Cross-Cultural
10.1177/1069397104267886
Jameson
/ REMARKS
Research / February 2005
On the Role of Culture
in Color Naming: Remarks
on the Articles of Paramei,
Kay, Roberson, and Hardin on
the Topic of Cognition,
Culture, and Color Experience
Kimberly A. Jameson
University of California, Irvine
University of California, San Diego
Comment on the preceding articles is now provided to highlight
some open questions and controversies raised in this issue.
The primary goal is to provide a context for examining the longaccepted views from the cross-cultural color categorization and
naming literature that are presented in the articles. For this reason alone, the articles by Hardin and Kay—which present new
analyses compatible with the established theory in the area—
receive more comment than do those of Roberson and Paramei—
which, comparatively speaking, are less concordant with the
established theory and reflect more culturally specific views of
color naming phenomena. In my opinion, each of the four articles
in this issue contributes important perspectives on cross-cultural
color naming research. Although the order of articles in this volume is Paramei, Kay, Roberson, and Hardin, the comments presented here are organized to emphasize the authors’ common viewCross-Cultural Research, Vol. 39 No. 1, February 2005 88-106
DOI: 10.1177/1069397104267886
© 2005 Sage Publications
88
Jameson / REMARKS
89
points and to provide an ordered development of the issues the
articles raise.
CLYDE L. HARDIN’S
“EXPLAINING BASIC COLOR CATEGORIES”
The first paragraph of Hardin’s article suggests that to assert
that no significant properties of color categories are biologically
based (“as some cultural relativists do”) is to disregard all that perceptual psychologists know about color vision mechanisms. Most
researchers would concur that sensible cultural relativists agree
with universalists that biology serves as an important constraint
for color perception. For example, consider that so-called colorblind individuals (who lack a type of color-sensitive cell in their retinae) perceive a reduced color category structure compared with
otherwise normal individuals who possess a full complement of
color-sensitive cells. In such a case, obviously retinal biology
shapes perceptual experience in the color-blind dichromat.
The remaining portion of Hardin’s argument is more difficult to
accept, however, because although his first sentence acknowledges
that there are “many properties of color categories that are a function of language and culture,” he does not explore these much, opting instead to present support for the view that pan-human color
processing is the basis for the universality seen in color naming
and categorization behaviors.1 A very different view, now gaining
popularity among researchers, considers visual processing models
inadequate as the primary basis for explaining complex color categorization behaviors seen within and across cultures. As an example, consider again the dichromat: Within a given culture, the
anomaly of “color blindness” does not prohibit an affected observer
from properly learning and using the culturally appropriate color
terms and concepts (e.g., polar opposition of red and green) that
refer to the color appearances that a dichromat observer has difficulty perceptually disambiguating (Izmailov & Sokolov, 1992;
Marmor, 1978; Shepard & Cooper, 1992). This ability to adequately
use and comprehend color language and color categories in the
absence of perceptual correlates that differentiate between
exemplars of various color terms and categories provides a clear
indication that a substantial (if not central) portion of color categorization conceptual behavior is not attributable to the usual biologically based visual processing explanations.
90
Cross-Cultural Research / February 2005
In my opinion, both individual and cultural categorization
behaviors require much more comprehensive explanations than
that Hardin offers. Thus, Hardin’s “accepted view” perspective is
incomplete as an explanation of such behaviors, despite his correct
assessment that empirical investigations of visual processing
mechanisms make important contributions toward understanding
the constraints and nonarbitrary features inherent in individual’s
early visual processing.
If the basis for cross-cultural color naming regularity is not universal biology, then what explains the frequently observed linkages described by Hardin between color naming universals and
opponent-process-based perceptually elementary hues? To pursue
this, we need to separately examine the links in the chain of the
accepted explanation, essentially asking whether either of the
nature or nurture views invoked by Hardin can be ruled out as possible (or partial) explanations for the apparent universality of perceptually elementary hues that seems to transcend the idiosyncrasies of particular languages (Hardin, 2005 [this issue]). Here, the
goal is not to prove a “correct” alternative explanation for the phenomenon; rather, it is to ask whether a plausible alternative view
can be specified, and if so, whether that alternative explanation
can be similarly analyzed and ruled out as incorrect compared
with the accepted view Hardin describes.
A simplified statement of the accepted chain of reasoning for the
biologically based explanation that Hardin describes can be summarized by two statements:
A. The hues most frequently and reliably named by individuals, and
most commonly named across cultures, are the perceptual elemental hues.
B. Pan-human uniformity in color vision (typically linked by color
naming researchers to opponent-process visual mechanisms) is the
basis for the widely shared significance of perceptual elemental
hues (as compared with nonelemental hues).
The problem is that although Hardin’s analyses (and much
work in the Berlin and Kay tradition) show that Statement A is often supported, it can be argued that the explanatory causal connection in Statement B is not supported for (at least) two reasons:
1. Existing neurophysiological and psychophysical data indicate that
individual variation in color vision mechanisms preclude the exist-
Jameson / REMARKS
91
ence of uniform “elemental” stimuli that could produce unitary perceptual experiences across observers.
2. Existing psychophysical data indicate that there is considerable
individual variation in empirically determined perceptual elemental hue settings across observers and to some extent across cultures
(Kuehni, 2004; Webster et al., 2002).
Given that the biologically based rationale in Statement B is not
the real explanation for observations in A, then why do so much existing data seem to suggest that “the hues . . . most commonly
named across cultures are the perceptual elemental hues”? Following Hardin’s usage, can we rule out the nurture-type explanation
for Statement A? Or can nurture or cultural influences produce
scenarios whereby perceptual elemental colors tend to be those
most commonly named across many cultures?
At least one plausible nurture scenario—one emphasizing color
naming and categorization as a cultural practice or artifact—
explains the tendency to name elemental colors. That is, consider
those cultures for which a linguistic gloss for color exists as an
autonomous lexical construct. In such a culture, the gloss for color
certainly operates as an important part of that culture’s strategy
for transmitting (to new generations of individuals) knowledge of
all the separate linguistic terms that make up the culture’s color
lexicon. Under one such transmission strategy, individuals undergoing the process of acquiring a culture’s color lexicon might first
implicitly master the concept of color by recognizing that what is
glossed by the term color is the culture’s particular notion of how
something appears or “looks.” Many languages that are classified
at or above Stage V (using the original framework of Berlin & Kay,
1969, or recent revisions) have a separate abstract concept for color
represented by a superordinate linguistic term; however, many
unwritten languages—those such as Bellona (Kuschel & Monberg
1974) and Yélî Dnye (Levinson, 2000)—do not have a native construct for color indexed by an unborrowed superordinate lexical
term and thus may not have an abstraction for color integral to
their cultural transmission of the color lexicon.
For cases where a superordinate term exists, during transmission of the concept color, differentiation of “pure” (or “unitary”) or
“blended” (or “binary”) forms of color may occur. A tendency to
begin by transmitting the simplest unitary cases first would imply
that “pure” cases of black, white, red, yellow, green and blue
(being appearances typified by less complex, or fewest number of,
92
Cross-Cultural Research / February 2005
dimensional components) would be learned by individuals first.
Indeed, this could be an advisable strategy for the cultural transmission of an abstract concept such as color, because using a simple
set of items sharing common dimensional features (integral to the
construct) may be optimal for the comprehension of the concept by
individuals (cf. Feldman, 2003).
Under this first transmission scenario, colors that are unlikely
to be classified as perceptually elemental hues would be categories
of greater complexity due to additional properties (or dimensions)
as defining aspects (Feldman, 2003). For example, appearances
glossed by the term gold might additionally require shimmer as a
defining aspect in addition to the typical lightness, purity, and hue
aspects. Similarly, color categories composed of exemplars that do
not reflect a representative range of exemplar variation along
principle dimensions (e.g., saturation and lightness; see Jameson,
1997) in the space would require additional specificity regarding
the defining aspects of the category. Thus, a defining aspect of a
greenish-yellowish, “chartreuse,” category could be distinctive
lightness and saturation levels—traits that are characteristic of
chartreuse exemplars (i.e., chartreuse exemplars rarely appear
dark and seldom unsaturated). These unipolar lightness and saturation features (or defining aspects) of chartreuse category exemplars imply that the category is conceptually more complex than
another color category that is composed of exemplars representing
a range of values from the principle dimensions of the space (i.e., a
full range of lightnesses, saturations, etc.) and which thereby is
more simply and strictly summarized by a dominant hue feature
as the defining aspect of the category (Jameson’s [in press–b]
contribution in the next issue discusses this further).
Although the foregoing scenario for cultural transmission in the
context of a superordinate term for color may seem slightly
counterintuitive, it reflects one plausible nurture-based hypothesis and is consistent with various developmental data indicating
unexplained difficulties in color-term acquisition (summarized by
Roberson, 2005 [this issue]).
As a consequence of this argument, one could assert that (a) a
color is more likely to be considered a perceptual elemental hue
when color space features needed to define that color category’s
best examples are minimal; (b) cultures’ linguistic systems play a
role in the definition of elemental colors when, as described, cultural variation in the occurrence of a native lexical term for a
superordinate color concept influences which color appearances
Jameson / REMARKS
93
are naturally assigned lexical labels and the cultural transmission
strategy for color lexicons; and (c) cultural practices and influences
can play a role in the definition of elemental colors when the
dimensions culturally deemed to have high pragmatic value (e.g.,
cultural preferences for hues; Saito, 1996a, 1996b) are strongly
biased in ways that de-emphasize the dimensions that the elemental color categories classically represent. In summary, it is suggested here that nurture-type explanations should not be ruled out
as a possible basis for why perceptual elemental hues are so widely
named across cultures and that nurture-type explanations are
plausible alternatives for explaining the highly cognitive and cultural behaviors observed in color categorization and naming
results relevant to color appearances popularized by Hering’s
theory of color opposites.
Note the relevant detail that some informants used for the languages described in Berlin and Kay (1969) were bilingual, with
English as one of their languages. Those authors acknowledge,
“While it can be argued that bilingualism in English affects the
results to some extent,” they do not support the possible objection
that bilingualism of their informants distorted their findings
(p. 12). Nevertheless, English bilingualism among those subjects
implies an awareness of English’s superordinate-term color concept, even though the foreign language tested may have lacked a
native term for that abstract concept. Whether assessment of
naive foreign language speakers who have no term for color in
their native language (and who have no familiarity with other languages with a term for that concept) would show a different tendency regarding the naming of elemental perceptual hues remains
an open empirical question.2
For languages that lack a superordinate color term, a nurturebased conceptual explanation can also play a role in the crosscultural similarity in elemental hues loci (e.g., those identified by
MacLaury, 1997b, and discussed by Hardin, 2005). For example, in
such languages, elemental hues can arise as salient colors for individuals and groups if either polar opposition or comparative color
purity relations are linguistically salient. Thus, MacLaury’s crossculturally observed elemental hue positions may arise through a
shared strategy of linguistically representing optimized polar
opposites formed by the purest, unblended, extremes available in
cognitive color appearance space. These two aspects of elemental
hues, (a) color “purity” and (b) polar-opposition pairing with
another “pure” color, may be a large part of developing a relational
94
Cross-Cultural Research / February 2005
structure that distinguishes elementals from other nonelemental
hues (see Jameson, in press–a, in press–b).
Thus, it is plausible that the difference distinguishing elemental hues from other hues could be a conceptual, or cognitive, difference, as opposed to a difference arising from a neural processing
substrate or similar “channel” linked to opponent-process chromatic response functions present in low-level color vision mechanisms. Indeed, the latter linkage is seriously undermined by the
fact that physiological response curve locations that the theory
links to unique hue experiences do not actually coincide with spectral positions described as pure elemental colors (see Jameson &
D’Andrade, 1997, pp. 301-308, for details; Gegenfurtner & Kiper,
2003).
Fortunately for the received view of color naming theory (i.e.,
the new variants of the Berlin & Kay, 1969, theory; see Kay & Maffi,
1999), describing elemental hues in this cognitive/conceptual way
remains entirely compatible with Hering’s original opponentcolors ideas (especially for color naming systems sharing the complexity of English), as well as with the principles on which the Natural Color System (NCS) is based (see Hardin’s [2005] discussion).
The only difference in the alternative explanation provided here is
that cross-cultural elemental hue naming and elementary color
salience are not determined by “biologically based” channels. Thus,
color-naming researchers can continue to use the Hering elementary colors as part of the explanation for color categorization and
naming results, as long as they do not rely on the typically implied
opponent-process neurophysiological mechanisms as the basis for
such results.
Hardin states: “Names for the Hering elementary colors are necessary and sufficient for naming all of the colors, a fact that justifies singling them out” (2005, this issue). The rationale for asserting this is strained because the same argument would apply
equally well to a subset of three elementary colors consisting of
red, yellow, and blue. Indeed, where strictly chromatic proportions
are concerned (i.e., those Hardin discusses for the NCS color samples), one can describe color appearances equally well as ratios of
only red, yellow, and blue (as Goethe and others have proposed),
similar to the way they are typically described using ratios of red,
yellow, blue, and green. In such a scenario, green colors could simply be described as proportional combinations (i.e., product mixtures) of blue and yellow. Thus, all green-containing colors, and any
other color in the NCS atlas, can be described just as well using
Jameson / REMARKS
95
three red, yellow, and blue proportions, as they can using Hering’s
four elementary colors.3 The question then is, Why is this strict
summing of named proportions for red, green, yellow, and blue primaries preferred as the operational definition for estimating NCS
color proportions when a simpler, more parsimonious subset of primaries is given by three primaries (red, yellow, and blue)? In my
view, these three primaries seem to serve just as well as a set of
“necessary and sufficient” components for describing NCS color
samples. The answer to this question may follow from a general
tendency to extend psychophysical color cancellation primaries
and their theoretical constructs into color naming research. In any
case, I do not think the argument Hardin provides for “singling
out” the names of Hering elementary colors is any more compelling
than one that could be developed for the alternative set of primaries suggested.
Hardin’s article also compares data from human and chimpanzee color naming (Matsuzawa, 1985). Hardin argues that similar
chimp and human color naming behaviors support the notion that
human color naming and categorization are based on the visual
processing mechanisms shared by both species. It is true that
chimpanzee and human color vision systems are highly comparable; however, another plausible nurture-like interpretation of
Matsuzawa’s (1985) findings is that these chimp and human color
naming data are similar because the chimp tested has acquired
through training a similarity structure for the Munsell colors that
resembles the similarity structure of human observers. Indeed, the
chimp’s acquisition of this similarity structure is implicit in its successful training (the set of training colors varied along all three
dimensions of Munsell hue (H), value (V), and chroma (C), and only
these three dimensions). Thus trained, generalizing such a similarity structure to categorizing other Munsell samples would minimally consist of recognizing stimulus differences in hue (H), lightness (V), and saturation (C)—a task for which the primate visual
system is ideally suited—and appropriately sorting such samples
into highly overlearned hue “bins.” Pigeons, which have a very different visual system from humans (see Varela, Palacios, & Goldsmith, 1993), can also be trained to sort color stimuli into the 11
basic color categories. Indeed, pigeons’ internal representation for
basic colors forms a hue circle, suggesting “that similarities among
colors for pigeons are organized in a manner generally like those
for humans. This is quite interesting in view of the anatomical differences between the species” (Schneider, 1972).4 Thus, whereas
96
Cross-Cultural Research / February 2005
such results are important for understanding generalization principles that are easily learned across species, they do not provide
proof that similarity in color chip sorting across species is a
consequence of shared color vision mechanisms.
Obviously, the present discussion of alternative nurture-like
hypotheses reveals that I oppose the view that the strongest argument for the primacy of the Hering elementary colors in color naming is a neural processing argument, and prefer instead a more
plausible higher-order cognitive argument. Instead of considering
that the “names for the Hering elementary colors are necessary
and sufficient for naming all of the colors, a fact that justifies singling them out” (Hardin, 2005, this issue), I believe the emphasis
should be placed on seeking empirical explanations as to why the
color opponents black/white, red/green, and yellow/blue seem to
represent phenomenal opposites and can serve as useful markers
in cognitive color space across a range of individuals who (a) vary
considerably regarding the physical stimuli that evoke unitary
hue experiences (e.g., Kuehni, 2004; Webster et al., 2002) and who
(b) often vary considerably regarding individuals’ point locations
for category best examples (MacLaury, 1997a). In sum, this is not
to say that Hardin’s arguments for nativistic explanations are
incorrect per se, rather it is to suggest that in light of plausible nurture-like explanations it seems too early for him to declare that the
biologically based explanations he describes are the only major
determinants of the cross-cultural color naming behaviors he
reviews.
PAUL KAY’S “COLOR CATEGORIES
ARE NOT ARBITRARY”
Kay’s article discusses two principle issues: First, he examines data originally published as a challenge to the universalevolutionary (UE) model, namely, the Berinmo data of Roberson,
Davies, and Davidoff (2000) and the Yélî Dnye data of Levinson
(2000). Second, he presents statistical analyses testing the plausibility that color naming systems tend to be based on primary basic
color categories (black, white, and the Hering opponent colors red/
green and yellow/blue).
Kay suggests that the results support two conclusions: (a) the
reexamined data do not challenge but support semantic universals
and evolutionary regularity (the UE model) through a correspon-
Jameson / REMARKS
97
dence with elemental hues (MacLaury, 1997b), and (b) Berinmo
and Yélî Dnye color naming show a statistical tendency toward a
basic primary colors bias.
With regard to the first claim, would Kay’s computed centroids
from Berinmo and Yélî Dnye data continue to closely correspond to
elemental hues points if the stimulus sets used permitted greater
freedom of indexing between color appearances and color names?
It is known that sampled empirical stimulus sets can produce patterns of naming uniformity that diminish if the available choice
options approximate the variety found in natural scenes (cf. Braun
& Julesz, 1998). The centroids Kay computes are not exempt from
influences of empirical stimulus properties and settings on informants’ response tendencies. Influences of this kind can lead participants to uniformly respond using an improper exemplar because
it is the best available among the choices offered. Various results in
the literature suggest that Kay would not likely observe the same
correspondence between his computed centroids and elemental
hues under different empirical settings or stimulus sets.
Second, Kay’s assumption of equal probability computationally
benefits his alternative hypothesis. He assumes that “each of the
320 cells has an equal probability of receiving a hue centroid.” This
benefits the computations needed to reject his null hypothesis and
accept his alternative hypothesis that the observed data closely
fit the predicted model. This uniform probability assumption
amounts to a zero-constraint scenario on the labeling of color—a
scenario that clashes with the sensible construct of “target areas”
that he defines.
Moreover, Kay’s highly improbable null hypothesis stating that
“hue naming centroids will bear no particular relation to the red,
yellow and green elementals” is impossible not to reject and in
essence requires that for categories roughly glossed red, yellow,
and green, the Yélî Dnye share no features with the many languages examined by MacLaury when identifying his universal elemental hue samples (MacLaury, 1997b). This seems to me a little
too much like a straw man hypothesis set-up to be easily rejected.
To summarize, both Hardin and Kay are rather adamant about
the fact that significant constraints on color naming exist. Kay
closes his article with the comment, “Neither the Berinmo nor the
Yélî Dnye data . . . weaken the hypothesis that there exists universal constraints on cross-language color naming; indeed, they
strengthen it,” and Hardin opens his article with the comment, “To
assert that no significant properties of color categories are
98
Cross-Cultural Research / February 2005
biologically based is to disregard what perceptual psychologists
know about color vision mechanisms.” The directness of these comments supporting universal constraints on color naming and categorization is understandable given the prolonged debate between
a group of scholars who espouse very culturally relative, antiempirical views (cf. Saunders & van Brakel, 1997) and those who
prefer to rely on what can be learned from the sometimes imperfect empirical data (authors in this journal issue and the next
issue). Research included in this two issue series represents the
empirically-oriented group mentioned and consists of a continuum
of views from more “nativist than relativist” (e.g., Kay, Hardin, and
MacLaury) to “more relativist than nativist” (Paramei, Roberson,
Alvarado, Jameson, and Dedrick). I believe that the heart of the difference in views expressed by the contributors of these issues, then,
is not so much whether human visual processing constraints exist
but rather to what degree those constraints determine color categorization and naming behaviors within a culture and across diverse
cultures. As seen in this issue’s articles, opinions diverge, with Kay
and Hardin vigorously advocating universal categories and focal
examples related to Hering opponent colors theory and Paramei
and Roberson arguing that certain social and linguistic factors varying across cultures contribute to color categorization and naming in
ways that supplant what others attribute to universally prevalent
opponent color appearance phenomenal channels. In my view, the
crux of the debate rests in specifying what forms of proof are amenable and appropriate for addressing the very different forms of phenomena—cultural, psychological, and neuropsychological—that
contribute at one time or another to the development and maintenance of color naming and categorization systems across cultures.
In contrast to the widely accepted establishment views presented by Hardin and Kay in this issue, Galina Paramei and Debi
Roberson’s separate articles describe empirical results that illustrate serious limitations of these views.
GALINA PARAMEI’S
“SINGING THE RUSSIAN BLUES: AN ARGUMENT
FOR CULTURALLY BASIC COLOR TERMS”
Paramei’s analysis of the Russian “blue” categories makes a
number of important observations regarding the dependence of
results on empirical methods and data analysis techniques
Jameson / REMARKS
99
employed. For example, her empirical review for the Russian color
terms Sinij and Goluboj illustrates the challenges inherent in generalizing “foci” findings across color order systems (e.g., the
Munsell, the Natural Color System, or the Color-Aid System) and
demonstrates that the stimulus set used greatly determines if a
color term tested is found to empirically satisfy “basicness” criteria. (For example, the 330-chip Munsell stimulus grid does not permit proper assessment of saturation and lightness variations that
play a large part in differentiating appearances named Sinij and
Goluboj—a stimulus-set attribute consistent with MacLaury’s
[1997a] findings for Goluboj). Paramei’s observations regarding
the biasing potential of stimuli are consistent with the known
stimulus-set effects in other perception and choice behaviors (cf.
Braun & Julesz, 1998, Kahneman & Tversky, 1984).
Touching on another methodological problem, Paramei implicitly cautions against making inferences about individual cognitive
and linguistic representation of color when aggregate data are
used—a point frequently illustrated with the data of MacLaury’s
(1997a) Mesoamerican color survey. This methodological consideration presents a challenge to the general practice of mapping discrete color category foci using aggregate data.
There is a natural question of whether color category boundaries are an issue that should garner as much attention as have
color category foci. In the Paramei article, this takes the form of a
discussion on the proper definition of color term “basicness,” and it
is implicit in her observation that individual differences in exemplar naming provide a starting point for examining category
boundary differences. Indeed, the question implicit in Paramei’s
analysis of the two Russian blue terms is, “What dictates that
there will always be not more than 11 basic color categories and
terms in a given culture’s color lexicon?”
In view of her in-depth analysis of Sinij and Goluboj, Paramei
raises the epistemological issue of whether a proper definition of
basicness exists. There are languages for which all the criteria that
are typically used to identify a basic term are satisfied by both the
term considered an alternate category label and that same category’s basic term. For example, such criteria apply to some languages in the Baltic and Caucasus regions when describing bluecategory stimuli. However, despite the fact that in such languages
two blue terms may satisfy the crucial basicness criteria proscribed by theory, there is tendency among some of them to limit
color term basicness to a single gloss for each category. Such a
100
Cross-Cultural Research / February 2005
practice seems to diminish the validity of the criteria used to define
basicness and further underscores the question that Paramei
raises of what actually are the proper grounds for defining basicness and nonbasicness of color terms across languages.
For improving color naming theory, Paramei suggests the use of
pragmatic psychometric measures and social and cultural constraints as important components for a theory of color naming and
categorization. She also generalizes the Interpoint Distance Model
(Jameson, in press-a; Jameson & D’Andrade, 1997) to argue for a
nonlinguistic cognitive basis for the emergence of Goluboj from the
natural partitioning of a large segment of perceptual color space
that is “unnamed,” according to basic color theory (see discussion
in Jameson, in press–b). She also argues for a greater recognition
that color naming systems are dynamic linguistic entities, so that a
given language’s color lexicon should be expected to vary over time
as locally salient cultural referents, linguistic drift, color term borrowing, and technological influences present pressure to change
the language that indexes color.
In the end, Paramei presents a convincing argument that Russian’s Sinij and Goluboj represent a challenging “special case” to
many of the accepted tenets still robust in the recent updates of the
Berlin and Kay Basic Color Term theory. As such, Paramei’s article
builds a strong argument for exploring new explanatory perspectives for what is typically viewed as a solved issue in cognition and
culture research, and she presents several leads as to what those
alternatives might be.
DEBI ROBERSON’S “COLOR CATEGORIES
ARE CULTURALLY DIVERSE IN COGNITION
AS WELL AS IN LANGUAGE”
Paramei’s analysis of two Russian blues asks, “Why does accepted theory dictate a limit of 11 universal basic color categories and
terms?” Roberson’s article asks, “What aspects of color category
universals are truly independent of language?” Roberson raises
this question through a review of empirical data showing that color
categories are greatly determined by a culture’s color language.
If, as Roberson suggests, cognitive color categories are highly
dependent on one’s learned language, and if color lexicons vary
c o n s i d e ra b l y a c r o s s c u l t u r e s, t h e n t h e b a s i s f o r t h e
Jameson / REMARKS
101
establishment’s panhuman universal color category explanation is
greatly undermined.
Bear in mind that Roberson’s position is not as neo-Whorfian as
it first sounds. Rather, Roberson and her colleagues believe that
learned linguistic categories facilitate recognition of categorical
partitions and correct perceptual classification of spectral continua, as opposed to strictly determining partitions of spectra. As
such, Roberson’s article represents a new version of linguistic relativism that contrasts with Kay and colleagues’ widely accepted
universalist perspective.
Specifically, Roberson examines the issue of whether speakers
of a language can have more and different cognitive color categories than are represented linguistically. She reviews results from
the larger field of categorization—some developmental, some
cross-cultural—that support a tight link between language and
thought. Through her review, she concludes that to the degree that
color classification and categorization resembles other classification tasks, it is very unlikely that cognitive color categories are
independent of the terms used to describe them.
She also presents convincing cross-cultural data comparing
native speakers of Berinmo and English. Her results suggest that
even in nonlinguistic sorting tasks, native speakers engage in categorical perception in a manner congruent with their native color
lexicon and that neither group can easily make consistent classifications when taught the other group’s color-language distinctions.
Her results support a tight link between linguistic labeling and
cognitive categories.
In the end, Roberson presents the explanation that in acquiring
color categories, individual learners master their culture’s specific
lexical designations for the continuously varying stimulus domain
of color. Her conclusion is that “there are no color categories that
are independent of the terms used to describe them.”
Roberson’s language-relatedness proposal is consistent with a
predicted result made earlier in the context of discussing the
Hardin article. That is, cultures that do not have a linguistic term
for the abstract concept color should have among its members a
less robustly shared set of the classical elemental hue constructs.
The rationale for this is that in the absence of an abstract concept
such as color-of-a-thing, a culture’s color labels should be largely
determined by contextualized pragmatics such as “good-to-eat
appearing,” “bad-to-eat appearing,” “dangerous appearing,” and so
102
Cross-Cultural Research / February 2005
on. In such a culture, the classical elemental hue constructs (unitary red, green, yellow, and blue) would likely be of less communicative value (and therefore less frequently used), when compared
with another culture that has a robust abstract concept for color
and a frequently used lexical term to express that concept.
Thus Roberson, like Paramei, calls for a reconsideration of the
more culturally dependent aspects of color categorization and
naming. In doing so, both reconfirm the views of others who have
suggested that the historical emphasis of perceptual processing
universals has hindered development of a full and proper explanation of color categorization across cultures (see also Jameson, in
press–b; Dedrick, in press; and Alvarado & Jameson, in press, for
similar or related views). In contrast to the Kay and Hardin articles, both Roberson and Paramei call for a more comprehensive
view of color categorization and naming—to perhaps curb the
overreliance on low-level visual processing as the explanatory
mechanism and to integrate into empirical research the aspects of
color naming and categorization that, to quote Paramei (2005 [this
issue]), find “impetus from sociocultural mechanisms.”
CONCLUSION
The four articles in this issue have had the common goal of trying to make sense of a range of empirical findings for color naming
and categorization across different ethnolinguistic cultures. As is
representative of the field, the explanations offered range from
universalistic to culturally relativisitic. A popular version of the
universalist perspective asserts that although color naming differences exist across cultures, they are largely explained by a model of
pan-human shared color experience, and this determines the
nonarbitrary basis by which all cultures categorize and name color
sensations. On the other hand, a moderate form of the relativist
perspective argues that too much of color naming and categorization behavior is shaped by sociocultural influences, and these
influences produce differences in cross-cultural color naming that
cannot be explained by the most popular universalist explanations. I do not think either opinion disputes that human visual processing imposes constraints on human color behaviors. However, I
do believe that enough questions are raised by the relativistic perspective to warrant the development of a more comprehensive
explanation of cross-cultural color naming phenomena, and I
Jameson / REMARKS
103
believe that cultural studies and cultural factors should play
greater roles in the new perspectives developed. The four articles
presented in the next installment of this two-issue series aim to
provide starting points and new approaches for developing a more
comprehensive perspective, and they set the goal as gaining an
improved explanation of the large corpus of existing color naming
and categorization data.
Notes
1. Hardin frames his support for the accepted view through an analysis
of several universal color processing results that are “primarily due to
nature” as opposed to those “mostly due to nurture” (this issue). For the
sake of comparing the two different views Hardin invokes, my comments
here frequently refer to this highly simplified nature/nurture dichotomy.
2. Jameson and Alvarado (2003) found that compared to monolingual
Vietnamese speakers tested, Vietnamese/English bilingual speakers
assessed in their native-language Vietnamese nevertheless showed patterns of color naming and categorization similar to monolingual English
speakers. The authors suggest that despite the context of Vietnamese
asessment, bilingual subjects’ Vietnamese naming behaviors clearly
reflected the influences attributable to second-language English color categorization structures.
3. One might criticize this alternative idea by asserting that estimating
the proportion of green in a sample is in some way an easier judgment than
estimating the proportion of yellow plus blue. To assert such a criticism,
comparative judgment difficulty should first be empirically examined.
Moreover, empirical judgment difficulty is not an explicit part of the rationale used by Hardin when he argues for the necessity and sufficiency of
the Hering color descriptors.
4. Schneider (1972) trained pigeons to sort 11 colors (importantly, spectral lights ranging in wavelength, not Munsell samples). His results suggest that human color categories represent a trainable, predictable subset
of pigeons’ categorical color perception, despite the tetrachromatic (or
even pentachromatic) properties of those birds’ color vision (Varela,
Palacios, & Goldsmith, 1993). Jacobs (1981) generally cautions the use of
human hue categories in cross-species research. However, as Schneider
implies, it seems likely that the color space dimensions arising from
human trichromacy are interpretable in terms of the color space dimensions of tetrachromat pigeons. Even so, pigeon (or chimp) sorting into
human color category bins does not imply color appearance equivalence
between observers from these different species.
104
Cross-Cultural Research / February 2005
5. Indeed, although the physiological interpretations he makes are not
correct from the point of view of our current understanding, in statistical
analyses Ember showed that cultural complexity combined with distance
from the equator are together a better explanation for color lexicon complexity than either factor alone. This implies that environment and culture are also contributing factors which the accepted view theory needs to
account for in explaining color lexicons across cultures (Ember, 1978).
References
Alvarado, N., & Jameson, K. A. (in press). Confidence judgments on color
category best exemplars. Cross-Cultural Research: The Journal of
Comparative Social Science, 39(2).
Berlin, B., & Kay, P. (1969). Basic color terms: Their universality and evolution. Berkeley: University of California Press.
Braun, J., & Julesz, B. (1998). Withdrawing attention at little or no cost:
Detection and discrimination tasks. Perception and Psychophysics,
60(1), 1-23.
Dedrick, D. (in press). Explanation and color naming research. Cross-Cultural Research: The Journal of Comparative Social Science, 39(2).
Ember, M. (1978). Size of color lexicon: Interaction of cultural and biological factors. American Anthropologist, 80, 364-367.
Feldman, J. (2003). The simplicity principle in human concept learning.
Current Directions in Psychological Science, 12, 227-232.
Gegenfurtner, K. R., & Kiper, D. C. (2003). Color vision. Annual Review of
Neuroscience, 26, 181-206.
Izmailov, C. A., & Sokolov, E. N. (1992). A semantic space of color names.
Psychological Science, 3, 105-110.
Jacobs, G. H. (1981). Comparative color vision. San Diego, CA: Academic
Press.
Jameson, K., & D’Andrade, R. G. (1997). It’s not really red, green, yellow,
blue: An inquiry into perceptual color space. In C. L. Hardin & L. Maffi
(Eds.), Color categories in thought and language (pp. 295-319). Cambridge, UK: Cambridge University Press.
Jameson, K. A. (1997). What Saunders and van Brakel chose to ignore in
color and cognition research. Behavioral and Brain Sciences, 20(2),
195-196.
Jameson, K. A. (in press–a). Culture and cognition: What is universal
about the representation of color experience? Journal of Cognition &
Culture.
Jameson, K. A. (in press–b). Why GRUE? An interpoint-distance model
analysis of composite categories. Cross-Cultural Research: The Journal
of Comparative Social Science, 39(2).
Jameson, K. A., & Alvarado, N. (2003). Differences in color naming and
color salience in Vietnamese and English. Color Research & Application, 28, 113-138.
Jameson / REMARKS
105
Kahneman, D., & Tversky, A. (1984). Choices, values and frames. American Psychologist, 39(4), 341-350.
Kay, P., & Maffi, L. (1999). Color appearance and the emergence and evolution of color lexicons. American Anthropologist, 101, 743-760.
Kuehni, R. (2004). Variability in unique hue selection: A surprising phenomenon. Color Research and Application, 29, 158-162.
Kuschel, R., & Monberg, T. (1974). “We don’t talk much about colour here”:
A study of colour semantics on Bellona Island. Man, 9, 213-242.
Levinson, S. (2000). Yélî Dnye and the theory of basic color terms. Journal
of Linguistic Anthropology, 10(1), 3-55.
MacLaury, R. E. (1997a). Color and cognition in Mesoamerican languages:
Constructing categories as vantages. Austin: University of Texas Press.
MacLaury, R. E. (1997b). Ethnographic evidence of unique hues and elemental colors. Behavioral and Brain Sciences, 20(2), 202-203.
Marmor, G. S. (1978). Age at the onset of blindness and the development of
the semantics of color names. Journal of Experimental Child Psychology, 25, 267-278.
Matsuzawa, T. (1985). Colour naming and classification in a chimpanzee
(Pan troglodytes). Journal of Human Evolution, 14, 283-291.
Paramei, G. (2005). Singing the Russian blues: An argument for culturally
basic color terms. Cross-Cultural Research: The Journal of Comparative Social Science, 39(1), 10-38.
Roberson, D. (2005). Color categories are culturally diverse in cognition as
well as in language. Cross-Cultural Research: The Journal of Comparative Social Science, 39(1), 56-71.
Roberson, D., Davies, I., & Davidoff, J. (2000). Colour categories are not
universal: Replications and new evidence from a stone age culture.
Journal of Experimental Psychology: General, 129, 360-398.
Saito, M. (1996a). Comparative studies on color preference in Japan and
other Asian regions, with special emphasis on the preference for white.
Color Research and Application, 21, 35-49.
Saito, M. (1996b). A comparative study of color preferences in Japan,
China and Indonesia with emphasis on the preference for white. Perceptual and Motor Skills, 83, 115-128.
Saunders, B. A. C., & van Brakel, J. (1997). Are there nontrivial constraints
on colour categorization? Behavioral and Brain Sciences, 20, 167-228.
Schneider, B. (1972). Multidimensional scaling of colour differences in the
pigeon. Perception & Psychophysics, 12, 373-378.
Shepard, R. N., & Cooper, L. A. (1992). Representation of colors in the blind,
color-blind and normally sighted. Psychological Science, 3, 97-110.
Varela, F. J., Palacios, A. G., & Goldsmith, T. H. (1993). Color vision of birds.
In H. P. Zeigler & H.-J. Bischof (Eds.), Vision, brain, and behavior in
birds (pp. 77-98). Cambridge, MA: MIT Press.
Webster, M. A., Webster, S. M., Bharadwaj, S., Verma, R., Jaikurnar, J.,
Madan, G., et al. (2002). Variations in normal color vision: Part III.
Unique hues in Indian and United States observers. Journal of the Optical Society of America A, 19, 1951-1962.
106
Cross-Cultural Research / February 2005
Kimberly A. Jameson has a Ph.D. in psychology from the University of California, Irvine. Her research seeks to understand the ways individuals and
groups conceptualize naturally occurring categories in the real world. This
includes empirical studies of color categorization and naming, the modeling of concept formation for perceptual stimuli (e.g., individual cognitive
organization of color sensations and its relation to linguistic classifiers),
evaluation of performance in tasks involving color-coded information, and
identification of cross-cultural cognitive universals of color appearance.