Available online at www.sciencedirect.com Language Sciences 32 (2010) 196–224 www.elsevier.com/locate/langsci Language, contact, and vantages: fifteen hundred years of Japanese color terms James Stanlaw Illinois State University, Department of Sociology and Anthropology, Campus Box 4660, Normal, IL 61790-4660, USA Abstract This paper describes the history of Japanese color naming and attempts to relate it to current models of color nomenclature research, particularly through Vantage Theory. I present evidence from early texts that reveals how this vocabulary might have operated 1500 years ago, and suggest a possible evolutionary sequence of Japanese color terminology from c. 400 C.E. to the present. I show how the Berlin and Kay standard model fails to account for some of this data. But the Japanese data do not necessary contradict the standard model so much as enrich it. What we are seeing is a very complex series of cognitive and social interactions, where color categories are in a state of flux. Color terms do not spontaneously generate, but are emergent and negotiated. And this, I argue, is based on speaker’s vantages. Ó 2009 Elsevier Ltd. All rights reserved. Keywords: Color nomenclature theory; Japanese; Berlin and Kay model; Language history; Color term evolutionary sequence; Vantage Theory 1. Introduction While Japanese was among the initial 20 languages examined by Berlin and Kay and their students some 30 years ago, the history of its color-term nomenclature system has yet to be fully described in formal terms either in English or Japanese. Not all the theoretical implications of the Japanese case for color nomenclature research have been fully explored, either. In this paper, I will examine the history of Japanese color naming and attempt to relate it to current models of color nomenclature research, both the “standard” model (Berlin and Kay, 1969, 1991) and what has come to be known as Vantage Theory (MacLaury, 1995, 2002). I argue that a possible set of Japanese of “basic” color terms, indicated throughout the paper using CAPITAL LETTERS, is as follows: 1. 2. 3. shiro kuro aka E-mail address: stanlaw@ilstu.edu 0388-0001/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.langsci.2009.10.005 WHITE BLACK RED 197 J. Stanlaw / Language Sciences 32 (2010) 196–224 4. 5. 6. 7. 8. 9. 10. 11. 12. ao ki-iro midori cha-iro murasaki pinku orenji hai-iro guree kon-iro BLUE YELLOW GREEN BROWN PURPLE PINK ORANGE GREY [DARK-BLUE] While this list has some intriguing elements – such as the presence of (the underlined) English loanwords, and a hypothesized DARK-BLUE basic term – even more interesting is how this inventory developed. I will offer evidence from several early texts that suggest how the Japanese color vocabulary system might have operated a millennium years ago, and suggest a possible developmental evolutionary sequence of Japanese color terminology from around 400 C.E. to the present. I will also offer evidence for why I believe that English loanwords are indeed now basic Japanese colors. I will present, too, reasons why it is likely that there might be a twelfth basic color in Japanese.1 I demonstrate that the Berlin and Kay standard model fails to account for some critical aspects of this data. For example, Japanese historically developed into a five-term system, c. 500 C.E., from earlier categories of LIGHT/CLEAR, DARK, WARM/BRIGHT, and COOL/OPAQUE. These terms were WHITE, BLACK, RED, YELLOW, and GRUE. However, instead of the finding the next expected color terms in the Berlin and Kay sequence – such as BLUE or GREEN – murasaki (PURPLE) appears by 650 C.E., as does momo (PINK) by 750 C.E. Even the predicted BROWN term is not found for another 250 years. These are serious challenges to the standard model, which at present offers no explanation for such findings. I believe that other explanations, particularly as offered by Vantage Theory, give more robust accounts, and elucidate much of this data, succinctly and parsimoniously. Besides offering further data for formal modeling, the Japanese case also offers at least two special advantages regarding the several attempts at refutation by those questioning the existence of color term universals. First, some doubt the efficacy of using any version of the formal standard model of color nomenclature because they cannot account for several anomalous cases (such as the early Japanese color term data presented here). Indeed, some critics, such as Saunders and van Brakel (1997), think it is too convenient that the evolutionary color term sequence is as parsimonious as it is. They believe that the 11 or 12 alleged basic color terms are either ad hoc at best or presumptions at worst, and represent little more than how Westerners, or those living in industrial societies, might experience the world. Japan, however, intentionally isolated itself until the mid-19th century, leaving a color vocabulary that developed with little outside influence – a rich source indeed. Others, such as Dedrick (1998), argue that the Berlin and Kay model is cognitively sterile. Secondly, formalists such as Cohen (2001) argue that ethnographic or historical undertakings are flawed because color sensations, being totally subjective, will never allow us to determine whether the color sensation of one 1 English discussions of Japanese color terms can be found in Haarman (1989, pp. 190–218), Hinds (1974), Kobayashi (1998), McNeill (1972), Stanlaw (1987, 1997a, 2004, 2007), Tanaka and Koike (1982), and Uchikawa and Boynton (1987). Japanese discussions on color history, culture, psychology, technology, gender, and aesthetics include Nagumo (2008), Ōmi (2008), Yoshioka (2002, 2007), Chijiwa (1987), Corona Books (2006), Hidaka (2003), Kawamoto (1978), Kobayashi (1974, 1981, 1984, 1990a,b), Matsuoka (1999), Nagata (2002), Ōoka (1979), Ōoka et al. (1980), Ōyama (1994), Ozaki (2003), Shibugawa and Takahashi (1983a), ujo (1984) is a smaller, Japanese, version of the Munsell Takimoto and Fujisawa (1977), and Yamanouchi (2001). Nihon Shikisai Kenky Book of Color. Approximations of the colors cited in the text can be viewed at HeartLand-Icho (2003) and Wa-shoku (2003). BG Japan (2000) is a film on the dying process of traditional Japanese colors. The Hepburn system of Romanization (Habein, 1984) has been used, and citations come from the following standard dictionaries: Arakawa (1977), Masuda (1974), and Shinmura (1998). The most detailed discussions on color nomenclature theory can be found in four edited collections: MacLaury et al. (2007), Biggam and Kay (2006), Biggam and Pitchford (2006), and Hardin and Maffi (1997). 198 J. Stanlaw / Language Sciences 32 (2010) 196–224 observer is the precise equivalent of the color sensation of another. A theory of vantages finesses such claims and ameliorates much of this problem. I will argue that the Japanese data presented here do not necessary contradict the standard model and its universalist claims so much as enrich it. Nonetheless, there appear to be aspects of Japanese color nomenclature that defy formal modeling. What we are seeing is a very complex series of cognitive and social interactions taking place, where color categories, ranges, connotations, and denotations are in a constant state of flux. Color terms do not spontaneously generate, nor are their “definitions” fixed. These are emergent and negotiated, just as is the rest of the semantic system in a language/culture. And this, I argue, is largely based on speaker’s vantages. I will begin by presenting some comments on the so-called Berlin and Kay model of color nomenclature that apply to the Japanese data. Next I will present this Japanese data in three stages: the historical development of the Japanese color lexicon, the impact of English on Japanese colors, and the arguments for an additional DARK-BLUE color term. I then show how the Japanese evolutionary sequence stands in contrast to the predictions made by Berlin and Kay. Finally, I end the paper by attempting to show that apparently exceptional cases like Japanese actually offer new opportunities to see how universalist constraints interact with relativist liberties in creating color language – a world restricted by certain formal structures but nonetheless allowing for an almost unbounded social creativity. 2. The BKM standard model The model of color nomenclature which is now commonly accepted as “standard” was first proposed in 1969 by Berlin and Kay (1969, 1991). I have called this the BKM model (Stanlaw, 2007) – the M standing for any or all of several important collaborators: Robert MacLaury, Luisa Maffi, William Merrifield and Chad McDaniel – and will continue to do so here. While the BKM model has faced a number of significant methodological and philosophical challenges over the past 30 years (some of which will be described later), it is still the starting point for any work on color nomenclature research. Indeed, no one today, be they linguist, psychologist, anthropologist, or even artist, can begin a discussion on the cross-cultural comparison of color without using the vocabulary of the BKM model. The BKM model has been described at length in many places (e.g. Kay et al., 1997; Kay and Maffi, 1999; MacLaury, 1997; Hardin and Maffi, 1997) and will not be repeated in great detail here save for the features that are germane to the discussion of the Japanese data. Briefly, in its original manifestation (Berlin and Kay, 1969), the BKM model claims that humans label colors using at most eleven “basic” color terms, though not every language necessarily has the same number of basic color terms. Color terminology systems of all the world’s languages are thought to develop in a very predictable evolutionary order, as shown in Fig. 1. The seven stages of development are labeled by Roman numerals I–VII. All languages, then, have two basic color terms, white and black. If a language has three basic color terms, they are white, black, and red. The next two terms found are either (1) green followed later by yellow or (2) yellow followed later by green. If a language has six color terms, the next term found is blue. Brown appears next in stage VI. The last four terms – pink, orange, purple, and grey – can appear in any order after stage VI. This order was proposed to be universal; that is, all languages in the course of their history must acquire new terms in the order prescribed in Fig. 1, and every current language would be located at one of these seven stages. Data for these claims were gathered using the so-called Munsell array of 330 color chips displayed in order of hue and brightness (much like house paint samples found in a hardware store). Fig. 2A gives a black and white green yellow red black pink &/or blue yellow brown green orange &/or purple &/or grey I II III IV V VI Fig. 1. The classic Berlin and Kay evolutionary sequence (1969). VII J. Stanlaw / Language Sciences 32 (2010) 196–224 199 Fig. 2. Color space and its partitions. white version of this array. This illustration is actually a two-dimensional representation of three-dimensional color space, just as the common “Mercator-projection” map is a flat version of a globe of the Earth (the details of which will explained later). Generally, informants were shown one of these color charts and asked to indicate the ranges of various color terms in their language (e.g., “Please circle all the X-colors you see here”). The original model proposed in Fig. 1 had an immediate impact, and inspired several hundred further studies, all generally supportive. It soon became obvious that what was being elicited was not just the appearance of new color terms, but ways in which speakers were categorizing, naming, and dividing up the whole of color space. That is, what researchers glossed as the basic color term RED were not just the English red colors, but often 200 J. Stanlaw / Language Sciences 32 (2010) 196–224 Fig. 2 (continued) wide ranges of red-like hues (oranges, yellows, pinks, crimsons, etc.). Likewise, BLACK could include the blues, purples, browns, or any dark colors, while WHITE could include the whites, yellows, and light reds. And one color term that covered all the blues and greens, often called GRUE by early researchers, was found to be very common. These color categories that incorporated broader ranges than those usually associated with the English labels were sometimes called “macro” colors. In other words, the whole of color space was being partitioned on these charts (rather just a few chips indicating “color X”). Fig. 2B–F shows that this color space divides more and more as a language acquires new color terms. J. Stanlaw / Language Sciences 32 (2010) 196–224 201 Fig. 2 (continued) It was also noted that the desaturated non-colored GREY term seemed to be a “wild card” that could appear almost anywhere after stage III. The original model, then, was modified in the early 1970s, often as shown in Fig. 3 (Witkowski and Brown, 1977), to incorporate this “macro” aspect of the ranges of color terms. As these macro or composite categories will come up several times in the figures and the discussion, and are often labeled by Berlin and Kay and others in slightly different ways, Table 1 might be useful. The proposed theory behind the BKM model was based on supposed human physiological universals and the physics of color. The evolutionary sequence was found in the order it was because these macro composite categories broke down into six, presumably biologically based, primaries. That is, at first color space is divided simply in two: (1) the MACRO-WHITES or LIGHT-WARM colors (the whites, reds, and yellows), versus (2) 202 J. Stanlaw / Language Sciences 32 (2010) 196–224 MACROWHITE YELLOW GRUE GREEN MACRORED or PINK &/or ORANGE &/or BROWN BLUE MACROBLACK YELLOW I II PURPLE GRUE ............................................... GREY III V IV ................................................. VI VII Fig. 3. The modified Berlin and Kay evolutionary sequence (1973). Table 1 Composite category names. Berlin and Kay label “Macro” nickname Covered hues LIGHT-WARM DARK-COOL WARM COOL MACRO-WHITE MACRO-BLACK MACRO-RED, RERROW! GRUE white–red–yellow black–green–blue red–yellow green–blue the MACRO-BLACKS or DARK-COOL colors (the blacks, greens, and blues), as shown in the hypothetical example in Fig. 2B. Later, the LIGHT-WARM category would split into WHITE and WARM (the reds and the yellows) as shown in an hypothetical example in Fig. 2C, and the DARK-COOL category would split into BLACK and COOL/GRUE (the greens and the blues), as shown in Fig. 2D. WARM would split into RED and YELLOW, as in Fig. 2E, and COOL/GRUE would split into GREEN and BLUE as in Fig. 2F. By stage V, then, the six color terms would be WHITE, BLACK, RED, YELLOW, GREEN, and BLUE. The remaining colors would appear as some sort of combination or intersection of these primaries: BROWN from YELLOW with BLACK, PURPLE from RED with BLUE, orange from RED with YELLOW, and PINK from RED with WHITE. The achromatic GREY remained a wild card. Though it was apparent that the composite categories were splitting up as the evolutionary sequence expanded, it was not clear just which splits would come first. For the next two decades, the original model underwent numerous refinements and several important revisions, as new data was uncovered and incorporated (Kay and McDaniel, 1978; Kay et al., 1991a, 1997; Kay and Maffi, 1999). Fig. 4 shows the development RED and YELLOW break MACRO-RED and WHITE break W/R/Y Bk/G/Bu W R/Y Bk/G/Bu W R/Y G/Bu Bk W R Y G/Bu Bk addition of PINK, ORANGE, PURPLE addition of BROWN W R Y G Bu Bk BLACK and GRUE BREAK W R Y G Bu Bk + Br GRUE break W R Y G Bu Bk Br + Pu + Pk + Pr ------------------------------------ I II III IV grey V -------------------------------------- VI VII Fig. 4. The amended Berlin and Kay mainline evolutionary sequence, proposed c. 1975. 203 J. Stanlaw / Language Sciences 32 (2010) 196–224 of the color term evolutionary sequence depicted as composite breakup. (In the charts that follow colors are abbreviated as: W = white, Bk = black, R = red, Y = yellow, G = green, Bu = blue, Br = brown, Pu = purple, Pk = pink. Or = orange.) At stage I we find two color terms: MACRO-WHITE (a combination of white, red, and yellow) and MACRO-BLACK (a combination of black, green, and blue). MACRO-RED splits off from WHITE at stage II. GRUE splits off from BLACK at stage III. MACRO-RED divides into RED and YELLOW at stage IV. GRUE breaks into GREEN and BLUE at stage V. BROWN gets added at STAGE VI. At stage VII, PINK, ORANGE, or PURPLE can appear any time after BROWN (in no predictable order). I have called Fig. 4 the “mainline” evolutionary sequence, as it appears that about 80% of the world’s languages follow this trajectory. Almost all of the rest of the 20% follow the “secondary” line in Fig. 5. For the BLACK and GRUE break MACRO-REDand WHITE break W/R/Y Bk/G/Bu W R/Y Bk/G/Bu W R Y Bk/G/Bu addition of BROWN W R Y G/Bu Bk W R Y G Bu Bk addition of PINK, ORANGE, PURPLE W R Y G Bu Bk + Br RED and YELLOW break W R Y G Bu Bk Br + Pu GRUE break + Pk + Pr ------------------------------------ grey -------------------------------------- ----- I II III IV V VI VII Fig. 5. The amended Berlin and Kay secondary line evolutionary sequence, proposed c. 1975. MACRO-BLACK and GREEN break addition of PINK, ORANGE, PURPLE addition of BROWN MACRO-REDWHITE break W/R/Y Bk/G/Bu W R/Y Bk/G/Bu W R Y Bk/G/Bu W R Y G Bk/Bu W R Y G Bu Bk RED and YELLOW break W R Y G Bu Bk + Br BLACK and BLUE break W R Y G Bu Bk Br + Pu + Pk + Pr ------------------------------------ I II III IV grey V -------------------------------------- VI Fig. 6. Berlin and Kay additional evolutionary sequence, proposed c. 1977. VII 204 J. Stanlaw / Language Sciences 32 (2010) 196–224 MACRO-RED and WHITE break W/R/Y Bk/G/Bu YELLOW-GREEN and BLUE break W R/Y Bk/G/Bu W R Y/G/Bu Bk W R Y/G Bu Bk RED and BLUEGREEN-YELLOW break W R Y G Bu Bk addition of BROWN addition of PINK, ORANGE, PURPLE W R Y G Bu Bk + Br W R Y G Bu Bk Br + Pu YELLOW and GREEN break + Pk + Pr ------------------------------------ I II III IV grey -------------------------------------- V VI VII Fig. 7. Berlin and Kay additional evolutionary sequence, proposed c. 1997. YELLOW and GRUE break addition of PINK, ORANGE, PURPLE addition of BROWN MACRO-RED and WHITE break W/R/Y Bk/G/Bu W R/Y Bk/G/Bu W R Y/G/Bu Bk W R Y G/Bu Bk RED and BLUEGREEN-YELLOW break W R Y G Bu Bk W R Y G Bu Bk + Br BLUE and GREEN break W R Y G Bu Bk Br + Pu + Pk + Pr ------------------------------------ I II III IV grey V -------------------------------------- VI VII Fig. 8. Berlin and Kay additional evolutionary sequence, proposed c. 1999. most part, the difference between the two sequences is the appearance of the GRUE term. In Fig. 4 GRUE appears for the YELLOW term, while in Fig. 5 YELLOW appears before the GRUE term. It became clear that the parsimonious sequences presented in Fig. 1 or Fig. 3 (or their later adjusted counterpart, Figs. 4 and 5) could not account for all the data that had been found by the mid-1990s. Though not numerous, there is at least some evidence to support the evolutionary sequences found in Figs. 6–8. Basically, they posit the existence of a BLUE–GREEN–YELLOW category (Figs. 7 and 8), or a BLUE–BLACK (vs. GREEN) category (in Fig. 6). The presence of a YELLOW–GREEN–BLUE composite category is often explained in the BKM model as follows: in some, admittedly unusual, cases a language did not fully partition color space at the earliest stages. That is, there may have been terms for WHITE, BLACK, and RED, but these were not broadly inclusive; for J. Stanlaw / Language Sciences 32 (2010) 196–224 205 example, BLACK would not include all of the green, blue, or other dark hues. This would leave much of color space unnamed. If a language later applied partitioning – in essence, naming what was unnamed before – the leftover yellows, blues, and greens would now be labeled, initially, as a single category. The YELLOW– GREEN–BLUE composite category could split into YELLOW–GREEN and BLUE (Fig. 7) or YELLOW and GRUE (Fig. 8). These cases, however, are controversial and not well attested (Fig. 8, for example, is based on only one language). However, these other paths suggest that there are either (1) exceptions to the motivations in the standard BKM model about how primary colors derive from the breaking up of composite macro categories, or (2) there are other parameters operating. 3. The Japanese data 1: Ancient colors 3.1. Incipient Japanese color sense I have described the historical evolution of Japanese colors in detail elsewhere (Stanlaw, 2007). Highlights are presented here. It is difficult to know what the basic color term inventory was for ancient Japanese. Even the cognitive concept “color” may have been different from today’s notion of a referent-detached abstract notion. Environmental and plant stimuli were likely to be resources for naming incipient abstract colors. The ancient Japanese shoku-butsu no iro (the colors probably divided colors into three broad categories (Nakae, 2003): (1) dōbutsu no iro (the color of animals), and (3) shizen no iro (the colors of nature). of plants), (2) An example of the last type would be yoake ( ), the color of dawn. Thus, tactile or other senses were also involved with color naming in early Japan. 3.2. Early texts and data sources What we know about early Japanese color nomenclature comes from written sources, though these were not always in pure Japanese. The Japanese did not have their own writing system until contact with the Asian mainland began when the Chinese came to the country in the fifth and sixth centuries. Different Chinese teachers often spoke different Chinese “dialects,” which were, in fact, often mutually unintelligible languages, so Japanese people learned several different ways of pronouncing a single character. Even now most Sino-Japanese characters have at least two pronunciations or readings: one based on one of the old Chinese pronunciations (called on-yomi) and one based on a Japanese pronunciation (called kun-yomi). Fig. 9 (cf. Stanlaw, 2007, p. 297) gives these readings, plus other information that we will return to later, for seventeen Japanese colors. The earliest extant Japanese texts were written in the first half of the eighth century. I have used the three largest and most important books as a source for Japanese color data: The Kojiki, 712 C.E. (Kanda and Ōta, 1962); The Nihon-shoki or Nihongi, 720 C.E. (Sakamoto et al., 1967); and the Man’yō-shu, 759 C.E. Kojima et al., 1971, 1975; Itō, 1985/2002).2 The first two comprise the scriptures of the Shinto religion, and the last is a collection of some 4500 poems (or possibly songs), and is the first book written in vernacular Japanese. Of these three texts, the Man’yō-sh u is the most important and revealing. Most of its poems are in the classic tanka form of 31 syllables written in five phrases of 5/7/5/7/7 syllables each. The following two poems, besides being noted for their vivid color imagery, are among the most famous (Stanlaw, 2007, p. 301). Princess Nukata was forced to marry the Emperor Tenchi, but was previously married to his brother Crown Prince Temmu, with whom she is still in love. At a hunt organized by Tenchi, where she will pass by Temmu, she wrote (end numbers refer to Man’yō-sh u poem numbers): 2 English translations of parts of the Kojiki have been published by Philippi (1968), parts of the Nihon-shoki by Aston (1986/1972), and parts of the Man’yō-sh u by Levy (1981) and Nippon Gakujutsu Shinkōkai (1965). The indexes and citations in Shin-pen Kokka Taikan (1983a,b, 1984a,b) are tremendous resources for examining poetry of the period. 206 J. Stanlaw / Language Sciences 32 (2010) 196–224 Fig. 9. Sino-Japanese characters, readings, frequencies, and year of first appearance for selected color terms in modern Japanese. (1) akane sasu murasaki-no yuki shime-no yuki nomori wa mizu ya kimi ga sode furu As you, my love, stand in front of the red setting sun and wave to me in the purple field, the guard is watching us (20) Temmu replied: (2) murasaki no nioeru imo o nikuku araba hitozuma yue ni ware koimeya mo Though you are now another’s wife, could I keep from being in love with you – you as beautiful as this purple grass – if I blamed you? (21) J. Stanlaw / Language Sciences 32 (2010) 196–224 207 Note that according to the standard BKM the colors like purple ( murasaki) and terms like akane ( a kind of red) are unexpected at this stage in Japanese color development. We will return to this problem shortly. 3.3. Japanese stage 1 (4 terms): Ancient colors, c. 400 C.E. Satake (n.d.) claims that the earliest Japanese had concepts of three abstract colors – white, black, and red. It is hard to know when a fourth color developed, or what it was. Blue, purple, or yellow are the likely candidates, and Kunihiro (citing Shibata in MacLaury (1997, p. 342)) argues for a composite blue–green–yellow term. It seems likely, then, that the early texts are naming at least four color concepts, if not actual categories, even if their labels were not clear color terms. In the Kojiki, these were brightness, darkness, clearness, and opacity, which eventually come to be named as aka (MACRO-RED), kuro (MACRO-BLACK), shiro (MACRO-WHITE), and ao (BLUE or more precisely, GRUE). 3.4. Japanese stage 2 (5 terms): Shōtoku colors, c. 500 C.E. ), the key cosmological Ancient Chinese philosophers apprehended reality via “five elements” (gogyō symbols of wood, fire, earth, metal, and water, which interacted with the opposing ying and yang forces (in and yō ). This system was adopted all over East Asia and Prince Shōtoku (574–622), the first unifier of Japan, institutionalized it by 600 C.E. Each element had an associated cardinal direction, season, planet, virtue, taste, sense, and color. In Japanese, these five colors were shiro ‘white’, kuro ‘black’, aka ‘red, ao lit. ‘blue’, and ki ‘yellow’. 3.5. Japanese stage 3 (6 terms): Taika colors, c. 650 C.E. The third stage of Japanese color term evolution saw the introduction of a sixth category, purple (murasaki). The character for murasaki ( ) certainly had been extant for hundreds of years in China (at least as a secondary color term) and was brought to Japan along with the other thousands of imported characters. Purple became the important national color, synonymous with the imperial court, after the so-called Taika Reforms, which unified and institutionalized the Japanese state. Certain purple hues became reserved for the aristocracy and were forbidden for commoners. Indeed, the ancient “five elements” was expanded, with a new virtue (toku – humanity or compassion) and color (purple) added. 3.6. Japanese stage 4 (8 terms): Man’yō colors, c. 750 C.E. The fourth, and in many ways the most interesting, stage of Japanese color term evolution came around the time the Man’yō-sh u was compiled. The Japanese color nomenclature system expanded tremendously. There were apparently eight full-fledged colors, with green and pink being added to the six terms of the Taika colors (white, black, red, blue, yellow, and purple). As an example, consider this Man’yō-shu poem (No. 2970) for the term momo/tōka (PINK): (3) tōka zome no asara no koromo asaraka ni omoite imo ni awanu monokamo Seeing the cloth softly colored pink, like that of the crested ibis, I wonder if I might meet the girl. 3.7. Japanese stages 5–7: Genji colors, Meiji colors, and modern colors During the Heian Period, (794–1185), one of Japan’s most fruitful literary times, the brown (cha-iro) and grey (nezumi-iro) terms appeared. Both of these terms are found in the Tale of Genji (c. 1000), Japan’s greatest 208 J. Stanlaw / Language Sciences 32 (2010) 196–224 literary masterpiece and the world’s first novel, as well as in other Heian works (Kondo, 2002).3 Little further change took place in Japan’s color vocabulary until the Meiji Period (1868–1912), when the orange (daidai) term developed. In the late 20th century, extensive English borrowing has provided a loanword counterpart for each native Japanese color term. The native terms for pink, orange, and grey have become almost archaic. I will now discuss these newer borrowings. 4. Japanese data 2: modern colors I have argued previously (Stanlaw, 1997a,b, 2004) that the Japanese color lexicon actually consists of two sets of mutually exclusive terms, one of native origin, the other borrowed from English. I suggest that English loanwords, for the colors PINK, ORANGE, and probably GREY, are replacing native Japanese color terms in reverse order in the Berlin and Kay evolutionary sequence, as shown, say, in Fig. 1. These claims are supported by ethnographic data and observation, as well as formal color elicitation tasks with over one hundred native Japanese informants (most from 1987). Details for this replacement hypothesis have been presented elsewhere (Stanlaw, 1987, 1997a,b, 2004) and only a cursory presentation of evidence will be given here. Fig. 10, for instance, shows the rank ordering of selected color terms of 91 informants in a frequency saliency task. For example, 88 people of various ages (or 97% of respondents) listed the Japanese term shiro (WHITE) when asked to write down those color terms they thought were most common or most important in everyday life in Japan. (English loanword terms are underlined in Fig. 10.) The first thing to notice is that for the categories PINK and ORANGE, the English loanwords pinku and orenji are much more salient than the native Japanese colors momo-iro and daidai-iro. In fact, younger informants were often at a loss to find a referent for daidai-iro. Also, for the GREY terms, the English loanword gurē was equally salient with one native Japanese term for GREY, hai-iro (lit. ‘ash-colored’) and twice as salient for the second proposed Japanese GREY basic color term nezumi-iro (lit. ‘ratcolored’). The literature suggests that frequency salience is actually a very good indicator of basicness (Hays et al., 1972; Bolton, 1978; Bolton et al., 1980; and English data found in Francis and Kucera, 1982 and Carroll et al., 1971).4 Frequency data of Japanese newspapers and magazines gathered by the Japanese National Lanujo, 1964, 1971, 1972, 1973, 1974), show that frequency/ guage Research Institute (Kokuritsu Kokugo Kenky salience seems to correlate with the Berlin and Kay evolutionary sequence (Stanlaw, 1987, pp.111–116). Thus, there appears to be strong evidence to believe that the data given in Fig. 10 reflect both basic color term status in Japanese and the evolutionary sequence. This suggests, then, that the three terms pinku, orenji, and gurē may be replacing their native Japanese counterparts for all practical purposes. Indeed, English loanword color terms may be in the process of replacing a number of other native Japanese color terms, and perhaps these substitutions are occurring in reverse order of the evolutionary sequence. For example, we might predict that native Japanese murasaki (PURPLE) or cha-iro (BROWN) could be the next color terms to be replaced (by the loanwords pāpuru and buraun, respectively). Informants were asked to perform a mapping task using the most frequently found terms in Fig. 10. For most informants, English loanword color terms are not mapped synonymously with native Japanese color terms. Best-example focal colors and the ranges vary substantially. In general, the number of chips chosen as focals for native Japanese color categories is different than those chosen for English loanword colors, and most focals seem to be brighter by at least one step of brightness level than their native Japanese counterparts. When considering the ranges of the color terms, a similar phenomenon is found. That is, in general, English loanword color terms seem to be thought of as brighter than their native Japanese correspondents (Stanlaw, 1997a, pp. 252–254; Stanlaw, 2004, pp. 222–232). 3 There are now three major English translations of the Tale of Genji. Seidensticker’s (1976) version is most often cited. While saliency is a good indication of basicness, Berlin and Kay (1969) originally posed several others at the same time: Basic color terms should be monolexemic and unanalyzable (thus excluding terms like “bluish”). The meaning of a basic color term should not be derivable from the meaning of its parts (thus excluding terms like “olive-green” or “sunburst”). The range of a basic color term should not be included within the range of another term (thus excluding “khaki” or “tan” as they are a kind of “brown”). A basic term must have wide applicability and not be restricted to a single referent (thus excluding “blonde”). And loanwords are held to be suspect. 4 209 J. Stanlaw / Language Sciences 32 (2010) 196–224 Fig. 10. Rank order of selected Japanese color terms. A. Apparant Synonymity of TAN and KHAKI B. C. KHAKI Terms Centering Around the TAN Terms Centering Around the “T” “K” Chip and Disseminating Downward Chip and Disseminating Upward scale of the degree of commitment to a color category: KHAKI only TAN and KHAKI (chip is definitely a member of the color category) (chip is increasing less likely to be a member of the color category) Fig. 11. Hypothetical relationship of coextension. In terms of vantages, how might this data be explained? Ethnographic observations indicate, for example, that the two Japanese terms for PINK, i.e. momo-iro (lit. ‘the color of peach flowers’) and pinku, both refer to slightly different ideas and demonstrate significant differences in usage. While it is possible to use both terms ), ‘a pink sweater’ – momoto, say, modify a sweater – as in {momo-iro/pinku} no seetaa ( ), ‘pink lipstick.’ iro cannot be used to modify lipstick: {*momo-iro/pinku} no kuchibeni (* 210 J. Stanlaw / Language Sciences 32 (2010) 196–224 Likewise, as one female informant said to me, “Who would like to buy a nezumi-iro (lit. ‘rat-colored’) dress?”, indicating presumably why the English loanword is the preferred choice. It is likely that MacLaury’s relationship of coextension characterizes most of these terms. Coextension is a semantic relationship “that did not fit our preconceptions of synonymy, near synonymy, inclusion, or complementation” (MacLaury (1997, p. 111)), and emerged when informants were asked to continue naming colors of the Munsell array (Fig. 2A) after they had initially finished doing so. In effect, it was found that informants would often use different words to label the same color, depending on their perspective or vantage. As an example (Stanlaw, 2004, p. 259, and I do not mean to imply that this analogy necessarily applies to the English color terms system), let us assume that a native English-speaker names the same several dozen color chips presented to them as either TAN or KHAKI, with maybe KHAKI being also applied to a few more colors (as shown in Fig. 11A, an excerpt of the Munsell array showing only the brown and beige colors). It would appear, then, that the two terms either label the same category (with TAN and KHAKI varying freely) or that KHAKI covers the TAN the category. However, if we asked informants to continue to name colors, we might find that these terms are hardly synonyms; it could be that an informant attends to each term differently. For instance, KHAKI terms might center around a light yellowish chip (K) and disseminate outward from it (Fig. 11B). Likewise, the TAN colors might be focused around some darker brown color chip (T) and proceed up from there (Fig. 11C). So the name of some particular color category is contingent on the vantage taken by the informant. If the informant calls the category KHAKI, he or she is coming at it from the light or yellowish side and extending it down into the darker brownish TANs. If it is called TAN, the category focuses around some ideal tan-color and extends upwards towards the yellows. This is not a mere case of two terms simply being applied to the same referents; presumably the two experiences are perceptually and psychologically different. As an example to see how English loanword color terms are interacting with the native Japanese counterparts, let us use the PINK case cited above. These two terms are present in Japanese, of course, and the difference in usage suggests that people may be using them differently according to vantage. We know that these two terms differ on at least two criteria: brightness and literalness. These indicate that the pinku and momo-iro are related to each other via coextension. Figs. 12 and 13 show this. In the first case, we see that pinku holds the dominant vantage. It is the difference in literalness that partly accounts for the different way they are used: Ground FIXED COORDINATES Figure MOBILE COORDINATES Entailments 1 pinku Similarity similarity in hues 2 Similarity momo-iro differential usage and sociolinguistic rules 3 momo-iro Difference difference in literalness PDMSS Fig. 12. How the English loanword pinku is related to the native Japanese color term momo-iro: coextension. Dominant vantage: pinku. Ground FIXED COORDINATES Figure MOBILE COORDINATES Entailments 1 momo-iro Difference difference in brightness 2 Difference pinku differential usage and sociolinguistic rules 3 pinku Similarity similarity in hues MSPDD Fig. 13. How the English loanword pinku is related to the native Japanese color term momo-iru: coextension. Recessive vantage: momo-iro. J. Stanlaw / Language Sciences 32 (2010) 196–224 211 pinku is the more abstract of the two terms because momo-iro means peach-colored. The recessive vantage (Fig. 13) accounts for the remainder. Here we see it is the difference in brightness that also contributes to the varying usages (and referents) of these two terms. Similar kinds of coextensive arguments can be made for gurē and hai-iro and nezumi-iro, and orenji and daidai-iro. 5. Japanese data 3: Are there 12 basic color terms in Japanese? The standard BKM model claims that the evolutionary sequence has stopped at 11 terms (the ones given in the last stages of, say, Fig. 1). Indeed, there is much evidence and anecdote to suggest that 11 terms are the maximum number. In fact, there are data to indicate that the 11 points labeled in color space by the English – and presumably only these 11 points – are perceptually “privileged”. Tests (Kay et al., 1991b) from the World Color Survey (about 110 unwritten languages spoken in small-scale non-industrial societies) indicate that there are clear cross-linguistic statistical tendencies for named color categories to cluster at certain privileged points in perceptual color space [. . . and] these privileged points are similar for the unwritten languages of non-industrialized communities and the written languages of industrialized societies. [. . . T]hese privileged points tend to lie near, although not always at, those colors named red, yellow, green, blue, purple, brown, orange, pink, black, white, and gray in English (Kay and Regier, 2003, p. 9089) Nonetheless, there is nothing inherent in either the BKM model or the sequence itself that necessary imply that the number of basic terms must be limited. There have been numerous possible counter-examples offered from a number of languages: e.g. Hungarian RED/DARK-RED, piros and vörös (MacLaury et al., 1997); Russian LIGHT-BLUE/DARK-BLUE, sinij and goluboj (Paramei, 2007; Corbet and Morgan, 1988); Turkish BLUE/DARK-BLUE, mavi and lacivert (Ögen and Davies, 1998); French BROWN, brun and marron (Forbes, 1979); Greek BLUE, galazio and ble (Androulaki et al., 2001); and Japanese BLUE/DARK-BLUE, ao and kon (Stanlaw, 1992a,b, 1997b). But MacLaury (1997, pp. 419–429), in his review of the literature with John Taylor and Henrietta Mondry on a possible cognitive ceiling of basic colors, concludes that there is really no evidence that any system will comprise more than the eleven basic color categories Berlin and Kay originally posited. He argues that this constraint is probably due to the limits of human biological perception. The differences between, say, light blue and dark blue “are too minimal to be overtaken by our capacity to distinguish on the basic level” (p. 428). MacLaury points out that a relationship of coextension characterizes most of these terms. However, I will argue that the relationship between Japanese ao (BLUE) and kon (DARK-BLUE) is not one of coextension. For color researchers, kon is an interesting term because it is of substantial antiquity (Fig. 9), highly salient (with informants volunteering to label it, even if you do not ask), monolexemic, and learned early. However, there is great variability when informants are asked to find it on the Munsell array (Fig. 2). On the other hand, when saturation is controlled for, there is much more consistency. The Munsell color system displays colors by three physical dimensions, hue, brightness, and saturation, calibrated so that they presumably represent equal psychological intervals.5 The Munsell chart, however, only 5 While the experience of color is a subjective psychological phenomena, the physical dimensions of color are a function of electromagnetic radiation, the “waves of light” that reflect off an object. Most color theories attend to three properties. (1) The wavelength determines hue, with longer waves producing the reds, medium wavelengths the yellows and greens, and the shorter waves the violets and purples. When an object reflects light, then, the wavelength of the majority of the waves being reflected – the dominant wavelength – will determine its hue, or what color we will call the item. (2) However, it is never the case that all the reflected light is of the same dominant wavelength; depending on the properties of the object, it may or may not be more uniform or “pure.” We experience this as the color being more deep and vivid, or pale and faded. This proportion of all the waves being of the same wavelength is called saturation or chroma. Technically, these terms are slightly different, but I will not address that here. (3) Brightness is dependent on the amplitude or the height of the wave of light. Those waves with higher amplitudes have higher energies, and are experienced as brighter by human observers. In everyday speech, brightness is sometimes described as how close a color is to white or black, and saturation as how close a color is to grey. In art, saturation is also sometimes defined as a color’s intensity, but this term technically applies more to luminance or the energy of a light wave. As described in the text, these three dimensions are often depicted as a color “sphere,” with the vertical axis representing brightness, the horizontal axis saturation, and hue being a third dimension making a circle as the colors of the spectrum blend into each other. It should be mentioned, too, that while hue seems to be the primary way people think about colors – at least, apparently, in most Western languages – MacLaury has shown that saturation (2007) and brightness (1992) are equally important. 212 J. Stanlaw / Language Sciences 32 (2010) 196–224 Fig. 14. Locating kon on the abbreviated Munsell array. displays two dimensions, hue and brightness. The third physical dimension of color, saturation or chroma, is collapsed and is not actually presented as a variable. Saturation is how vivid or “deep” a color is. A pair of jeans that has been washed many times is not a different shade of blue, but it certainly is more faded than when first bought. Every chip shown in the array is given at maximum saturation, like the surface of a globe of the world. However, as the Munsell chart (indeed like a globe) is a two-dimensional projection of a three dimensional object, only the “surface” chips are shown. But because of the physical characteristics of light and human physiology, this “globe” of color space is hardly a perfect sphere. In other words, though each chip is given at maximum saturation, there are pronounced differences between the saturation of chips of different hues. Bright yellow chips, for instance, are naturally more highly saturated than dark purple ones. MacLaury (1997, p. 11) gives the saturation values for the chips displayed on the Munsell array. In an effort to see if saturation was a factor in the naming of kon, I constructed three new Munsell charts, each holding brightness constant by varying by hue and saturation. One chart was made for chips at brightness level 4 (i.e. row G) on the chart in Fig. 2. Another chart was made for brightness level 3 (i.e. row H). And one chart (Fig. 14) was made for brightness level 2 (i.e. row I). The thick black bar in Fig. 14 represent the “border,” or surface, of the Munsell array (the chips running along it being the ones seen in the chart). The chips appearing to the left of this are not seen in the usual array, but are the third dimension “going into” the page of Fig. 2. I displayed all three charts to three dozen native Japanese informants, and asked them to label the chips they thought were kon. These ranges were indicated by shadings, and the numbers indicate the number of times a chip was picked as a focal color or “best example” of kon. As is seen in Fig. 14, there was substantial J. Stanlaw / Language Sciences 32 (2010) 196–224 213 Fig. 15. A possible evolutionary sequence of Japanese color terms. Note: these are not the stages in the BKM model (which use roman numerals: stage I, etc.). agreement on several less saturated chips as representing kon, the range was remarkably compact.6 Thus, I claim that the relationship of kon to ao (Japanese BLUE) is not one of coextension but one of complementation. I also posit that kon is at least an incipient, if not full-fledged, twelfth color term in Japanese. 6. Explaining the evolutionary sequence of Japanese color terms 6.1. Early BKM violations Putting all the evidence together from the three sets of data presented above, we can now offer a possible developmental sequence of Japanese color terms, from around 400 C.E. to the present. As seen in Fig. 15, the Ancient colors of 400 C.E. and the Shōtoku colors of 500 C.E. follow the mainline trajectory of the BKM model (Fig. 4). The first discrepancy comes with the Taika colors, because murasaki (PURPLE) appears at an unpredicted time. In the Man’yō colors we also unexpectedly find momo (PINK); midori (GREEN) also appears and ao (BLUE/GRUE) splits. The Genji colors (c. 1000 C.E.) of medieval Japan find BROWN (cha-iro) and GREY (nezumi-iro) appearing. ORANGE (daidai) appears in the Meiji Period (1868–1911). These last stages are consistent with the standard model. Finally, in the late 20th century, we find the appearance of English loanword synonyms for every basic Japanese color. The critical issues are obviously at Japanese stage 3 (the Taika colors) and stage 4 (the Man’yō colors), and how they relate to the BKM stage IV to stage VI transitions. The explanatory mechanisms of the BKM model discussed previously (i.e. first partition color space into WHITE-WARM and BLACK-COOL, then split WHITE from WARM/MACRO-RED, then split WARM into RED and YELLOW, then split COOL/ GRUE from BLACK, then split COOL/GRUE) simply cannot account for the development found in Fig. 15, even if the secondary path (Fig. 5), or the YELLOW–GREEN–BLUE exceptions (Figs. 7 and 8) and BLACK/BLUE exceptions (Fig. 6) are considered. The first problem is that neither PURPLE nor PINK should appear before BROWN. More important is why the PURPLE murasaki term should not appear here. Fig. 16 shows the composite and derived colors in the main line of the BKM model. PURPLE theoretically develops out of a mixing of RED and BLUE 6 No chips were chosen by informants for kon in the first two charts I made for brightness level 4 (row G) or brightness level 3 (row H). 214 J. Stanlaw / Language Sciences 32 (2010) 196–224 Fig. 16. Evolution of composites and derived colors in the “mainline” BKM model. Fig. 17. Early evolution of Japanese composites and derived colors, c. 400–700 C.E. Note: these are not the stages in the BKM model (which use roman numerals: stage I, etc.). sometime at stage VII (on the bottom of the figure). However, the Japanese COOL/GRUE term (ao) has not yet split into its component parts at Japanese stage 3, Taika colors, seen in Fig. 17 (showing composite and derived Japanese color categories; cf. Stanlaw (2007, p. 308)). Even at the Japanese stage 4 of the Man’yō colors (Figs. 15 and 17), there seems to be incomplete splitting of the COOL category. The relationship between the ao and midori terms and the GREEN and BLUE (and even at times, BLACK) color categories are unclear (and will be discussed shortly). Some other mechanisms must be operating for these early Japanese color data. 6.2. The Man’yō colors terms, c. 759 C.E. As Fig. 15 indicates, there are eight Man’yō colors. What does this figure tell us? J. Stanlaw / Language Sciences 32 (2010) 196–224 215 1. WHITE The WHITE (shiro or shira) term seems straightforward, though BLACK (kuro) sometimes overlaps with ao (BLUE) and/or midori (GREEN). Aka (RED), while apparently basic, competes with several other frequent and salient red terms (e.g., compare akane in the Man’yō-shu poem, Example 1, given previously with u poems). kurenai in other Man’yō-sh 2. PINK PINK (momo or tōka) appears to name only a restricted very light range of pink connected to the hues of significant flowers such as cherry blossoms. 3. YELLOW The character for the YELLOW category ( ) has several spoken forms (ki momichi, or even kariyasu), and its range floats around, usually including modern BROWNs and sometimes interacting with the BLUE and GREEN terms. It is possible that the character appeared before the ki reading. 4. PURPLE The importance of the murasaki term is still noted today, but was an extremely important color in early Japan. It was the color of the nobility and apparently had common light (nise-murasaki) and dark (honmuraski) forms which might have almost been basic color categories themselves. 5. BLUE The ao (BLUE) term is apparently the GRUE term, referring to blue and green, and even sometimes black and grey colors. Its Sino-Japanese character ( ) was one of the five “official colors” of East Asia. However, in texts where actual vernacular Japanese was used, midori (GREEN) is co-contemporaneous with ao. For example, we find both ao-sora (lit. ‘blue sky’) and midori-sora (lit. ‘green sky’) in Japanese poems. This at least suggests that in the spoken language ao and midori both labeled the GRUE category, and might not have completely split apart until much later. 6. GRUE The Japanese GRUE category has always been an intriguing problem. It is often assumed that ao names this category. Even today the character used to write Japanese ao ( ) can refer to several different hues, such as green, blue, black, and grey, in both Japanese and Chinese (Stanlaw, 2007, p. 309). Thus, historically, it is probable that ao was equally as broad. This term is certainly fluid, at times labeling everything from black foxes, to grey horses, to blue skies, and green beans. The Manyō-shu explicitly even says that ao was a white color with a bit of grey added (hai-iro ga katta shiro-iro). Thus, it appears that ao is the exemplar GRUE term. 7. GREEN It is clear from numerous poems in the Man’yō-shu that midori named green hues. Uemura and Yamazaki (1950, p. 50) claim that the first appearance of the character for midori is in the Nihon-shoki in the section “Midori-No” (lit. ‘green fields’) in the chapter on the Keikō emperor. They go on to state that “In the Man’yō-sh u, midori ( ‘green’) is also called the ‘color of duck’s wings.’ However, in general, this color [of duck’s wings] was called ao ( ‘blue’).” Does this mean, then, that midori was contained within ao? Perhaps not. In terms of written Japanese (not Chinese) texts, midori (GREEN) and ao (BLUE) appear together. Some philologists (Nakae, 2003, pp. 64–65) claim that the origin of the midori green term can be traced back to the earlier term mizu ‘water’, and indeed can be derived from it through consonant assimilation. Also, “sometimes, the midori-iro green color ( ) was used as one of the five official color terms, and sometimes it meant ao ( blue) and even ai ( dark blue)” (Uemura and Yamazaki (1950, p. 50)). And, as we have seen, the sky could at times be both BLUE (ao-sora) and GREEN (midori-sora). These examples would suggest that while ao may have often been more broad, midori was near synonymous with it: we can find cases of green hues being called ao and blues being called midori. 6.3. Finding the early Japanese GRUE term The BKM model does not lend itself to easily interpreting a finding such as the ao and midori case here, where two labels apparently name, or almost name, a single category. The BKM model might claim that 216 J. Stanlaw / Language Sciences 32 (2010) 196–224 the names are in free variation, or perhaps the category is in the process of splitting (e.g. Berlin and Kay (1969, pp. 42–43), for the Japanese case). However, how and why this split is occurring is left unexplained. Vantage Theory provides a way: this is another example of coextension, a situation where two terms label the same category but with distinct stresses (MacLaury (1997, p. 111)). Therefore, the question “Was the ancient Japanese GRUE category labeled by ao or midori?” is not the correct way to phrase things. As we have seen, different people used different terms at different times. MacLaury (1997, p. 292) has argued that with respect to GRUE/COOL categories, it is neither green-hues nor blue-hues by themselves that shape the category, so much as the manner in which their relations are constructed or perceived. In the case of Japanese, we first notice that the blue hues tend to be brighter in Munsell values than the green hues. Thus, at least some of the time, the ao term might be naming brightness rather than hue vis-à-vis midori. There are also some special connotations attached to ao. First, as mentioned, it is one of the five “official” colors of the pan-East Asian culture area, and some of its historical associations still carry to this day. Second, ao carries with it feelings of youth, inexperience, change, and freshness. What the Japanese seem to be encoding in their use of ao is the idea of “starting,” or “beginning-ness.” As several informants told me, when they are speaking of things already grown, they use midori, but “when we want to indicate things that are in the process of growth, we use ao.” Indeed, it is likely that the Japanese people were invoking this notion of “starting” or “freshness” when they decided to use ao as the name for the green-light in their traffic signal (Stanlaw, 1997a). A “definition,” then, of ao would include all these things, besides just its hue. We can make an analogy to “red” in American English. A fire engine, for example, is still called red even if it is actually more brownish or orangeish or yellowish rather than prototypical “fire engine red.” The flashing lights that a policeman uses to pull a car over for a traffic stop are now blue, or blue and red, in most communities in the United States. Still, any American knows the feeling when someone says “I saw those flashing red lights in the rearview mirror and my heart skipped a beat.” Context, situation, and the nature of the categories themselves give individuals great variety in selecting a color term. Such processes can be explained by vantages. Figs. 18–21 show the relationship of coextension that exist between ao and midori, depending on a speaker’s vantage. In Fig. 18, we see the case from the perspective of the ao being the dominant vantage (i.e. being seen as the broader GRUE term), with midori being in the recessive vantage (Fig. 19). It is probable that the attention to brightness is the underlying motivation for this. It is also possible, however, that we see the category from the “fresh” changing vantage. Here the dominant vantage of midori is shown in Fig. 20 and the recessive vantage is found in Fig. 21. 6.4. Early derived categories: PURPLE and PINK Besides being limited in explaining how composite categories split in the Japanese case (as we have seen in the previous discussions), the BKM model also has some difficulties explaining another aspect of the Japanese sequence: the early and unexpected appearance of derived categories, such as those labeled by murasaki (PURPLE) and momo-iro (PINK). Derived basic terms are believed to name colors that are perceptually most distinct from each other (PURPLE being maximally far from both RED and BLUE; PINK far from RED Ground FIXED COORDINATES Figure MOBILE COORDINATES Entailments 1 ao-GRUE Similarity similarity in hues 2 Similarity midori including midori in aoGRUE 3 midori Difference difference in brightness ASSMD Fig. 18. How ao-GRUE is related to midori: coextension. Dominant vantage: ao-GRUE. J. Stanlaw / Language Sciences 32 (2010) 196–224 Ground FIXED COORDINATES Figure MOBILE COORDINATES Entailments 1 midori Difference difference in brightness 2 Difference ao-GRUE differential usage and sociolinguistic rules 3 ao-GRUE Similarity similarity in hues 217 MDDAS Fig. 19. How ao-GRUE is related to midori: coextension. Dominant vantage: midori. Ground FIXED COORDINATES Figure MOBILE COORDINATES Entailments 1 midori-GRUE Similarity similarity in hues 2 Similarity ao including ao in midoriGRUE 3 ao Difference difference in freshness MSSAD Fig. 20. How midori-GRUE is related to ao: coextension. Dominant vantage: midori-GRUE. Ground FIXED COORDINATES Figure MOBILE COORDINATES Entailments 1 ao Difference difference in freshness 2 Difference midori differential usage and sociolinguistic rules 3 midori Similarity similarity in hues ADDMS Fig. 21. How midori is related to ao: coextension. Recessive vantage: ao. and WHITE). MacLaury (2001, pp. 1228–1236) argues that one (or a few at most) of the derived basic color terms can commonly appear at stage III or stage IV of the BKM sequence. Assuming the “maximally perceptual distant” hypothesis is true (and color mapping tasks, logic, and anecdotal evidence gives us little reason to think otherwise), the expected order would be BROWN, PURPLE, PINK, ORANGE, or GREY. In Japanese, the early appearance of the derived category murasaki (PURPLE) among the Taika colors coincides with the appearance of the composite GRUE category (ao) and precedes BROWN by many centuries. Among the Man’yō colors, momo (PINK) appears, also centuries before BROWN, at around the same time we see evidence of the ao-midori GRUE coextension split. There are probably no formal BKM mechanisms to predict or explain these “anomalies” in the ancient Japanese color term development. Instead, we must examine other factors. For example, consider this part of a Man’yō-sh u poem (Stanlaw, 2007, p. 313): (11) ... ... . . . murasaki kusa no nioeru . . . . . . your lover’s smell of the purple grass . . . 218 J. Stanlaw / Language Sciences 32 (2010) 196–224 Fig. 22. The split of the purple colors in classical and medieval Japan. We note here that neither “blue grass” nor “green grass” is used (cf. the Man’yō-shu poems given previously in Examples 1 and 2). While this is not definitive evidence for the possible cognitive primacy of Man’yō PURPLE over GRUE or BLUE or GREEN, this allusion does show the importance of murasaki. Purple, being the most virtuous of colors at the time, took on great prestige and value, as we discussed previously. As purple grew in significance, it became rather important for everyone to be able to discern just what kind of purple was acceptable for everyday wear for the aristocracy and warrior caste, and what kind was forbidden for commoners. Over the course of history of the dyeing process, then, the purple colors apparently became ‘deep purple’) and asa-murasaki ( ‘light purple’), which divided into the two groups, fuka-murasaki ( were important secondary, if not quasi-basic, color terms themselves in medieval Japan. Fig. 22 depicts the evolution of these two purple terms (Stanlaw, 2007, p. 314). Nagasaki (1983, 2001, pp. 72, 258–260) argues that one shade of fukai murasaki (‘deepish purple’, or the color often known as “royal purple” in English [5P 3/7 in Munsell color notation]) eventually became the prototypical purple color. This color came to be known as hon-murasaki (i.e., primary, original, purple). But generally these colors were simply called murasaki, though no one besides the aristocracy or Buddhist priests ‘forbidden colors’) for others.7 could could wear them: they were kin-jiki ( 7 The fact that the term kin-jiki( notions can still linger even today. ‘forbidden colors’) was the title of novelist Yukio Mishima’s 1968 influential novel shows how such J. Stanlaw / Language Sciences 32 (2010) 196–224 219 Fig. 23. The multi-layered relationship between Japanese purple colors. It should be noted here that relations between “real purple” hon-murasaki and “fake” nise-murasaki are not those of inclusion or near-synonymy. But neither are they really complementary, with no overlap or connection. At the same time, they cannot be reduced to simple coextension. Instead, the relationships are more of the order of “two-dimensional” coextension. Fig. 23 shows how this might look. The top of the figure shows the development of the “fake” ersatz purples from the light purplish hues. Presumably along the way, each pair of circles (that is, any two contiguous terms) started off in near synonymy, and drifted apart through various steps of coextension. Likewise for the “real” purples on the bottom of the chart. But while each line of circles (relations) moves along its own path, each circle is connected to its counterpart above (or below) it throughout the developmental journey. And these are likely to be coextensive as well, but not as predictable as is usually the case. Thus, we can say, if nothing else, that the “meta-chromatic” properties of the purple colors affected color choice in various indeterminate and unpredictable ways. They likely also affected how Japanese color nomenclature development for the next centuries. It is impossible to know if the terms in Fig. 22 were secondary or basic color terms, but their existence no doubt was a major factor in the growth and development of the Japanese color vocabulary. Another interesting case is PINK. As we saw, the term momo appears very early in the history of Japanese color term development. However, its range seems to have been rather restricted, referring mostly to light hues (especially to various flowers and plants). We do not know how this term interacted with RED (aka). In the Man’yō-sh u poems, there were five or ten other common red-terms that might have been rivals for basic status, and PINK might have appeared out of this contest. Regardless whether or not it is basic, it became important because of its poetic connotations and frequent use in verse. 7. Conclusion There are two fundamental – and as yet, still outstanding – questions regarding the Berlin and Kay evolutionary sequence: (1) why does the evolutionary sequence exist in the first place? (2) what are the mechanisms that cause a language to move along this sequence? The commonly accepted answer to the first question is that the evolutionary sequence is due to human physiological or neurological universals.8 The second question remains a highly debated and empirical issue. The Japanese data suggest that the standard BKM model needs to be extended in several crucial ways. 8 To be sure there have been other proposals as to why the BKM naming sequence is found. For example, Jameson and D’Andrade (1997) have argued that the evolution of color terms is strongly related to the irregularity of perceptual color space. In human perception, hue interacts with brightness and saturation to form a color “sphere”. However, as described in the text, because of the sensitivities of the human and the properties of light, this sphere is actually tilted, looking more like a hollow egg than a perfectly round hollow ball. These “lumps” in perceptual space encourage languages to develop terms that are maximally distant from those terms that already exist. Their case is convincing, and I have even made similar arguments myself (Stanlaw, 1987). Nonetheless, I would say that such proposals are additions to, or refinements of, the universalism of the standard BKM model. 220 J. Stanlaw / Language Sciences 32 (2010) 196–224 Berlin and Kay tried to explain the dynamics of the evolutionary color term sequence as due to cultural and technological complexity. They argued that in small local environments, secondary color terms are not only sufficient, but actually advantageous. If all people know “plant X,” then the secondary color term “color of plant X” carries more information than an abstract color name. When technology and group size increase, however – so the argument goes – abstract color terms are required for people who may not have the same referent in their environment. Increasing color technology such as dyeing, painting, religious decoration, and so on, would require even more emphasis on abstract color terms. These arguments no doubt contain some truth. Berlin and Kay (1969, p. 16) showed that cultures with small populations and limited technologies tend to have fewer basic color terms, industrialized societies the most. But do cultures stop evolving color terms when they reach a certain level of technological sophistication and linguistic development? There is no a priori reason to assume so. The Japanese data suggest at least three techniques which could encourage further development of the evolutionary sequence: (1) a language/culture could create new basic color categories (such as a “dark blue” or a “yellow–green”); (2) a language/culture could increase the number of terms available for basic color term status through extensive borrowing of loanwords; and (3) a language/culture could replace native terms in the evolutionary sequence with loanwords. However, that said, it must be admitted that the BKM standard model has a pronounced difficulty in accounting for significant aspects of the Japanese color term data. Indeed, it is for reasons like these that some have questioned the validity of the BKM formal standard model. While the three main sources of color nomenclature data – i.e., the World Color Survey, the Mesoamerican Color Survey, and the original Berlin and Kay data and its offshoots – all appear theoretically and empirically sound, and show similar results, there are nevertheless some who pose philosophical challenges (e.g. Davis, 2000; Dedrick, 1998; Gage, 1999; Saunders, 1992, 2000, 2007; Saunders and van Brakel, 1988, 1997, 2002).9 These critics think that it is too convenient that the evolutionary color term sequence is as parsimonious as it is, and oddly reflects the color terminology of European languages. Formalists also argue that ethnographic or historical undertakings are flawed because “color sensations have multitudes of names, and being totally subjective, we shall never determine whether the color sensation of one observer is the precise equivalent of the color sensation of another observer, even when both use the same descriptor” (Cohen, 2001, p. 6; Kelly and Judd, 1976). I believe, however, that the Japanese data enrich, rather than contradict, the standard model. What we are seeing is a multifaceted series of social and cognitive associations and exchanges, where color categories, color terms and their ranges, as well as color connotation and denotation, are in a continuous state of fluctuation and change. Color terms do not spontaneously appear or disappear and their defining parameters are not fixed. They emerge out of the complex interplay of social discourse and negotiated meanings, but mediated and constrained by formal elements such as vantages and human biological and psychological universals. Colors cannot be just any combination of hues one might imagine. But from these formal limitations, human color creativity blossoms, whether orally as in poetry or visually as in the fine arts. There is another reason we might mention for why the Japanese case is important to color nomenclature theory. Saunders and van Brakel have strenuously argued that the world’s present-day color term vocabulary is due to Western imperialism. The Japanese case, due to its color vocabulary having developed for the most part in isolation, could offer support for the claims of Saunders and van Brakel – yet it does not. Indeed, we can see in the Japanese case in the late nineteenth and twentieth centuries what happens when a language/culture not only tolerates but in fact encourages contact (Stanlaw, 1987, 2004). In ancient times in Japan, “color” was often seen as synonymous with dyeing. But because dyeing was expensive, “color” (like writing and other things borrowed from mainland Asia) was a generally a high-class concern. The lower classes and the farming majority did not have the opportunity to make new colors, and their clothes were basically drab whites or plain browns. High class men in classical and medieval Japan were peacocks, with extremely loud and bright clothes (Yoshioka, 2002, p. 111) wearing such colors as beni ( ‘crimson red’), kōbai ( ‘light pink or ‘plum blossom red’), sakura ( ‘cherry blossom red’), or haji-momiji 9 These debates have sometimes been rather tumultuous. For some reactions to these criticisms see Hardin and Maffi (2000), Kay (2000), Kay and Berlin (1997), and Stanlaw (1993). J. Stanlaw / Language Sciences 32 (2010) 196–224 221 ( lit. ‘colors of the start of fall’). Color, then, was a marker of social class rather than gender, as it often seems to be in Japan and the West today. While there are no hard and fast fashion rules in the 21st century post-modern world, many people still associate lighter brighter colors with feminine dress and conservative darker colors with masculinity (the pink for girl babies and blue for boy babies stereotype being just one example). But regardless of this, the point here is that “color” was as much a social product as a biological response to the early Japanese (as indeed, it is in every culture). The Japanese color terms presented here expand and contract, depending on the period of time under consideration. Indeed, some have disappeared. Their referents and ranges at times overlap, sometimes separate, sometime reunite, and even morph into something new or unpredicted. There are social and cultural exigencies and contingencies to every cross-linguistic universal. The Japanese case here offers a few such examples, but many more remain to be discovered. Acknowledgements I wish to thank Adam Głaz and Keith Allan, and two anonymous reviewers, for their very thoughtful reading and comments. Also, Adam and Keith need to be thanked once again for organizing the publication of this special issue. I also need to thank several other people for their help in preparing this manuscript: Masanori Yoshida of Nihon University has continued to provide me with Japanese color sources, films, and books. He also helped arranged for me to be a visiting scholar at Nihon University in the summer of 2007. Simply put, his help has been invaluable. Roger Thomas, a specialist on Japanese poetry, provided me with many rare sources, new ideas, and was a patient listener. And what can I say about the late Robert MacLaury? Those of us who knew him know how much our papers have suffered without his advice and comments. There is no academic colleague or friend that I miss more. Finally, once again without Nobuko Adachi’s knowledge and patience this paper would not have been possible. References Androulaki, A., Pestaña, N.G., Lillo, J., Davies, I.R., 2001. Greek colour terms: evidence for a natural category boundary within blue. Perception 30 ECVP Abstract Supplement. Arakawa, Sōbei, 1977. Kadokawa Gairaigo Jiten [Kadokawa loanword dictionary], 2nd ed. Kadogawa Shoten, Tokyo. Aston, William George, 1986/1972. Nihongi: Chronicles of Japan from the Earliest Times to AD 697. Charles E. Tuttle Company, Tokyo. Berlin, Brent, Kay, Paul, 1969. Basic Color Terms: Their Universality and Growth. University of California Press, Berkeley. Berlin, Brent, Kay, Paul, 1991. Basic Color Terms: Their Universality and Growth, Updated ed. University of California Press, Berkeley. BG Japan (Nikei Ēijō, producers), 2000. Koto no Juni-shoku [Twelve basic colors of the old capital]. Nikei Ēijō [film]. Biggam, Carole P., Kay, Christian J. (Eds.), 2006. Progress in Colour Studies. Language and Culture, vol. I. John Benjamins Publishing Company, Amsterdam, Philadelphia. Biggam, Carole P., Pitchford, Nicola (Eds.), 2006. Progress in Colour Studies. Psychological Aspects, Vol. II. John Benjamins Publishing Company, Amsterdam/Philadelphia. Bolton, Ralph, 1978. Black, white, and red all over: The riddle of color term salience. Ethnology 17, 287–312. Bolton, Ralph, Curtis, Ann, Thomas, Lynn, 1980. Nepali color terms: Salience on a listing task. Journal of the Steward Anthropological Society 12, 309–322. Carroll, John B., Davies, Peter, Richman, Barry, 1971. Word Frequency Book. Houghton Mifflin/American Heritage, Boston. Chijiwa, Hideaki, 1987. Color Harmony: A Guide to Creative Color Combinations. Rockport Publishers, Rockport, MA, North Light Books, Cincinnati, OH. Cohen, Jozef B., 2001. Visual Color and Color Mixture. University of Illinois Press, Urbana, IL. Corbet, Greville, Morgan, Gerry, 1988. Colour terms in Russian: reflections of typological constraints in a single language. Journal of Linguistics 24, 31–64. Corona Books Editorial Staff, 2006. Nihon no Iro [Japanese colors]. Heibun-sha, Tokyo. Davis, Steven (Ed.), 2000. Color Perception: Philosophical, Psychological, Artistic, and Computational Perspectives. Oxford University Press, New York. Dedrick, Don, 1998. Naming the Rainbow: Colour Language, Colour Science, and Culture. Kluwer Academic Publishers, Dordrecht, The Netherlands. Edward Seidensticker, 1976. The Tale of Genji by Murasaki Shikibu. Trans., Alfred A. Knopf, New York. Forbes, Isabel, 1979. The terms brun and marron in modern standard French. Journal of Linguistics 15, 295–305. Francis, W. Nelson, Kucera, Henry, 1982. American Heritage Word Frequency Book. American Heritage, New York. Gage, John, 1999. Color and Meaning: Art, Science, and Symbolism. University of California Press, Berkeley. Haarman, Harald, 1989. Symbolic Values of Foreign Language Use. Mouton de Gruyter, Berlin. 222 J. Stanlaw / Language Sciences 32 (2010) 196–224 Habein, Yaeko Sato, 1984. The History of the Japanese Written Language. University of Tokyo Press, Tokyo. Hardin, C.L., Maffi, Luisa, 2000. Color categories: the docudrama. American Anthropologist 102, 321. Hardin, C.L., Maffi, Luisa (Eds.), 1997. Color Categories in Thought and Language. Cambridge University Press, Cambridge. Hays, David G., Margolis, Enid, Narrol, Raoul, Perkins, Dale Revere, 1972. Color Term Salience. American Anthropologist 74, 1107– 1121. HeartLand-Icho, 2003. Japanese colors. <http://geocities.co.jp/Heartland-Icho/9109/irome.html> Hidaka, Kaoru, 2003. Nihon Bujutsu no Kotoba Annai [A guide to the words of Japanese art]. Shoogakukan, Tokyo. Hinds, John, 1974. Make mine BURAKU. Language Research 10 (2), 92–108. Itō, Haku, 1985/2002. Man’yō-sh u, 2 vols. Kadokawa, Tokyo. Jameson, Kimberly, D’Andrade, Roy, 1997. It’s not really red, green, yellow, blue: An inquiry into pereceptual color space. In: Hardin, C.L., Maffi, Luisa (Eds.), Color Categories in Thought and Language. Cambridge University Press, Cambridge, pp. 295-319. Kanda, Hideo, Ōta, Yoshimaro, 1962. Nihon Koten Zensh u: Kojiki [Collected Japanese classics: The Kojiki], 2 vols. Asahi Shimbun-sha, Tokyo. Kawamoto, Shigeo, 1978. Kotoba no Shikisai [Words of coloration]. Iwanami-shinsho, Tokyo. Kay, Paul, 2000. In defense of color categories in thought and language: a response to BAC. Saunders’s review. American Anthropologist 102, 321–323. Kay, Paul, Berlin, Brent, 1997. There are non-trivial constraints of color categorization. Behavioral and Brain Sciences 20, 196–202. Kay, Paul, Berlin, Brent, Merrifield, William, 1991a. Bicultural implications of systems of color naming. Journal of Linguistic Anthropology 1, 12–25. Kay, Paul, Berlin, Brent, Merrifield, William, 1991b. World Color Survey. Summer Institute of Linguistics, San Antonio. Kay, Paul, Berlin, Brent, Maffi, Luisa, Merrifield, William, 1997. Color naming across languages. In: Hardin, C.L., Maffi, Luisa (Eds.), Color Categories in Thought and Language. Cambridge University Press, Cambridge, pp. 21–56. Kay, Paul, Maffi, Luisa, 1999. Color appearance and the emergence of basic color lexicons. American Anthropologist 101, 743–760. Kay, Paul, McDaniel, Chad, 1978. The linguistic significance of the meanings of basic color terms. Language 54, 610–646. Kay, Paul, Regier, Terry, 2003. Resolving the question of color naming universals. Proceedings of the National Academy of Sciences 100 (15), 9085–9089. Kelly, Kenneth, Judd, Deane, 1976. Color: Universal Language and Dictionary of Names. US Department of Commerce, Washington, DC. Kobayashi, Shigenobu, 1974. Nihon-jin no Kokoro to Iro [The heart of Japan and color]. Kodansha International, Tokyo. Kobayashi, Shigenobu, 1984. A Book of Colors. Kodansha International, Tokyo. Kobayashi, Shigenobu, 1981. The aim and method of the color image scale. Color Research and Application 6 (2), 93–107. Kobayashi, Shigenobu, 1990a. Karā-imēji Sukēru [Color image scale]. Kodansha, Tokyo. Kobayashi, Shigenobu, 1990b. Color Image Scale. Kodansha International, Tokyo. Kobayashi, Shigenobu, 1998. Colorist: A Practical Handbook for Personal and Professional Use. Kodansha International, Tokyo. Kojiki, 1968. Trans. by Donald Philippi. Princeton University Press, Princeton, NJ. Kojima, Noriyuki, Kinoshita, Masatoshi, Satake, Akihiro, 1971–1975. Nihon Koten Bungaku Zensh u: Man’yō-sh u [The collection of Japanese literary classics: The Man’yō-sh u], vols, I–IV. Shōgakukan, Tokyo. Kokuritsu Kokugo Kenky ujo (Japanese National Language Research Institute), 1964. Gendai Zasshi Ky ujusshu no Yōji Yōgo [The vocabulary and Chinese characters of ninety Magazines of Today]. Sh uei Shuppan, Tokyo. Kokuritsu Kokugo Kenky ujo (Japanese National Language Research Institute), 1971. Denki-Keisanki ni Yoru Shimbun no Goi Chōsa [Studies on the vocabulary of modern newspapers], vol. I. Sh uei Shuppan, Tokyo. Kokuritsu Kokugo Kenky ujo (Japanese National Language Research Institute), 1972. Denki-Keisanki ni Yoru Shimbun no Goi Chōsa [Studies on the vocabulary of modern newspapers], vol. II. Sh uei Shuppan, Tokyo. Kokuritsu Kokugo Kenky ujo (Japanese National Language Research Institute), 1973. Denki-Keisanki ni Yoru Shimbun no Goi Chōsa [Studies on the vocabulary of modern newspapers], vol. III. Sh uei Shuppan, Tokyo. Kokuritsu Kokugo Kenky ujo (Japanese National Language Research Institute), 1974. Denki-Keisanki ni Yoru Shimbun no Goi Chōsa [Studies on the vocabulary of modern newspapers], vol. IV. Sh uei Shuppan, Tokyo. ujo, Tokyo. Kondo, Tomie, 2002. Fukusō de Tanoshimu Genji Monogatari [Appreciating the clothing of the ‘Tale of Genji’]. PHP Kenky Levy, Ian Hideo, 1981. Man’yō-sh u, vol. 1. University of Tokyo Press, Tokyo. McNeill, Noriko, 1972. Colour and colour terminology. Journal of Linguistics 8, 21–33. MacLaury, Robert E., 1992. From brightness to hue: an explanatory model of color-category evolution. Current Anthropology 33, 137–186. MacLaury, Robert E., 1995. Vantage theory. In: Taylor, John R., MacLaury, Robert E. (Eds.), Language and the Cognitive Construal of the World. Mouton de Gruyter, Berlin and New York, pp. 231–276. MacLaury, Robert E., 1997. Color and Cognition in Mesoamerica: Constructing Categories as Vantages. University of Texas Press, Austin. MacLaury, Robert E., 2001. Color terms. In: Haspelmath, Martin, König, Ekkehard, Oesterreicher, Wulf, Raible, Wolfgang (Eds.), Laguage Typology and Universals: An International Handbook. Mouton de Gruyter, Berlin and New York, pp. 1227–1251. MacLaury, Robert E., 2007. Categories of desaturated-complex color: sensorial, perceptual, and cognitive models. In: MacLaury, Robert E., Paramei, Galina., Dedrick, Don. (Eds.), The Anthropology of Color: Interdisciplinary Multilevel Modeling. John Benjamins, Amsterdam/Philadelphia, pp. 125–150. MacLaury, Robert E., Paramei, Galina, Dedrick, Don (Eds.), 2007. The Anthropology of Color: Interdisciplinary Multilevel Modeling. John Benjamins, Amsterdam/Philadelphia. J. Stanlaw / Language Sciences 32 (2010) 196–224 223 MacLaury, Robert E., Almási, Judit, Kövecses, Zoltán, 1997. Hungarian piros and vörös: color from points of view. Semiotica, 67–81. MacLaury, Robert E. (Ed.), 2002. Vantage theory: applications in linguistic cognition. Language Sciences 24 (5–6) (special issue). Masuda, Koh (Ed.), 1974. Kenkyusha’s New Japanese–English Dictionary. Kenkyusha, Tokyo. Matsuoka, Takeshi, 1999. Shikisai to Shinri Omoshiro Jiten (Shite Yaku Tatsu) [Interesting dictionary of hue and mental state (which are useful to know)], revised ed. Mikasa-shobō, Tokyo. Nagasaki, Seiki, 1983. Nihon Dentō Shikisai-kō [A study of Japanese traditional colors]. Kyoto-shoin, Tokyo. Nagasaki, Seiki, 2001. Nihon no Dentō Shoku: Sono Iro to Shiki-shō [Japanese traditional colors: Their color names and color tones]. Seigensha, Tokyo. Nagata, Yasuhiro, 2002. Iro no Techō: Nihon no Go-hyaku-shoku [Notes on color: 500 Japanese colors]. Shōkagukan, Tokyo. Nagumo, Haruyoshi, 2008. Iro no Atarashi Toraekata: Genba de “Tsukaeru” Shikisairon [New ways of viewing color: Color theory for use in the real world]. Shōkagukan, Tokyo. Nakae, Katsumi, 2003. Iro no Namae de Yomi Toku Nihon-shi [Understanding Japanese history through color names]. Seishun-shuppan, Tokyo. Nihon Shikisai Kenky ujo (Japan Color Research Institute), 1984. Shikimei Shōji-ten, Kaiteiban [Concise manual of color names, revised version]. Nihon Shiki Kenjigyō Kabushikigaisha, Tokyo. Nippon Gakujutsu Shinkōkai (Japan Society for the Promotion of Scientific Research), 1965. The Man’yō-sh u. Columbia University Press, New York. Ögen, E., Davies, I.R.L., 1998. Turkish color terms: tests of Berlin and Kay’s theory of color universals and linguistic relativity. Linguistics 36, 919–956. Ōmi, Gentaroō, 2008. Iro no Namae ni Kokoro Yomu: Shikimeigaku Ny umon [Understanding the thought behind color naming: An introduction to color nomenclature theory]. Kenshusha, Tokyo. Ōoka, Makoto (Ed.), 1979. Nihon no Iro: Iro no Reikishi to Bunka [Japanese color: The history and culture of color]. Asahi-shimbun-sha, Tokyo. Ōoka, Makoto et al., 1980. Nihon no Iro: Reikishi to Bunka o-Irodoru [Japanese color: Coloring history and culture]. Kōdansha, Tokyo. Ōyama, Tadasu, 1994. Shikisai Shinrigaku Ny umon [Introduction to color psychology]. Ch uō Kōron-shinsha, Tokyo. Ozaki, Saeko, 2003. Onna o Migaku Chishiki-gaku [The science of how color develops womanhood]. Riyon-sha, Tokyo. Paramei, Galina, 2007. Russian “blues:” Controversies of basicness. In: MacLaury, Robert E., Paramei, Galina, Dedrick, Don (Eds.), The Anthropology of Color: Interdisciplinary Multilevel Modeling. John Benjamins, Amsterdam/Philadelpia, pp. 75–106. Sakamoto, Tarō., Ienaga, Saburō., Inoue, Mitsusada., Ōno, Susumu. (Eds.), ihon Koten Bungaku Taikei: Nihon-shoki [The great collection of Japanese literary classics: The Nihon-shoki], vol. 2, Springer, berlin, 1967. Satake, Akihiro, n.d. [The character of color terms in ancient Japanese.] Ms. Saunders, Barbara A.C., 1992. The Invention of Basic Color Terms. ISOR (Interdisciplinary Social Science Research Institute), Utrecht. Saunders, Barbara A.C., 2000. Reply to Hardin, Maffi and Kay. American Anthropologist 102, 323. Saunders, Barbara A.C., 2007. Towards a new typology of color. In: MacLaury, Robert E., Paramei, Galina, Dedrick, Don (Eds.), The Anthropology of Color: Interdisciplinary Multilevel Modeling. John Benjamins, Amsterdam/Philadelpia, pp. 467–479. Saunders, Barbara A.C., van Brakel, Jaap, 1988. Re-evaluating basic color terms. Cultural Dynamics 1, 359–378. Saunders, Barbara A.C., van Brakel, Jaap, 1997. Are there non-trivial constraints on color categorization? Brain and Behavioral Sciences 20 (2), 167–212. Saunders, Barbara A.C., van Brakel, Jaap (Eds.), 2002. Theories, Technologies, Instrumentalities of Color: Anthropological and Historiographic Perspectives. University Press of America, Lanham, NY. Shibugawa, Ikuyoshi, Takahashi, Yumi, 1983a. Hai-shoku Jiten [Color scheme dictionary]. Kazan Shoboo Shin-sha, Tokyo. Shin-pen Kokka Taikan (The National Literary Collection, New Edition), 1983a. Chokusensh u-hen: Kash u, [Imperial collections: Poems], vol. I. Kadokawa, Tokyo. Shin-pen Kokka Taikan (The National Literary Collection, New Edition), 1983b. Chokusensh u-hen: Sakuin, [Imperial collections: Index], vol. II. Kadokawa, Tokyo. Shin-pen Kokka Taikan (The National Literary Collection, New Edition), 1984a. Shisens u-hen: Kash u[Private collections: Poems], vol. I. Kadokawa, Tokyo. Shin-pen Kokka Taikan (The National Literary Collection, New Edition), 1984b. Shisensh u-hen: Sakuin [Private collections: Index], vol. II. Kadokawa, Tokyo. Shinmura, Izuru (Ed.), 1998. Kōjien [Comprehensive Japanese Dictionary], fifth ed. Iwanami Shoten, Tokyo. Stanlaw, James, 1987. Colors, culture, and contact: English loanwords and problems of color nomenclature in modern Japanese. Ph.D. dissertation, Department of Anthropology, University of Illinois. Stanlaw, James, 1992a. Commentary on MacLaury’s From brightness To hue: An explanatory model of color-category evolution. Current Anthropology 33 (2), 167–168. Stanlaw, James, 1992b. Make mine BURAKU: Shedding light on brightness in Japanese-English color terminology. Paper presented at the 91st Annual Meetings of the American Anthroplogical Association. San Francisco, December 2–7. Stanlaw, James, 1993. Review of The Invention of Basic Color Terms by Barbara Ann Saunders. American Anthropologist 95, 183–185. Stanlaw, James, 1997a. Two observations on culture contact and the Japanese color nomenclature system. In: Hardin, C.L., Maffi, Luisa (Eds.), Color Categories in Thought and Language. Cambridge University Press, Cambridge, pp. 240–260. Stanlaw, James, 1997b. Making light of keeping color categories in the dark: Some arguments against the notion of trivial constraints in color nomenclature of Saunders and van Brakel. Behavioral and Brain Research 20 (2), 208–209. Stanlaw, James, 2004. Japanese and English Language and Culture Contact. Hong Kong University Press, Hong Kong. 224 J. Stanlaw / Language Sciences 32 (2010) 196–224 Stanlaw, James, 2007. Japanese color terms, from 400 CE to the present: Literature, orthography, and language contact in light of current cognitive theory. In: MacLaury, Robert E., Paramei, Galina, Dedrick, Don (Eds.), The Anthropology of Color: Interdisciplinary Multilevel Modeling. John Benjamins, Amsterdam/Philadelpia, pp. 295–318. Takimoto, Takao, Fujisawa, Hideaki, 1977. Ny umon-shikisai Shinri-gaku [Introduction to color psychology]. Dai-nihon Tosho, Tokyo. Tanaka, Ikko, Koike, Kazuko, 1982. Japan Color. Libro Port Co., Ltd., Tokyo. Uchikawa, Keiji, Boynton, Robert, 1987. Categorical color perception of Japanese observers: Comparisons with that of Americans. Vision Research 27, 1825–1833. Uemura, Rokuro, Yamazaki, Katuhiro (Eds.), 1950. (Kaitei) Nihon Shikimei Takian [(Revised) Japanese color term encyclopedia]. Yōtokusha, Kyoto. ), 2003. Wa-shoku list of Japanese color terms. <http://tenka1.jp/~karasu_2/jcolor_list.htm>. Wa-shoku, Risuto, ( Witkowski, Stanley, Brown, Cecil, 1977. An explanation of color nomenclature universals. American Anthropologist 79, 50–57. ] Serabı [Color therapy.] KK Rongu Serāzu, Tokyo. Yamanouchi, Nobuko, 2001. Karā [but using the characters for Yoshioka, Sachio, 2002. Nihon no Iro o Someru (Dyeing Japanese colors). Iwanami-shinsho, Tokyo. Yoshioka, Sachio, 2007. Nihon no Iro o Aruku [A walk through Japan’s color’s world]. Heibonsha Shinsho, Tokyo.