The evolutionary view of the authors (quoted verbatim at the beginning of this
review) is stated in sufficiently loose terms as to make it relatively vacuous. Since
we do NOT yet have objective means to assess the relative complexity of culture
and technology, and since in the authors' own findings Zuni and Dinka, among
others, are grouped with English, Russian, Spanish, and Japanese as displaying
stage VII systems, the view is best left in such terms for the present. Similarly,
a cautious comparison of the development of colour lexicons with the early
Jakobson-Halle (1956) theory of phonological development must perforce
remain cautious.
The pendulum has swung in Anthropology and Linguistics from the Boasian
emphasis on description to a strong emphasis on theorizing. The authors are
sensitive to the dangers of theorizing without a sufficiently large empirical base
(Berlin, 1970), but they are still willing to go out on a theoretical limb. It's a sign
of the times.
262
 REVIEWS

The experimental data upon which the study rests are limited - only twenty
languages, usually only one informant per language with the exception of Tzeltal,
as far as I can tell. Granted that seventy-eight languages were added to the
sample from the literature, more experimental data are in order to confirm the
hypotheses. But an excellent start has been made. We are in the debt of Berlin and
Kay. The price of the book seems high for its size (perhaps due in part to the
excellent colour chart which accompanies it), but the reader will be rewarded in
the reading of it.
Inasmuch as the reviewer resides at a research centre of the Summer Institute
of Linguistics (SIL) frequented by native speakers of several languages of Mexico,
he did some experimenting of his own along the lines set down by Berlin and
Kay in an attempt to find either additional verification or contradiction of their
conclusions. Eight SIL teams, all working in Otomanguean languages, had in-
formants at the centre at this time. I tried to gather the data under conditions
similar to those of the pilot study, but there was some lowering of standards.
Most important, I used only the printed colour chart that accompanies the book
rather than the Munsell colour chips used by Berlin and Kay. This created the
minor technical problem that, almost without exception, the informants accepted
only the colour of the glossy white paper of the chart as a referent for their terms
for white. The lightest shades of neutral hue were almost invariably rejected as
other than white. More than one informant had difficulty rinding what he con-
sidered true referents for certain colour terms. They apparently sought brighter
colours than they found on the chart. Perhaps a high-intensity artificial light
source, such as that used by Berlin and Kay, would have solved this problem
(did they really use that among the Tzeltals?). Our light source was a bright
midday Oaxaca sky, the remarkable qualities of which are unknown to most
denizens of large American cities like Berkeley, but which undoubtedly can be
improved upon for certain limited purposes such as mapping colour terms on a
printed colour chart. Instead of using clear acetate to protect the chart, I prepared
strips of thin clear plastic cut from plastic bags available locally. A black indelible
felt marking pencil was used rather than a grease pencil.
The experimental data gathered by the reviewer follow. I found them to fall
quite in line with the ordering suggested by the authors. They are presented
below as an addition to their initial sample, and as an illustration of their format.
I have not gone to the trouble of preparing charts of the foci and ranges for each
term as the authors have done for the twenty languages they investigated experi-
mentally. I hope to send Berlin and Kay copies of the mappings done by our
informants to add to their file.

Language Xochistlahuaca Amuzgo (Stage VII)

Linguistic class Otomanguean, Amuzgoan

H 263
 JOURNAL OF LINGUISTICS

Area Mexico (Abasolo, Guerrero)

Source Carmelo Valtierra L<Spez, Amy Buaernschmidt
Basic colour terms kancHi? [white]
ntom [black]
wee [red]
BW, kabn? [green]
kaha [yellow]
tshs* [blue]
kalu, kahue [brown]
kacioo? [purple]
Una [pink]
handya* [orange]
tsa [grey]
Discussion The two terms for green have different foci but the same range.
kahti? is the name of a parrot, but the informant insisted it was the more basic
term, wii having a strong connotation of 'nature' or green, living things. The term
for orange also means picoso 'peppery hot'. The term for grey includes the pale
blues as well as the neutral hues. One of the terms for brown is from Spanish
cafe. It was elicited early, kalu being given as an afterthought after the mapping
of the terms was complete. The informant also offered the term koine for the
colour of metallic or transparent objects that reflect light.

Language Comaltepec Chinantec (Problematic stage V)

Linguistic class Otomanguean, Chinantecan
Area Mexico (Ixtldn de Juirez, Oaxaca)
Source Manuel Rivera L6pez, Judith Anderson, Wanda Pace
Basic colour terms tee, tiang [white]
df, W, leng? [black]
y°"g M l
rA>, rMigf [green]
nee", niang [yellow]
nlfl, mongf [blue]
Hyd, ?leng* [grey]
Discussion All Chinantec languages distinguish two genders, inanimate and
animate. There are thus two forms for each colour term in each language. Most
informants automatically gave only the inanimate forms. The Comaltepec in-
formant, however, presented both sets, indicating adjacent but non-overlapping
ranges for each member of a pair. In further experimentation with several in-
formants for each Chinantec language, the animate terms should be mapped on
one piece of acetate or plastic and the inanimate terms on another. It is very

264
 REVIEWS

possible that the reference of animate terms does not correspond exactly to that
of inanimate terms either in focus or range. There are two inanimate Comaitepec
terms for black. ?urif is cognate with a Palantla term which denotes dark, or black
hair. The reference has apparently flip-flopped in Comaitepec, le ? having reference
to dark colours, and ?wif to true black. The Comaitepec terms for blue are
cognate with a Palantla term meaning dark (as in darkness). The two terms for
grey are atypical in that they map the same area. They are also problematic for
the proposals of Berlin and Kay. Brown is supposed to be the next colour term
to appear after blue, and although the informant did refer to this seventh colour
in Spanish as media cafe" he mapped it onto a colour area far removed from the
browns. r{?, from Spanish rubio, is the colour of blond hair.

Language Palantla Chinantec (Stage V)

Linguistic class Otomanguean, Chinantecan
Area Mexico (Valle Nacional, Tuxtepec, Oaxaca)
Source Ernesto Perez Martinez, Alfred Anderson, William R
Merrifield
Basic colour terms teg2 [white]
&'» [black]
gyfg2 [red]
r»?2 [green]
nig2 [yellow]
mi*cyu3, mi*cyp2 [blue]
Discussion There are two dialectal variants of the term for blue. Pretonic mi*-
is a common stem formative.

Language Tepetotutla Chinantec (Stage VI)

Linguistic class Otomanguean, Chinantecan
Area Mexico (Usila, Tuxtepec, Oaxaca)
Source Leoncio Hernandez Osorio, David Westley
Basic colour terms tiog2 [white]
Ua*2 [black]
gfigi [red]
ri?2 [green]
niog2 [yellow]
lip [blue]
si92 [brown]

Language Tlacoatzintepec Chinantec (Partial stage VII)

Linguistic class Otomanguean, Chinantecan

265
 JOURNAL OF LINGUISTICS
Area Mexico (Cuicatlan, Oaxaca)
Source Roberto Regules, Jacqueline Bernhardt, Mabel Lewis
Basic colour terms tyu* [white]
lya** [black]
dau3* [red]
re** [green]
meghuhfi [yellow]
Ao?» [blue]
sag** [brown]
Geg^ purple]
Discussion The pretonic meg*- of the term for yellow is a stem formative.
Spanish cafe, accommodated to Chinantec sound patterns, denotes 'coffee' or
the same portion of the colour map as sag**.

Language Ayutla Mixtec (Partial stage VII)

Linguistic class Otomanguean, Popotecan, Mixtecan
Area Mexico (Allende, Guerrero)
Source Esteban Morales, Leo Pankratz
Basic colour terms ya?a} [white]
tu*u*\ ndya^aK [black]
kwaS*a^ [red]
kuriU** [green]
kwaxax [yellow]
ndi>*e*i [blue]
brown]
[grey]
Discussion The terms for white and black are better glossed 'light' and 'dark'
colours. The two terms for black map the same colour area, tu^u*1 is related to
tffiuto 'soot'. The informant did not accept any colour on the chart for his terms
for blue and brown. He did accept a royal blue found on a book cover which is
very close to the blue given by the Chayuco Mixtec informant. The brown term
was described as the colour of a wooden table left out in the weather, but the
informant also accepted the colour of the socks the reviewer was wearing which is
roughly that of the focal point given for Tepetotutla Chinantec brown. £«>/»> is
said to be a mixed colour, something like the colour of ashes. Several modifiers
were suggested to identify fresh, pale, or strong colours of various hues.

Language Chayuco Mixtec (Partial stage VII)

Linguistic class Otomanguean, Popotecan, Mixtecan
266
 REVIEWS

Area Mexico (Jamiltepec, Oaxaca)

Source Cresencio Serrano, Audrey Johnson, Brenda Pensinger
Basic colour terms kwisi [white]
t$ [black]
ktvd?d [red]
ktoii [green]
kwdd [yellow]
aOu [blue]
ndydd [brown-purple]
yd?d [purple]
tisd?d [grey]
Discussion The term for blue is from Spanish azul. That for brown-purple was
given a disjunctive mapping: a deep brown between red and green, and a deep
purple midway between and below blue and violet. The term for purple maps
a deep purple to the right of this area on the chart below the lavenders and
violets. The term tlUnda 'shellfish' is used for still another purple, ndydd is
cognate with the Ayutla term which is best glossed as 'dark' colour. A possible
internal reconstruction which would accord with the developmental hypothesis of
Berlin and Kay is as follows: ndydd originally mapped all the dark colours. The
word for soot (tfy) was taken over for true black, being modified as a colour word
to tpff. The reference of ndydd was then gradually limited to its present colour
referents.

Language Mixtepec Mixtec (Stage V)

Linguistic class Otomanguean, Popotecan, Mixtecan
Area Mexico (Juxtlahuaca, Oaxaca)
Source Agustfn L6pez Reyes, Thomas Ibach
Basic colour terms yaa, kivici [white]
nlaH [black]
krva?a [red]
kwii [green]
ktvpa [yellow]
ntaa [blue]
Discussion Two terms, both having cognates in sister languages, map the
same area of white, but refer to distinct sets of objects, yaa is used for paper,
human skin, etc.; kivici refers to the colour of certain beans, etc.

REFERENCES
Berlin, B. (1970). A universalist-evolutionary approach in ethnographic semantics. In A.
Fischer (ed.), Current directions in anthropology. Bulletins of the American Anthro-
pological Association, Vol. 3, No. 3, Part 2, 3-18.

267
 JOURNAL OF LINGUISTICS

Gleason, H. A., Jr. (1961). An introduction to descriptive linguistics. New York: Holt,
Rinchart & Winston. (Rev. ed., 1966.)
Goodenough, W. H. (1965). Yankee kinship terminology; a problem in componential
analysis. In E. A. Hammel (ed.), Format semantic analysis. Special Publication of the
American Anthropological Association, Vol. 67, No. 5, Part 2, 259-287.
Jakobson, R. & Halle, M. (1956). Fundamentals of language. The Hague: Mouton.
Lenneberg, E. H. & Roberts, J. M. (1956). The language of experience: a study in method-
ology. Memoir 13 of IJAL.
Longacre, R. E. (1956). Review of W. M. Urban, Language and reality: the philosophy of
language and the principles af symbolism; and of B. L. Whorf, Four Articles on meta-
linguistics. Lg 3a. 298-308.
Nida, E. A. (1959). Principles of translation as exemplified by Bible translating. In R. A.
Brower (ed.), On translation. Cambridge: Harvard University Press, pp. 11-31.
[Reprinted in The Bible Translator 10. 148-164 (1959).]
Merrifield, W. R. (1966). Linguistic clues for the reconstruction of Chinantec prehistory.
In A. Pompa y Pompa (ed.), Summa Anthropologica en homenaje a Roberto J. Weitlaner.
Mexico, D. F., 579-596.

Reviewed by W. R. MERRIFIELD,
Summer Institute of Linguistics,
{Received 12 October 1970) Mitla, Oaxaca, Mtxico.

Mario Wandnuzka, Sprachen: vergleichbar und unvergleichlich. Munich: R.

Piper & Co. Verlag, 1969. Pp. 541.

Comparative work on translations from and into German, English and the major
Romance languages (French, Italian, Spanish, Portuguese) has always been M.
Wandruszka's main field of research. Many of the articles he has written within
the last ten years deal with this topic and most of the dissertations he has super-
vised during this period are concerned with comparative work of this kind. The
book under review is the magnum opus in which the results of this research are
presented in a detailed and comprehensive description.
The book comprises thirty chapters of text, a short preface, a short summary
and two lists of references. In the first list we find a few general books on transla-
tions and all the dissertations mentioned above, the second enumerates all the
books from which the material is taken. This list comprises seventy-eight titles
plus their translations. All of these are well-known literary works. The author
has selected original texts from all six languages listed above and compares them
with their translations into the other five. Thus the focus of comparison changes
continuously, depending on the point under discussion.
Wandruszka's objectives, his reasons for writing the book, are stated in the
preface. It is the main purpose of the book to investigate the stylistic resources
and peculiarities of the languages under comparison ('was die Sprachen fur
uns leisten und wie sie es leisten', 7: 11). A few further aims are pursued too.
The author wants to show that languages are irregular to a large extent, that they
contain synonymies, polysemies, redundancies and deficiencies. The author is

268