Multilingual/Bilingual Color Naming/Categories
A bilingual person is a person fluent in more than one language, either from birth or acquired later in life. A multilingual person is fluent in more than two languages. Color naming is the process of assigning color terms to refer to color appearances in the world. Color categories are mental representations of the regions of color space that are designated by particular color terms or names.
The influence of culture on color perception and categorization is most often studied by comparing the use of single-word (monolexemic) color terms to name color categories across different languages, as in the World Color Survey [1, 2]. Color categorization and naming have been studied across a wide range of cultures to explore questions about the universality of basic color categories, relativism of color naming and/or perception, and the ways in which color perception contribute to color naming. However, such comparisons between groups of individuals speaking different languages in different cultures necessarily confound differences in individual experience and individual perception with the linguistic properties of the languages spoken and the categorization practices of the cultures themselves, resulting in a difficulty identifying the distinct contribution of each source of influence on color-naming behavior. Comparison of the categorization and naming behavior of multilingual or bilingual individuals using each of their distinct languages provides a way of controlling for individual differences so that the influence of culture and linguistics on naming can be more readily observed.
Comparisons of multilingual/bilingual individuals performing identical color-naming or observation tasks separately in each of their languages depend on the level of fluency of the individual in each language. In some cases, a person may know two or more languages because he or she is an immigrant to a new culture, in the process of acquiring a new language while using the native language or previously acquired languages less frequently. In other cases, a person is a participant in a multicultural society in which different languages are spoken in different contexts and for different purposes but are generally spoken frequently enough to be refreshed in memory. The sequence in which languages were acquired and the relative proficiency in each language are important and should be measured in studies of multilingual or bilingual color naming. An individual in the process of assimilation to a new culture and language may experience interference from previous languages. He or she may be motivated to forget or experience pressure to forget previously acquired languages, affecting naming behavior. Studies comparing bilingual naming to monolingual naming show increased consensus of response among bilinguals, as well as a preference for use of more basic terms, perhaps reflecting a focus on what is shared across languages or perhaps a vocabulary loss affecting the less frequently used, more differentiated terms needed to describe subtleties of color samples [3, 4]. Alvarado and Jameson  observed the same phenomenon for emotion terms. For this reason, bilingual color categorization and naming behavior should be compared to both monolingual immigrant and nonimmigrant populations. Few studies have made such comparisons, even in the literature on bilingualism.
Code-switching occurs when bilingual or multilingual individuals spontaneously change languages while speaking. The switching suggests that some concept might be better expressed using a particular language or that access to vocabulary is easier in that language. Study of when and how this occurs is an emerging field in linguistics . In comparisons of bilingual color naming, subjects are required to switch languages to perform the same task twice but are constrained to respond in one language. It is unclear how much code-switching may be occurring mentally as subjects perform tasks. Further, the requirement to respond in a single language will have an impact on speeded tasks that must be considered in interpretation of results. In studies comparing multilingual/bilingual subjects across languages, the focus is generally not on the factors affecting voluntary choice of terms but on how a subject’s perceptual choice behavior or categorization and naming of a color sample are affected by use of a specific language. Researchers have asked whether perceptual boundaries match the categories existing within a language and whether perceptual experience changes to fit a second language’s categories when a new language is acquired. As yet, no studies have investigated whether an individual might prefer to name colors using the language with the richer color vocabulary, the more differentiated category structure, or perhaps the terms most relevant to performing a particular task.
Languages vary in how many color terms they contain and in where they draw color category boundaries . Languages with fewer basic terms using Berlin and Kay’s evolutionary-stage classification system may have broader categories with different focal hues. For example, it is common for many Southeast Asian and Asian languages to combine the green and blue categories which are distinct in English into a single category (commonly referred to as a “Grue” category). The Tarahumara people of northern Mexico use far fewer basic terms and form broader categories using a lightness-based naming system. This does not imply that individuals using those languages will be unable to recognize differences between green and blue color samples nor that the Tarahumara cannot perceive hues. Asian languages with a single term for the broad green/blue range of color samples use a basic term (such as Xanh for Grue in Vietnamese) combined with modifiers to specify the distinct hue (thus Xanh lá Cây for Leaf green in Vietnamese).
In general, when bilingual individuals perform color-naming tasks in two languages, their behavior suggests a dissociation between cognitive processing based on meaning (e.g., perceptual experience) and language choices dictated by grammar or language structure. Thus, a bilingual person who is assimilating to a new culture can make meaning-based choices based on one culture while using language-related naming behaviors typical of the other culture. For example, Pavlenko and Driagina  found that English-speaking bilinguals followed a Russian pattern of intransitive verb use when speaking in Russian but switched to an adjectival pattern in English. In contrast, Jameson and Alvarado  found that Vietnamese bilinguals tended to impose both the grammar and category structure of English onto Vietnamese responses when responding in Vietnamese. Thus, bilingual subjects responding in Vietnamese tended to name the category orange (a distinct category in English but not Vietnamese) using the object gloss (the fruit name) in both English and Vietnamese, whereas those monolingual in Vietnamese (in both the United States and Vietnam) named the same color sample using a modified basic term (e.g., Vàng Dam or dark yellow, with the adjective following the noun). On the other hand, native Vietnamese speakers responding in English were more likely to use Vietnamese language patterns such as more modified basic terms and compound terms (e.g., yellow-brown) than native speakers of English, who were more likely to use monolexemes (tan for yellowish-brown colors) and object glosses (Bark, Baby Puke). Thus bilinguals retained the language structure but not the meaning-based categorization patterns [4, 5]. Alvarado and Jameson  found a similar dissociation for emotion terms. It is possible this behavior may be specific to the domain of color terms, since Pavlenko  has identified several studies in which a domain of terms (in her case, emotion terms) has been treated differently than other categories of concrete and abstract nouns in a language. This suggests the need for greater caution in use of language-based properties such as concreteness/abstraction as a definition for basic color status and consideration of how the content of a domain might interact with language structure. Within-individual comparisons of naming behavior can help identify such differences in language structure as applied to a specific domain of terms, such as color names.
Color surveys that have emphasized monolexemic naming in a search for single-word basic color terms have disadvantaged those languages employing other strategies for making distinctions between colors, including use of modifiers, suffixes and prefixes to stem terms, intensifiers such as word repetition, and use of multiple-character combinations to name basic colors in Asian languages with nonphonetic writing systems. These complications of grammar with meaning complicate cross-cultural comparisons used to form theories about basic color naming but can be clarified by examining behavior of bilingual subjects because meaning-based choices are affected differently than grammar-based choices.
As noted previously , MacLaury  suggests that Berlin and Kay’s evolutionary stages (in which cultures are ranked according to the number and order of emergence of basic color terms) reflect a transition from a lightness-based naming system to a hue-based system. He also suggests that cultures evolve in their naming due to a shift in emphasis from describing similarity to describing difference. Schirillo  suggests that evolutionary stages may also reflect a transition from use of contextualized names to use of more abstract terms (basic terms are abstract by definition). Van Brakel  suggests that stages reflect the influence of Western culture on indigenous naming behaviors and are thus a transition to the Western color system from a variety of viable alternatives. However, in studies of unconstrained naming, where subjects are not confined to use of basic color terms, both English speakers and those speaking other languages greatly prefer to describe colors using secondary names or modified basic terms, not basic terms alone. If cultures were evolving toward a greater emphasis on describing difference, those cultures with more basic terms should also have more modifiers. The opposite seems to be true – use of modifiers and use of contextualized basic terms seem to vary inversely, suggesting that these are alternative strategies for describing differences in color percepts.
The behavior of multilingual/bilingual speakers may recapitulate this evolutionary process on the individual level to the extent that immigrants speaking multiple languages are in transition to English from one or more other languages. It can be argued that generalizing from individual language learning to evolution of language stages at the cultural level is justified because (a) the culture’s language practices are determined by the behavior of individuals speaking that language; (b) as with individuals, cultures change in part due to contact with other cultures; and (c) the culture’s evolutionary stage is measured by assessing the language behaviors of participants sampled from that culture. Otherwise, the processes by which cultures change are not hypothesized to be the same as those by which individuals acquire language abilities. In support of this idea, Jameson and Alvarado  found that Vietnamese bilinguals used a basic term for orange while Vietnamese monolinguals did not and that the focal samples for bilinguals responding in Vietnamese had shifted toward those of monolingual English speakers. However, for these Vietnamese bilinguals, highly specific Vietnamese object glosses and modifiers tended to fall into disuse before basic color terms. Thus bilingual speakers were unable to fully emulate the pattern of naming used by monolingual English speakers because they did not have access to a large vocabulary of object glosses in Vietnamese (probably due to disuse). A comparison of multicultural multilingual naming with immigrant multilingual naming is needed to show whether those with richer language experience show an increased tendency to use their wider vocabularies to describe additional dimensions of perceptual experience.
Alvarado and Jameson  suggest that cultures may not evolve from lightness to hue but rather from a single-dimensional naming system (based on lightness) to a multidimensional naming system capturing distinctions related to hue and saturation in addition to lightness. If so, the interaction between dimensions may produce more sharply defined category boundaries than a single dimension alone. In that case, modifier use should increase as cultures attend to additional dimensions. Further, modifiers tend to evolve along with basic terms. As noted, Tarahumara terms like Very and Somewhat evolve to Light and Bright, terms which are specific to color naming. Vietnamese uses the term Fresh (Tươi) to describe saturation, a modifier not used in English but equivalent to terms such as Vivid or Deep. Areas of color space described by lightness modifiers do not vary as much as those described by modifiers capturing other dimensions. This suggests that an evolutionary-stage theory based on a broader spectrum of naming behavior, capturing more aspects of spontaneous naming behavior, might produce a different ordering of stages.
Comparisons of naming behavior of multilingual individuals across languages provide an approach to testing assumptions of current theories of color naming, such as the assumption that monolexemic naming using basic color terms is a valid approach to comparing languages with widely varying strategies for describing color appearances. For example, as described by Burgess et al. , the Tarahumara language used a postposed bound modifier (a modifier added to the end of a stem term) that specifies the relation of the currently named color to the center of the category. This system is different from Chinese, Vietnamese, and English , where modifiers are separate words and relational distinctions are subordinated to other distinctions. In contrast, Chinese characters are frequently constructed by compounding several characters into a single character with a new meaning, and all color terms consist of two-character pairs in which the first character specifies the color and the second character specifies that it is a noun referring to the color appearance itself (decontextualized, abstract). When the color name is used as an adjective, the second character is the name of the object taking that color. In Japanese, a different second character is used to differentiate chromatic and achromatic color names . This complicates comparison of the basicness of terms when the focus is on monolexemic naming and on the properties of the language instead of the behavior of the individuals speaking that language. Emphasis on a few basic color categories, in both study and theory, neglects the richness of color-naming behavior.
It has been argued that immigrant monolingual speakers in the United States (or elsewhere) are not a suitable population for cross-cultural investigations. When monolingual speakers in immigrant communities do not acquire a new language, there appears to be a close similarity between their responses and those of monolingual speakers in their original country of residence. This may be most true for immigrants living in large insulated neighborhoods in urban areas who can conduct daily business without needing to speak an additional language. In two studies of different domains [4, 6], monolingual Vietnamese speakers in the United States and Saigon produced responses that were closely similar, suggesting that the groups responding in the United States were a close proxy for their counterparts in Vietnam. Cultural contamination cannot be automatically presumed to invalidate every investigation. Although it might also be argued that the availability of Western media may have changed the Vietnamese lexicon, resulting in the close similarity of results among the studies, differences were noted for certain contextualized terms which parallel the differences observed between the American and Vietnamese cultures themselves. For example, differences in categorization and naming of the category orange were observed for bilingual subjects but not for monolingual immigrants living in the United States, who would presumably have been exposed to Western media. Given that such cultural differences were not eliminated by exposure to the media, we suggest that perhaps everyday social interaction and communication have a greater impact on language than passive media exposure. Findings of the persistence of regional accents in Great Britain in the face of homogenous media speech support this conclusion. To assess this, the level of assimilation should be measured in studies of multilingual/bilingual versus monolingual immigrant speakers.
Study of the cognitive language processing of multilingual/bilingual individuals in color-naming tasks has the potential to widen our understanding of how language is used to describe perceptual experience by providing a glimpse into the cognitive processing that can be attributed to meaning distinct from the processing that is related to language, holding individual differences and cultural differences constant. However, to conduct such studies, comparisons must be broadened to compare bilingual and monolingual subjects in both the native country and the country to which such subjects have emigrated, as well as comparisons between multilingual individuals within multicultural societies in which multiple languages are regularly spoken independent of immigrant experience.
- 1.Kay, P., Berlin, B., Maffi, L., Merrifield, W., Cook, R.: The World Color Survey. CSLI Publications, Stanford (2009)Google Scholar
- 2.Berlin, B., Kay, P.: Basic Color Terms: Their Universality and Evolution. University of California Press, Berkeley (1969)Google Scholar
- 6.Bullock, B.E., Toribio, A.J. (eds.): The Cambridge Handbook of Linguistic Code-Switching, 1st edn. Cambridge University Press, London (2009)Google Scholar
- 8.Pavlenko, A.: Emotion and emotion-laden words in the bilingual lexicon. Biling. Lang. Cogn. 11, 147–164 (2008)Google Scholar
- 11.Van Brakel, J.: Comment on “from brightness to hue”. Curr. Anthropol. 33, 169–172 (1992)Google Scholar